OpenCloudOS-Kernel/drivers/gpu/drm/i915/i915_irq.c

4704 lines
131 KiB
C
Raw Normal View History

/* i915_irq.c -- IRQ support for the I915 -*- linux-c -*-
*/
/*
* Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
* All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sub license, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial portions
* of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
* IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
* ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/sysrq.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
#include <linux/circ_buf.h>
#include <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_trace.h"
#include "intel_drv.h"
/**
* DOC: interrupt handling
*
* These functions provide the basic support for enabling and disabling the
* interrupt handling support. There's a lot more functionality in i915_irq.c
* and related files, but that will be described in separate chapters.
*/
static const u32 hpd_ilk[HPD_NUM_PINS] = {
[HPD_PORT_A] = DE_DP_A_HOTPLUG,
};
static const u32 hpd_ivb[HPD_NUM_PINS] = {
[HPD_PORT_A] = DE_DP_A_HOTPLUG_IVB,
};
static const u32 hpd_bdw[HPD_NUM_PINS] = {
[HPD_PORT_A] = GEN8_PORT_DP_A_HOTPLUG,
};
static const u32 hpd_ibx[HPD_NUM_PINS] = {
[HPD_CRT] = SDE_CRT_HOTPLUG,
[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG,
[HPD_PORT_B] = SDE_PORTB_HOTPLUG,
[HPD_PORT_C] = SDE_PORTC_HOTPLUG,
[HPD_PORT_D] = SDE_PORTD_HOTPLUG
};
static const u32 hpd_cpt[HPD_NUM_PINS] = {
[HPD_CRT] = SDE_CRT_HOTPLUG_CPT,
[HPD_SDVO_B] = SDE_SDVOB_HOTPLUG_CPT,
[HPD_PORT_B] = SDE_PORTB_HOTPLUG_CPT,
[HPD_PORT_C] = SDE_PORTC_HOTPLUG_CPT,
[HPD_PORT_D] = SDE_PORTD_HOTPLUG_CPT
};
static const u32 hpd_spt[HPD_NUM_PINS] = {
[HPD_PORT_A] = SDE_PORTA_HOTPLUG_SPT,
[HPD_PORT_B] = SDE_PORTB_HOTPLUG_CPT,
[HPD_PORT_C] = SDE_PORTC_HOTPLUG_CPT,
[HPD_PORT_D] = SDE_PORTD_HOTPLUG_CPT,
[HPD_PORT_E] = SDE_PORTE_HOTPLUG_SPT
};
static const u32 hpd_mask_i915[HPD_NUM_PINS] = {
[HPD_CRT] = CRT_HOTPLUG_INT_EN,
[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_EN,
[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_EN,
[HPD_PORT_B] = PORTB_HOTPLUG_INT_EN,
[HPD_PORT_C] = PORTC_HOTPLUG_INT_EN,
[HPD_PORT_D] = PORTD_HOTPLUG_INT_EN
};
static const u32 hpd_status_g4x[HPD_NUM_PINS] = {
[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_G4X,
[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_G4X,
[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
};
static const u32 hpd_status_i915[HPD_NUM_PINS] = {
[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I915,
[HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I915,
[HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
[HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
};
/* BXT hpd list */
static const u32 hpd_bxt[HPD_NUM_PINS] = {
[HPD_PORT_A] = BXT_DE_PORT_HP_DDIA,
[HPD_PORT_B] = BXT_DE_PORT_HP_DDIB,
[HPD_PORT_C] = BXT_DE_PORT_HP_DDIC
};
/* IIR can theoretically queue up two events. Be paranoid. */
#define GEN8_IRQ_RESET_NDX(type, which) do { \
I915_WRITE(GEN8_##type##_IMR(which), 0xffffffff); \
POSTING_READ(GEN8_##type##_IMR(which)); \
I915_WRITE(GEN8_##type##_IER(which), 0); \
I915_WRITE(GEN8_##type##_IIR(which), 0xffffffff); \
POSTING_READ(GEN8_##type##_IIR(which)); \
I915_WRITE(GEN8_##type##_IIR(which), 0xffffffff); \
POSTING_READ(GEN8_##type##_IIR(which)); \
} while (0)
#define GEN5_IRQ_RESET(type) do { \
I915_WRITE(type##IMR, 0xffffffff); \
POSTING_READ(type##IMR); \
I915_WRITE(type##IER, 0); \
I915_WRITE(type##IIR, 0xffffffff); \
POSTING_READ(type##IIR); \
I915_WRITE(type##IIR, 0xffffffff); \
POSTING_READ(type##IIR); \
} while (0)
/*
* We should clear IMR at preinstall/uninstall, and just check at postinstall.
*/
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
static void gen5_assert_iir_is_zero(struct drm_i915_private *dev_priv,
i915_reg_t reg)
{
u32 val = I915_READ(reg);
if (val == 0)
return;
WARN(1, "Interrupt register 0x%x is not zero: 0x%08x\n",
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_mmio_reg_offset(reg), val);
I915_WRITE(reg, 0xffffffff);
POSTING_READ(reg);
I915_WRITE(reg, 0xffffffff);
POSTING_READ(reg);
}
#define GEN8_IRQ_INIT_NDX(type, which, imr_val, ier_val) do { \
gen5_assert_iir_is_zero(dev_priv, GEN8_##type##_IIR(which)); \
I915_WRITE(GEN8_##type##_IER(which), (ier_val)); \
I915_WRITE(GEN8_##type##_IMR(which), (imr_val)); \
POSTING_READ(GEN8_##type##_IMR(which)); \
} while (0)
#define GEN5_IRQ_INIT(type, imr_val, ier_val) do { \
gen5_assert_iir_is_zero(dev_priv, type##IIR); \
I915_WRITE(type##IER, (ier_val)); \
I915_WRITE(type##IMR, (imr_val)); \
POSTING_READ(type##IMR); \
} while (0)
static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir);
/* For display hotplug interrupt */
static inline void
i915_hotplug_interrupt_update_locked(struct drm_i915_private *dev_priv,
uint32_t mask,
uint32_t bits)
{
uint32_t val;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(bits & ~mask);
val = I915_READ(PORT_HOTPLUG_EN);
val &= ~mask;
val |= bits;
I915_WRITE(PORT_HOTPLUG_EN, val);
}
/**
* i915_hotplug_interrupt_update - update hotplug interrupt enable
* @dev_priv: driver private
* @mask: bits to update
* @bits: bits to enable
* NOTE: the HPD enable bits are modified both inside and outside
* of an interrupt context. To avoid that read-modify-write cycles
* interfer, these bits are protected by a spinlock. Since this
* function is usually not called from a context where the lock is
* held already, this function acquires the lock itself. A non-locking
* version is also available.
*/
void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv,
uint32_t mask,
uint32_t bits)
{
spin_lock_irq(&dev_priv->irq_lock);
i915_hotplug_interrupt_update_locked(dev_priv, mask, bits);
spin_unlock_irq(&dev_priv->irq_lock);
}
/**
* ilk_update_display_irq - update DEIMR
* @dev_priv: driver private
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
void ilk_update_display_irq(struct drm_i915_private *dev_priv,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
uint32_t new_val;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(enabled_irq_mask & ~interrupt_mask);
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
return;
new_val = dev_priv->irq_mask;
new_val &= ~interrupt_mask;
new_val |= (~enabled_irq_mask & interrupt_mask);
if (new_val != dev_priv->irq_mask) {
dev_priv->irq_mask = new_val;
I915_WRITE(DEIMR, dev_priv->irq_mask);
POSTING_READ(DEIMR);
}
}
/**
* ilk_update_gt_irq - update GTIMR
* @dev_priv: driver private
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
static void ilk_update_gt_irq(struct drm_i915_private *dev_priv,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(enabled_irq_mask & ~interrupt_mask);
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
return;
dev_priv->gt_irq_mask &= ~interrupt_mask;
dev_priv->gt_irq_mask |= (~enabled_irq_mask & interrupt_mask);
I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
POSTING_READ(GTIMR);
}
void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
{
ilk_update_gt_irq(dev_priv, mask, mask);
}
void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask)
{
ilk_update_gt_irq(dev_priv, mask, 0);
}
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
static i915_reg_t gen6_pm_iir(struct drm_i915_private *dev_priv)
{
return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IIR(2) : GEN6_PMIIR;
}
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
static i915_reg_t gen6_pm_imr(struct drm_i915_private *dev_priv)
{
return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IMR(2) : GEN6_PMIMR;
}
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
static i915_reg_t gen6_pm_ier(struct drm_i915_private *dev_priv)
{
return INTEL_INFO(dev_priv)->gen >= 8 ? GEN8_GT_IER(2) : GEN6_PMIER;
}
/**
* snb_update_pm_irq - update GEN6_PMIMR
* @dev_priv: driver private
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
uint32_t new_val;
WARN_ON(enabled_irq_mask & ~interrupt_mask);
assert_spin_locked(&dev_priv->irq_lock);
new_val = dev_priv->pm_irq_mask;
new_val &= ~interrupt_mask;
new_val |= (~enabled_irq_mask & interrupt_mask);
if (new_val != dev_priv->pm_irq_mask) {
dev_priv->pm_irq_mask = new_val;
I915_WRITE(gen6_pm_imr(dev_priv), dev_priv->pm_irq_mask);
POSTING_READ(gen6_pm_imr(dev_priv));
}
}
void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
{
drm/i915: mask RPS IRQs properly when disabling RPS Atm, igt/gem_reset_stats can trigger the recently added WARN on left-over PM_IIR bits in gen6_enable_rps_interrupts(). There are two reasons for this: 1. we call intel_enable_gt_powersave() without a preceeding intel_disable_gt_powersave() 2. gen6_disable_rps_interrupts() doesn't mask interrupts in PM_IMR 1. means RPS interrupts will remain enabled and can be serviced during the HW initialization after a GPU reset. 2. means even if we called gen6_disable_rps_interrupts() any new RPS interrupt during RPS initialization would still propagate to PM_IIR too early (though wouldn't be serviced). This patch solves the 2. issue by also masking interrupts in PM_IMR, the following patch fixes 1. getting rid of the WARN. This also makes intel_enable_gt_powersave() and intel_disable_gt_powersave() more symmetric. Since gen6_disable_rps_interrupts() is called during driver loading with i915 interrupts disabled add a new version of gen6_disable_pm_irq() that doesn't WARN for this. Also while at it, get the irq_lock around the whole PM_IMR/IER/IIR programming sequence and make sure that any queued PM_IIR bit is also cleared. The WARN was caught by PRTS after I sent my previous RPS sanitizing patchset and I could easily reproduce it on HSW. To actually fix it we also need the next patch. Reported-by: He, Shuang <shuang.he@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-21 05:01:47 +08:00
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
return;
snb_update_pm_irq(dev_priv, mask, mask);
}
drm/i915: mask RPS IRQs properly when disabling RPS Atm, igt/gem_reset_stats can trigger the recently added WARN on left-over PM_IIR bits in gen6_enable_rps_interrupts(). There are two reasons for this: 1. we call intel_enable_gt_powersave() without a preceeding intel_disable_gt_powersave() 2. gen6_disable_rps_interrupts() doesn't mask interrupts in PM_IMR 1. means RPS interrupts will remain enabled and can be serviced during the HW initialization after a GPU reset. 2. means even if we called gen6_disable_rps_interrupts() any new RPS interrupt during RPS initialization would still propagate to PM_IIR too early (though wouldn't be serviced). This patch solves the 2. issue by also masking interrupts in PM_IMR, the following patch fixes 1. getting rid of the WARN. This also makes intel_enable_gt_powersave() and intel_disable_gt_powersave() more symmetric. Since gen6_disable_rps_interrupts() is called during driver loading with i915 interrupts disabled add a new version of gen6_disable_pm_irq() that doesn't WARN for this. Also while at it, get the irq_lock around the whole PM_IMR/IER/IIR programming sequence and make sure that any queued PM_IIR bit is also cleared. The WARN was caught by PRTS after I sent my previous RPS sanitizing patchset and I could easily reproduce it on HSW. To actually fix it we also need the next patch. Reported-by: He, Shuang <shuang.he@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-21 05:01:47 +08:00
static void __gen6_disable_pm_irq(struct drm_i915_private *dev_priv,
uint32_t mask)
{
snb_update_pm_irq(dev_priv, mask, 0);
}
drm/i915: mask RPS IRQs properly when disabling RPS Atm, igt/gem_reset_stats can trigger the recently added WARN on left-over PM_IIR bits in gen6_enable_rps_interrupts(). There are two reasons for this: 1. we call intel_enable_gt_powersave() without a preceeding intel_disable_gt_powersave() 2. gen6_disable_rps_interrupts() doesn't mask interrupts in PM_IMR 1. means RPS interrupts will remain enabled and can be serviced during the HW initialization after a GPU reset. 2. means even if we called gen6_disable_rps_interrupts() any new RPS interrupt during RPS initialization would still propagate to PM_IIR too early (though wouldn't be serviced). This patch solves the 2. issue by also masking interrupts in PM_IMR, the following patch fixes 1. getting rid of the WARN. This also makes intel_enable_gt_powersave() and intel_disable_gt_powersave() more symmetric. Since gen6_disable_rps_interrupts() is called during driver loading with i915 interrupts disabled add a new version of gen6_disable_pm_irq() that doesn't WARN for this. Also while at it, get the irq_lock around the whole PM_IMR/IER/IIR programming sequence and make sure that any queued PM_IIR bit is also cleared. The WARN was caught by PRTS after I sent my previous RPS sanitizing patchset and I could easily reproduce it on HSW. To actually fix it we also need the next patch. Reported-by: He, Shuang <shuang.he@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-21 05:01:47 +08:00
void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, uint32_t mask)
{
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
return;
__gen6_disable_pm_irq(dev_priv, mask);
}
void gen6_reset_rps_interrupts(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = gen6_pm_iir(dev_priv);
spin_lock_irq(&dev_priv->irq_lock);
I915_WRITE(reg, dev_priv->pm_rps_events);
I915_WRITE(reg, dev_priv->pm_rps_events);
POSTING_READ(reg);
dev_priv->rps.pm_iir = 0;
spin_unlock_irq(&dev_priv->irq_lock);
}
void gen6_enable_rps_interrupts(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
spin_lock_irq(&dev_priv->irq_lock);
WARN_ON(dev_priv->rps.pm_iir);
WARN_ON(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events);
drm/i915: sanitize rps irq disabling When disabling the RPS interrupts there is a tricky dependency between the thread disabling the interrupts, the RPS interrupt handler and the corresponding RPS work. The RPS work can reenable the interrupts, so there is no straightforward order in the disabling thread to (1) make sure that any RPS work is flushed and to (2) disable all RPS interrupts. Currently this is solved by masking the interrupts using two separate mask registers (first level display IMR and PM IMR) and doing the disabling when all first level interrupts are disabled. This works, but the requirement to run with all first level interrupts disabled is unnecessary making the suspend / unload time ordering of RPS disabling wrt. other unitialization steps difficult and error prone. Removing this restriction allows us to disable RPS early during suspend / unload and forget about it for the rest of the sequence. By adding a more explicit method for avoiding the above race, it also becomes easier to prove its correctness. Finally currently we can hit the WARN in snb_update_pm_irq(), when a final RPS work runs with the first level interrupts already disabled. This won't lead to any problem (due to the separate interrupt masks), but with the change in this and the next patch we can get rid of the WARN, while leaving it in place for other scenarios. To address the above points, add a new RPS interrupts_enabled flag and use this during RPS disabling to avoid requeuing the RPS work and reenabling of the RPS interrupts. Since the interrupt disabling happens now in intel_suspend_gt_powersave(), we will disable RPS interrupts explicitly during suspend (and not just through the first level mask), but there is no problem doing so, it's also more consistent and allows us to unify more of the RPS disabling during suspend and unload time in the next patch. v2/v3: - rebase on patch "drm/i915: move rps irq disable one level up" in the patchset Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-19 21:30:04 +08:00
dev_priv->rps.interrupts_enabled = true;
I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) |
dev_priv->pm_rps_events);
gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
spin_unlock_irq(&dev_priv->irq_lock);
}
u32 gen6_sanitize_rps_pm_mask(struct drm_i915_private *dev_priv, u32 mask)
{
/*
* SNB,IVB can while VLV,CHV may hard hang on looping batchbuffer
* if GEN6_PM_UP_EI_EXPIRED is masked.
*
* TODO: verify if this can be reproduced on VLV,CHV.
*/
if (INTEL_INFO(dev_priv)->gen <= 7 && !IS_HASWELL(dev_priv))
mask &= ~GEN6_PM_RP_UP_EI_EXPIRED;
if (INTEL_INFO(dev_priv)->gen >= 8)
mask &= ~GEN8_PMINTR_REDIRECT_TO_NON_DISP;
return mask;
}
void gen6_disable_rps_interrupts(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: sanitize rps irq disabling When disabling the RPS interrupts there is a tricky dependency between the thread disabling the interrupts, the RPS interrupt handler and the corresponding RPS work. The RPS work can reenable the interrupts, so there is no straightforward order in the disabling thread to (1) make sure that any RPS work is flushed and to (2) disable all RPS interrupts. Currently this is solved by masking the interrupts using two separate mask registers (first level display IMR and PM IMR) and doing the disabling when all first level interrupts are disabled. This works, but the requirement to run with all first level interrupts disabled is unnecessary making the suspend / unload time ordering of RPS disabling wrt. other unitialization steps difficult and error prone. Removing this restriction allows us to disable RPS early during suspend / unload and forget about it for the rest of the sequence. By adding a more explicit method for avoiding the above race, it also becomes easier to prove its correctness. Finally currently we can hit the WARN in snb_update_pm_irq(), when a final RPS work runs with the first level interrupts already disabled. This won't lead to any problem (due to the separate interrupt masks), but with the change in this and the next patch we can get rid of the WARN, while leaving it in place for other scenarios. To address the above points, add a new RPS interrupts_enabled flag and use this during RPS disabling to avoid requeuing the RPS work and reenabling of the RPS interrupts. Since the interrupt disabling happens now in intel_suspend_gt_powersave(), we will disable RPS interrupts explicitly during suspend (and not just through the first level mask), but there is no problem doing so, it's also more consistent and allows us to unify more of the RPS disabling during suspend and unload time in the next patch. v2/v3: - rebase on patch "drm/i915: move rps irq disable one level up" in the patchset Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-19 21:30:04 +08:00
spin_lock_irq(&dev_priv->irq_lock);
dev_priv->rps.interrupts_enabled = false;
spin_unlock_irq(&dev_priv->irq_lock);
cancel_work_sync(&dev_priv->rps.work);
drm/i915: mask RPS IRQs properly when disabling RPS Atm, igt/gem_reset_stats can trigger the recently added WARN on left-over PM_IIR bits in gen6_enable_rps_interrupts(). There are two reasons for this: 1. we call intel_enable_gt_powersave() without a preceeding intel_disable_gt_powersave() 2. gen6_disable_rps_interrupts() doesn't mask interrupts in PM_IMR 1. means RPS interrupts will remain enabled and can be serviced during the HW initialization after a GPU reset. 2. means even if we called gen6_disable_rps_interrupts() any new RPS interrupt during RPS initialization would still propagate to PM_IIR too early (though wouldn't be serviced). This patch solves the 2. issue by also masking interrupts in PM_IMR, the following patch fixes 1. getting rid of the WARN. This also makes intel_enable_gt_powersave() and intel_disable_gt_powersave() more symmetric. Since gen6_disable_rps_interrupts() is called during driver loading with i915 interrupts disabled add a new version of gen6_disable_pm_irq() that doesn't WARN for this. Also while at it, get the irq_lock around the whole PM_IMR/IER/IIR programming sequence and make sure that any queued PM_IIR bit is also cleared. The WARN was caught by PRTS after I sent my previous RPS sanitizing patchset and I could easily reproduce it on HSW. To actually fix it we also need the next patch. Reported-by: He, Shuang <shuang.he@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-21 05:01:47 +08:00
spin_lock_irq(&dev_priv->irq_lock);
I915_WRITE(GEN6_PMINTRMSK, gen6_sanitize_rps_pm_mask(dev_priv, ~0));
drm/i915: mask RPS IRQs properly when disabling RPS Atm, igt/gem_reset_stats can trigger the recently added WARN on left-over PM_IIR bits in gen6_enable_rps_interrupts(). There are two reasons for this: 1. we call intel_enable_gt_powersave() without a preceeding intel_disable_gt_powersave() 2. gen6_disable_rps_interrupts() doesn't mask interrupts in PM_IMR 1. means RPS interrupts will remain enabled and can be serviced during the HW initialization after a GPU reset. 2. means even if we called gen6_disable_rps_interrupts() any new RPS interrupt during RPS initialization would still propagate to PM_IIR too early (though wouldn't be serviced). This patch solves the 2. issue by also masking interrupts in PM_IMR, the following patch fixes 1. getting rid of the WARN. This also makes intel_enable_gt_powersave() and intel_disable_gt_powersave() more symmetric. Since gen6_disable_rps_interrupts() is called during driver loading with i915 interrupts disabled add a new version of gen6_disable_pm_irq() that doesn't WARN for this. Also while at it, get the irq_lock around the whole PM_IMR/IER/IIR programming sequence and make sure that any queued PM_IIR bit is also cleared. The WARN was caught by PRTS after I sent my previous RPS sanitizing patchset and I could easily reproduce it on HSW. To actually fix it we also need the next patch. Reported-by: He, Shuang <shuang.he@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-21 05:01:47 +08:00
__gen6_disable_pm_irq(dev_priv, dev_priv->pm_rps_events);
I915_WRITE(gen6_pm_ier(dev_priv), I915_READ(gen6_pm_ier(dev_priv)) &
~dev_priv->pm_rps_events);
spin_unlock_irq(&dev_priv->irq_lock);
synchronize_irq(dev->irq);
}
/**
* bdw_update_port_irq - update DE port interrupt
* @dev_priv: driver private
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
static void bdw_update_port_irq(struct drm_i915_private *dev_priv,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
uint32_t new_val;
uint32_t old_val;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(enabled_irq_mask & ~interrupt_mask);
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
return;
old_val = I915_READ(GEN8_DE_PORT_IMR);
new_val = old_val;
new_val &= ~interrupt_mask;
new_val |= (~enabled_irq_mask & interrupt_mask);
if (new_val != old_val) {
I915_WRITE(GEN8_DE_PORT_IMR, new_val);
POSTING_READ(GEN8_DE_PORT_IMR);
}
}
/**
* bdw_update_pipe_irq - update DE pipe interrupt
* @dev_priv: driver private
* @pipe: pipe whose interrupt to update
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
void bdw_update_pipe_irq(struct drm_i915_private *dev_priv,
enum pipe pipe,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
uint32_t new_val;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(enabled_irq_mask & ~interrupt_mask);
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
return;
new_val = dev_priv->de_irq_mask[pipe];
new_val &= ~interrupt_mask;
new_val |= (~enabled_irq_mask & interrupt_mask);
if (new_val != dev_priv->de_irq_mask[pipe]) {
dev_priv->de_irq_mask[pipe] = new_val;
I915_WRITE(GEN8_DE_PIPE_IMR(pipe), dev_priv->de_irq_mask[pipe]);
POSTING_READ(GEN8_DE_PIPE_IMR(pipe));
}
}
/**
* ibx_display_interrupt_update - update SDEIMR
* @dev_priv: driver private
* @interrupt_mask: mask of interrupt bits to update
* @enabled_irq_mask: mask of interrupt bits to enable
*/
void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
uint32_t interrupt_mask,
uint32_t enabled_irq_mask)
{
uint32_t sdeimr = I915_READ(SDEIMR);
sdeimr &= ~interrupt_mask;
sdeimr |= (~enabled_irq_mask & interrupt_mask);
WARN_ON(enabled_irq_mask & ~interrupt_mask);
assert_spin_locked(&dev_priv->irq_lock);
if (WARN_ON(!intel_irqs_enabled(dev_priv)))
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
return;
I915_WRITE(SDEIMR, sdeimr);
POSTING_READ(SDEIMR);
}
static void
__i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
u32 enable_mask, u32 status_mask)
{
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = PIPESTAT(pipe);
u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(!intel_irqs_enabled(dev_priv));
if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
status_mask & ~PIPESTAT_INT_STATUS_MASK,
"pipe %c: enable_mask=0x%x, status_mask=0x%x\n",
pipe_name(pipe), enable_mask, status_mask))
return;
if ((pipestat & enable_mask) == enable_mask)
return;
dev_priv->pipestat_irq_mask[pipe] |= status_mask;
/* Enable the interrupt, clear any pending status */
pipestat |= enable_mask | status_mask;
I915_WRITE(reg, pipestat);
POSTING_READ(reg);
}
static void
__i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
u32 enable_mask, u32 status_mask)
{
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = PIPESTAT(pipe);
u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
assert_spin_locked(&dev_priv->irq_lock);
WARN_ON(!intel_irqs_enabled(dev_priv));
if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
status_mask & ~PIPESTAT_INT_STATUS_MASK,
"pipe %c: enable_mask=0x%x, status_mask=0x%x\n",
pipe_name(pipe), enable_mask, status_mask))
return;
if ((pipestat & enable_mask) == 0)
return;
dev_priv->pipestat_irq_mask[pipe] &= ~status_mask;
pipestat &= ~enable_mask;
I915_WRITE(reg, pipestat);
POSTING_READ(reg);
}
static u32 vlv_get_pipestat_enable_mask(struct drm_device *dev, u32 status_mask)
{
u32 enable_mask = status_mask << 16;
/*
* On pipe A we don't support the PSR interrupt yet,
* on pipe B and C the same bit MBZ.
*/
if (WARN_ON_ONCE(status_mask & PIPE_A_PSR_STATUS_VLV))
return 0;
/*
* On pipe B and C we don't support the PSR interrupt yet, on pipe
* A the same bit is for perf counters which we don't use either.
*/
if (WARN_ON_ONCE(status_mask & PIPE_B_PSR_STATUS_VLV))
return 0;
enable_mask &= ~(PIPE_FIFO_UNDERRUN_STATUS |
SPRITE0_FLIP_DONE_INT_EN_VLV |
SPRITE1_FLIP_DONE_INT_EN_VLV);
if (status_mask & SPRITE0_FLIP_DONE_INT_STATUS_VLV)
enable_mask |= SPRITE0_FLIP_DONE_INT_EN_VLV;
if (status_mask & SPRITE1_FLIP_DONE_INT_STATUS_VLV)
enable_mask |= SPRITE1_FLIP_DONE_INT_EN_VLV;
return enable_mask;
}
void
i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
u32 status_mask)
{
u32 enable_mask;
if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
enable_mask = vlv_get_pipestat_enable_mask(dev_priv->dev,
status_mask);
else
enable_mask = status_mask << 16;
__i915_enable_pipestat(dev_priv, pipe, enable_mask, status_mask);
}
void
i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
u32 status_mask)
{
u32 enable_mask;
if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
enable_mask = vlv_get_pipestat_enable_mask(dev_priv->dev,
status_mask);
else
enable_mask = status_mask << 16;
__i915_disable_pipestat(dev_priv, pipe, enable_mask, status_mask);
}
/**
* i915_enable_asle_pipestat - enable ASLE pipestat for OpRegion
* @dev: drm device
*/
static void i915_enable_asle_pipestat(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv->opregion.asle || !IS_MOBILE(dev))
return;
spin_lock_irq(&dev_priv->irq_lock);
i915_enable_pipestat(dev_priv, PIPE_B, PIPE_LEGACY_BLC_EVENT_STATUS);
if (INTEL_INFO(dev)->gen >= 4)
i915_enable_pipestat(dev_priv, PIPE_A,
PIPE_LEGACY_BLC_EVENT_STATUS);
spin_unlock_irq(&dev_priv->irq_lock);
}
/*
* This timing diagram depicts the video signal in and
* around the vertical blanking period.
*
* Assumptions about the fictitious mode used in this example:
* vblank_start >= 3
* vsync_start = vblank_start + 1
* vsync_end = vblank_start + 2
* vtotal = vblank_start + 3
*
* start of vblank:
* latch double buffered registers
* increment frame counter (ctg+)
* generate start of vblank interrupt (gen4+)
* |
* | frame start:
* | generate frame start interrupt (aka. vblank interrupt) (gmch)
* | may be shifted forward 1-3 extra lines via PIPECONF
* | |
* | | start of vsync:
* | | generate vsync interrupt
* | | |
* ___xxxx___ ___xxxx___ ___xxxx___ ___xxxx___ ___xxxx___ ___xxxx
* . \hs/ . \hs/ \hs/ \hs/ . \hs/
* ----va---> <-----------------vb--------------------> <--------va-------------
* | | <----vs-----> |
* -vbs-----> <---vbs+1---> <---vbs+2---> <-----0-----> <-----1-----> <-----2--- (scanline counter gen2)
* -vbs-2---> <---vbs-1---> <---vbs-----> <---vbs+1---> <---vbs+2---> <-----0--- (scanline counter gen3+)
* -vbs-2---> <---vbs-2---> <---vbs-1---> <---vbs-----> <---vbs+1---> <---vbs+2- (scanline counter hsw+ hdmi)
* | | |
* last visible pixel first visible pixel
* | increment frame counter (gen3/4)
* pixel counter = vblank_start * htotal pixel counter = 0 (gen3/4)
*
* x = horizontal active
* _ = horizontal blanking
* hs = horizontal sync
* va = vertical active
* vb = vertical blanking
* vs = vertical sync
* vbs = vblank_start (number)
*
* Summary:
* - most events happen at the start of horizontal sync
* - frame start happens at the start of horizontal blank, 1-4 lines
* (depending on PIPECONF settings) after the start of vblank
* - gen3/4 pixel and frame counter are synchronized with the start
* of horizontal active on the first line of vertical active
*/
static u32 i8xx_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
{
/* Gen2 doesn't have a hardware frame counter */
return 0;
}
/* Called from drm generic code, passed a 'crtc', which
* we use as a pipe index
*/
static u32 i915_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t high_frame, low_frame;
u32 high1, high2, low, pixel, vbl_start, hsync_start, htotal;
struct intel_crtc *intel_crtc =
to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
const struct drm_display_mode *mode = &intel_crtc->base.hwmode;
htotal = mode->crtc_htotal;
hsync_start = mode->crtc_hsync_start;
vbl_start = mode->crtc_vblank_start;
if (mode->flags & DRM_MODE_FLAG_INTERLACE)
vbl_start = DIV_ROUND_UP(vbl_start, 2);
/* Convert to pixel count */
vbl_start *= htotal;
/* Start of vblank event occurs at start of hsync */
vbl_start -= htotal - hsync_start;
high_frame = PIPEFRAME(pipe);
low_frame = PIPEFRAMEPIXEL(pipe);
/*
* High & low register fields aren't synchronized, so make sure
* we get a low value that's stable across two reads of the high
* register.
*/
do {
high1 = I915_READ(high_frame) & PIPE_FRAME_HIGH_MASK;
low = I915_READ(low_frame);
high2 = I915_READ(high_frame) & PIPE_FRAME_HIGH_MASK;
} while (high1 != high2);
high1 >>= PIPE_FRAME_HIGH_SHIFT;
pixel = low & PIPE_PIXEL_MASK;
low >>= PIPE_FRAME_LOW_SHIFT;
/*
* The frame counter increments at beginning of active.
* Cook up a vblank counter by also checking the pixel
* counter against vblank start.
*/
return (((high1 << 8) | low) + (pixel >= vbl_start)) & 0xffffff;
}
static u32 g4x_get_vblank_counter(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
return I915_READ(PIPE_FRMCOUNT_G4X(pipe));
}
/* I915_READ_FW, only for fast reads of display block, no need for forcewake etc. */
static int __intel_get_crtc_scanline(struct intel_crtc *crtc)
{
struct drm_device *dev = crtc->base.dev;
struct drm_i915_private *dev_priv = dev->dev_private;
const struct drm_display_mode *mode = &crtc->base.hwmode;
enum pipe pipe = crtc->pipe;
int position, vtotal;
vtotal = mode->crtc_vtotal;
if (mode->flags & DRM_MODE_FLAG_INTERLACE)
vtotal /= 2;
if (IS_GEN2(dev))
position = I915_READ_FW(PIPEDSL(pipe)) & DSL_LINEMASK_GEN2;
else
position = I915_READ_FW(PIPEDSL(pipe)) & DSL_LINEMASK_GEN3;
/*
* On HSW, the DSL reg (0x70000) appears to return 0 if we
* read it just before the start of vblank. So try it again
* so we don't accidentally end up spanning a vblank frame
* increment, causing the pipe_update_end() code to squak at us.
*
* The nature of this problem means we can't simply check the ISR
* bit and return the vblank start value; nor can we use the scanline
* debug register in the transcoder as it appears to have the same
* problem. We may need to extend this to include other platforms,
* but so far testing only shows the problem on HSW.
*/
if (HAS_DDI(dev) && !position) {
int i, temp;
for (i = 0; i < 100; i++) {
udelay(1);
temp = __raw_i915_read32(dev_priv, PIPEDSL(pipe)) &
DSL_LINEMASK_GEN3;
if (temp != position) {
position = temp;
break;
}
}
}
/*
* See update_scanline_offset() for the details on the
* scanline_offset adjustment.
*/
return (position + crtc->scanline_offset) % vtotal;
}
static int i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe,
unsigned int flags, int *vpos, int *hpos,
ktime_t *stime, ktime_t *etime,
const struct drm_display_mode *mode)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
int position;
drm/i915: Fix scanout position for real Seems I've been a bit dense with regards to the start of vblank vs. the scanline counter / pixel counter. After staring at the pixel counter on gen4 I came to the conclusion that the start of vblank interrupt and scanline counter increment happen at the same time. The scanline counter increment is documented to occur at start of hsync, which means that the start of vblank interrupt must also trigger there. Looking at the pixel counter value when the scanline wraps from vtotal-1 to 0 confirms that, as the pixel counter at that point reads hsync_start. This also clarifies why we see need the +1 adjustment to the scaline counter. The counter actually starts counting from vtotal-1 on the first active line. I also confirmed that the frame start interrupt happens ~1 line after the start of vblank, but the frame start occurs at hblank_start instead. We only use the frame start interrupt on gen2 where the start of vblank interrupt isn't available. The only important thing to note here is that frame start occurs after vblank start, so we don't have to play any additional tricks to fix up the scanline counter. The other thing to note is the fact that the pixel counter on gen3-4 starts counting from the start of horizontal active on the first active line. That means that when we get the start of vblank interrupt, the pixel counter reads (htotal*(vblank_start-1)+hsync_start). Since we consider vblank to start at (htotal*vblank_start) we need to add a constant (htotal-hsync_start) offset to the pixel counter, or else we risk misdetecting whether we're in vblank or not. I talked a bit with Art Runyan about these topics, and he confirmed my findings. And that the same rules should hold for platforms which don't have the pixel counter. That's good since without the pixel counter it's rather difficult to verify the timings to this accuracy. So the conclusion is that we can throw away all the ISR tricks I added, and just increment the scanline counter by one always. Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:44 +08:00
int vbl_start, vbl_end, hsync_start, htotal, vtotal;
bool in_vbl = true;
int ret = 0;
drm/intel: Push get_scanout_position() timestamping into kms driver. Move the ktime_get() clock readouts and potential preempt_disable() calls from drm core into kms driver to make it compatible with the api changes in the drm core. The intel-kms driver needs to take the uncore.lock inside i915_get_crtc_scanoutpos() and intel_pipe_in_vblank(). This is incompatible with the preempt_disable() on a PREEMPT_RT patched kernel, as regular spin locks must not be taken within a preempt_disable'd section. Lock contention on the uncore.lock also introduced too much uncertainty in vblank timestamps. Push the ktime_get() timestamping for scanoutpos queries and potential preempt_disable_rt() into i915_get_crtc_scanoutpos(), so these problems can be avoided: 1. First lock the uncore.lock (might sleep on a PREEMPT_RT kernel). 2. preempt_disable_rt() (will be added by the rt-linux folks). 3. ktime_get() a timestamp before scanout pos query. 4. Do all mmio reads as fast as possible without grabbing any new locks! 5. ktime_get() a post-query timestamp. 6. preempt_enable_rt() 7. Unlock the uncore.lock. This reduces timestamp uncertainty on a low-end HP Atom Mini netbook with Intel GMA-950 nicely: Before: 3-8 usecs with spikes > 20 usecs, triggering query retries. After : Typically 1 usec (98% of all samples), occassionally 2 usecs (2% of all samples), with maximum of 3 usecs (a handful). v2: Fix formatting of new multi-line code comments. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-10-30 12:13:08 +08:00
unsigned long irqflags;
if (WARN_ON(!mode->crtc_clock)) {
DRM_DEBUG_DRIVER("trying to get scanoutpos for disabled "
"pipe %c\n", pipe_name(pipe));
return 0;
}
htotal = mode->crtc_htotal;
drm/i915: Fix scanout position for real Seems I've been a bit dense with regards to the start of vblank vs. the scanline counter / pixel counter. After staring at the pixel counter on gen4 I came to the conclusion that the start of vblank interrupt and scanline counter increment happen at the same time. The scanline counter increment is documented to occur at start of hsync, which means that the start of vblank interrupt must also trigger there. Looking at the pixel counter value when the scanline wraps from vtotal-1 to 0 confirms that, as the pixel counter at that point reads hsync_start. This also clarifies why we see need the +1 adjustment to the scaline counter. The counter actually starts counting from vtotal-1 on the first active line. I also confirmed that the frame start interrupt happens ~1 line after the start of vblank, but the frame start occurs at hblank_start instead. We only use the frame start interrupt on gen2 where the start of vblank interrupt isn't available. The only important thing to note here is that frame start occurs after vblank start, so we don't have to play any additional tricks to fix up the scanline counter. The other thing to note is the fact that the pixel counter on gen3-4 starts counting from the start of horizontal active on the first active line. That means that when we get the start of vblank interrupt, the pixel counter reads (htotal*(vblank_start-1)+hsync_start). Since we consider vblank to start at (htotal*vblank_start) we need to add a constant (htotal-hsync_start) offset to the pixel counter, or else we risk misdetecting whether we're in vblank or not. I talked a bit with Art Runyan about these topics, and he confirmed my findings. And that the same rules should hold for platforms which don't have the pixel counter. That's good since without the pixel counter it's rather difficult to verify the timings to this accuracy. So the conclusion is that we can throw away all the ISR tricks I added, and just increment the scanline counter by one always. Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:44 +08:00
hsync_start = mode->crtc_hsync_start;
vtotal = mode->crtc_vtotal;
vbl_start = mode->crtc_vblank_start;
vbl_end = mode->crtc_vblank_end;
if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
vbl_start = DIV_ROUND_UP(vbl_start, 2);
vbl_end /= 2;
vtotal /= 2;
}
ret |= DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_ACCURATE;
drm/intel: Push get_scanout_position() timestamping into kms driver. Move the ktime_get() clock readouts and potential preempt_disable() calls from drm core into kms driver to make it compatible with the api changes in the drm core. The intel-kms driver needs to take the uncore.lock inside i915_get_crtc_scanoutpos() and intel_pipe_in_vblank(). This is incompatible with the preempt_disable() on a PREEMPT_RT patched kernel, as regular spin locks must not be taken within a preempt_disable'd section. Lock contention on the uncore.lock also introduced too much uncertainty in vblank timestamps. Push the ktime_get() timestamping for scanoutpos queries and potential preempt_disable_rt() into i915_get_crtc_scanoutpos(), so these problems can be avoided: 1. First lock the uncore.lock (might sleep on a PREEMPT_RT kernel). 2. preempt_disable_rt() (will be added by the rt-linux folks). 3. ktime_get() a timestamp before scanout pos query. 4. Do all mmio reads as fast as possible without grabbing any new locks! 5. ktime_get() a post-query timestamp. 6. preempt_enable_rt() 7. Unlock the uncore.lock. This reduces timestamp uncertainty on a low-end HP Atom Mini netbook with Intel GMA-950 nicely: Before: 3-8 usecs with spikes > 20 usecs, triggering query retries. After : Typically 1 usec (98% of all samples), occassionally 2 usecs (2% of all samples), with maximum of 3 usecs (a handful). v2: Fix formatting of new multi-line code comments. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-10-30 12:13:08 +08:00
/*
* Lock uncore.lock, as we will do multiple timing critical raw
* register reads, potentially with preemption disabled, so the
* following code must not block on uncore.lock.
*/
spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
drm/i915: Fix scanout position for real Seems I've been a bit dense with regards to the start of vblank vs. the scanline counter / pixel counter. After staring at the pixel counter on gen4 I came to the conclusion that the start of vblank interrupt and scanline counter increment happen at the same time. The scanline counter increment is documented to occur at start of hsync, which means that the start of vblank interrupt must also trigger there. Looking at the pixel counter value when the scanline wraps from vtotal-1 to 0 confirms that, as the pixel counter at that point reads hsync_start. This also clarifies why we see need the +1 adjustment to the scaline counter. The counter actually starts counting from vtotal-1 on the first active line. I also confirmed that the frame start interrupt happens ~1 line after the start of vblank, but the frame start occurs at hblank_start instead. We only use the frame start interrupt on gen2 where the start of vblank interrupt isn't available. The only important thing to note here is that frame start occurs after vblank start, so we don't have to play any additional tricks to fix up the scanline counter. The other thing to note is the fact that the pixel counter on gen3-4 starts counting from the start of horizontal active on the first active line. That means that when we get the start of vblank interrupt, the pixel counter reads (htotal*(vblank_start-1)+hsync_start). Since we consider vblank to start at (htotal*vblank_start) we need to add a constant (htotal-hsync_start) offset to the pixel counter, or else we risk misdetecting whether we're in vblank or not. I talked a bit with Art Runyan about these topics, and he confirmed my findings. And that the same rules should hold for platforms which don't have the pixel counter. That's good since without the pixel counter it's rather difficult to verify the timings to this accuracy. So the conclusion is that we can throw away all the ISR tricks I added, and just increment the scanline counter by one always. Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:44 +08:00
drm/intel: Push get_scanout_position() timestamping into kms driver. Move the ktime_get() clock readouts and potential preempt_disable() calls from drm core into kms driver to make it compatible with the api changes in the drm core. The intel-kms driver needs to take the uncore.lock inside i915_get_crtc_scanoutpos() and intel_pipe_in_vblank(). This is incompatible with the preempt_disable() on a PREEMPT_RT patched kernel, as regular spin locks must not be taken within a preempt_disable'd section. Lock contention on the uncore.lock also introduced too much uncertainty in vblank timestamps. Push the ktime_get() timestamping for scanoutpos queries and potential preempt_disable_rt() into i915_get_crtc_scanoutpos(), so these problems can be avoided: 1. First lock the uncore.lock (might sleep on a PREEMPT_RT kernel). 2. preempt_disable_rt() (will be added by the rt-linux folks). 3. ktime_get() a timestamp before scanout pos query. 4. Do all mmio reads as fast as possible without grabbing any new locks! 5. ktime_get() a post-query timestamp. 6. preempt_enable_rt() 7. Unlock the uncore.lock. This reduces timestamp uncertainty on a low-end HP Atom Mini netbook with Intel GMA-950 nicely: Before: 3-8 usecs with spikes > 20 usecs, triggering query retries. After : Typically 1 usec (98% of all samples), occassionally 2 usecs (2% of all samples), with maximum of 3 usecs (a handful). v2: Fix formatting of new multi-line code comments. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-10-30 12:13:08 +08:00
/* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */
/* Get optional system timestamp before query. */
if (stime)
*stime = ktime_get();
if (IS_GEN2(dev) || IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
/* No obvious pixelcount register. Only query vertical
* scanout position from Display scan line register.
*/
position = __intel_get_crtc_scanline(intel_crtc);
} else {
/* Have access to pixelcount since start of frame.
* We can split this into vertical and horizontal
* scanout position.
*/
position = (I915_READ_FW(PIPEFRAMEPIXEL(pipe)) & PIPE_PIXEL_MASK) >> PIPE_PIXEL_SHIFT;
/* convert to pixel counts */
vbl_start *= htotal;
vbl_end *= htotal;
vtotal *= htotal;
drm/i915: Fix scanout position for real Seems I've been a bit dense with regards to the start of vblank vs. the scanline counter / pixel counter. After staring at the pixel counter on gen4 I came to the conclusion that the start of vblank interrupt and scanline counter increment happen at the same time. The scanline counter increment is documented to occur at start of hsync, which means that the start of vblank interrupt must also trigger there. Looking at the pixel counter value when the scanline wraps from vtotal-1 to 0 confirms that, as the pixel counter at that point reads hsync_start. This also clarifies why we see need the +1 adjustment to the scaline counter. The counter actually starts counting from vtotal-1 on the first active line. I also confirmed that the frame start interrupt happens ~1 line after the start of vblank, but the frame start occurs at hblank_start instead. We only use the frame start interrupt on gen2 where the start of vblank interrupt isn't available. The only important thing to note here is that frame start occurs after vblank start, so we don't have to play any additional tricks to fix up the scanline counter. The other thing to note is the fact that the pixel counter on gen3-4 starts counting from the start of horizontal active on the first active line. That means that when we get the start of vblank interrupt, the pixel counter reads (htotal*(vblank_start-1)+hsync_start). Since we consider vblank to start at (htotal*vblank_start) we need to add a constant (htotal-hsync_start) offset to the pixel counter, or else we risk misdetecting whether we're in vblank or not. I talked a bit with Art Runyan about these topics, and he confirmed my findings. And that the same rules should hold for platforms which don't have the pixel counter. That's good since without the pixel counter it's rather difficult to verify the timings to this accuracy. So the conclusion is that we can throw away all the ISR tricks I added, and just increment the scanline counter by one always. Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:44 +08:00
/*
* In interlaced modes, the pixel counter counts all pixels,
* so one field will have htotal more pixels. In order to avoid
* the reported position from jumping backwards when the pixel
* counter is beyond the length of the shorter field, just
* clamp the position the length of the shorter field. This
* matches how the scanline counter based position works since
* the scanline counter doesn't count the two half lines.
*/
if (position >= vtotal)
position = vtotal - 1;
drm/i915: Fix scanout position for real Seems I've been a bit dense with regards to the start of vblank vs. the scanline counter / pixel counter. After staring at the pixel counter on gen4 I came to the conclusion that the start of vblank interrupt and scanline counter increment happen at the same time. The scanline counter increment is documented to occur at start of hsync, which means that the start of vblank interrupt must also trigger there. Looking at the pixel counter value when the scanline wraps from vtotal-1 to 0 confirms that, as the pixel counter at that point reads hsync_start. This also clarifies why we see need the +1 adjustment to the scaline counter. The counter actually starts counting from vtotal-1 on the first active line. I also confirmed that the frame start interrupt happens ~1 line after the start of vblank, but the frame start occurs at hblank_start instead. We only use the frame start interrupt on gen2 where the start of vblank interrupt isn't available. The only important thing to note here is that frame start occurs after vblank start, so we don't have to play any additional tricks to fix up the scanline counter. The other thing to note is the fact that the pixel counter on gen3-4 starts counting from the start of horizontal active on the first active line. That means that when we get the start of vblank interrupt, the pixel counter reads (htotal*(vblank_start-1)+hsync_start). Since we consider vblank to start at (htotal*vblank_start) we need to add a constant (htotal-hsync_start) offset to the pixel counter, or else we risk misdetecting whether we're in vblank or not. I talked a bit with Art Runyan about these topics, and he confirmed my findings. And that the same rules should hold for platforms which don't have the pixel counter. That's good since without the pixel counter it's rather difficult to verify the timings to this accuracy. So the conclusion is that we can throw away all the ISR tricks I added, and just increment the scanline counter by one always. Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:44 +08:00
/*
* Start of vblank interrupt is triggered at start of hsync,
* just prior to the first active line of vblank. However we
* consider lines to start at the leading edge of horizontal
* active. So, should we get here before we've crossed into
* the horizontal active of the first line in vblank, we would
* not set the DRM_SCANOUTPOS_INVBL flag. In order to fix that,
* always add htotal-hsync_start to the current pixel position.
*/
position = (position + htotal - hsync_start) % vtotal;
}
drm/intel: Push get_scanout_position() timestamping into kms driver. Move the ktime_get() clock readouts and potential preempt_disable() calls from drm core into kms driver to make it compatible with the api changes in the drm core. The intel-kms driver needs to take the uncore.lock inside i915_get_crtc_scanoutpos() and intel_pipe_in_vblank(). This is incompatible with the preempt_disable() on a PREEMPT_RT patched kernel, as regular spin locks must not be taken within a preempt_disable'd section. Lock contention on the uncore.lock also introduced too much uncertainty in vblank timestamps. Push the ktime_get() timestamping for scanoutpos queries and potential preempt_disable_rt() into i915_get_crtc_scanoutpos(), so these problems can be avoided: 1. First lock the uncore.lock (might sleep on a PREEMPT_RT kernel). 2. preempt_disable_rt() (will be added by the rt-linux folks). 3. ktime_get() a timestamp before scanout pos query. 4. Do all mmio reads as fast as possible without grabbing any new locks! 5. ktime_get() a post-query timestamp. 6. preempt_enable_rt() 7. Unlock the uncore.lock. This reduces timestamp uncertainty on a low-end HP Atom Mini netbook with Intel GMA-950 nicely: Before: 3-8 usecs with spikes > 20 usecs, triggering query retries. After : Typically 1 usec (98% of all samples), occassionally 2 usecs (2% of all samples), with maximum of 3 usecs (a handful). v2: Fix formatting of new multi-line code comments. Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2013-10-30 12:13:08 +08:00
/* Get optional system timestamp after query. */
if (etime)
*etime = ktime_get();
/* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */
spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
in_vbl = position >= vbl_start && position < vbl_end;
/*
* While in vblank, position will be negative
* counting up towards 0 at vbl_end. And outside
* vblank, position will be positive counting
* up since vbl_end.
*/
if (position >= vbl_start)
position -= vbl_end;
else
position += vtotal - vbl_end;
if (IS_GEN2(dev) || IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
*vpos = position;
*hpos = 0;
} else {
*vpos = position / htotal;
*hpos = position - (*vpos * htotal);
}
/* In vblank? */
if (in_vbl)
ret |= DRM_SCANOUTPOS_IN_VBLANK;
return ret;
}
int intel_get_crtc_scanline(struct intel_crtc *crtc)
{
struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
unsigned long irqflags;
int position;
spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
position = __intel_get_crtc_scanline(crtc);
spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags);
return position;
}
static int i915_get_vblank_timestamp(struct drm_device *dev, unsigned int pipe,
int *max_error,
struct timeval *vblank_time,
unsigned flags)
{
struct drm_crtc *crtc;
if (pipe >= INTEL_INFO(dev)->num_pipes) {
DRM_ERROR("Invalid crtc %u\n", pipe);
return -EINVAL;
}
/* Get drm_crtc to timestamp: */
crtc = intel_get_crtc_for_pipe(dev, pipe);
if (crtc == NULL) {
DRM_ERROR("Invalid crtc %u\n", pipe);
return -EINVAL;
}
if (!crtc->hwmode.crtc_clock) {
DRM_DEBUG_KMS("crtc %u is disabled\n", pipe);
return -EBUSY;
}
/* Helper routine in DRM core does all the work: */
return drm_calc_vbltimestamp_from_scanoutpos(dev, pipe, max_error,
vblank_time, flags,
&crtc->hwmode);
}
static void ironlake_rps_change_irq_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 busy_up, busy_down, max_avg, min_avg;
u8 new_delay;
spin_lock(&mchdev_lock);
I915_WRITE16(MEMINTRSTS, I915_READ(MEMINTRSTS));
new_delay = dev_priv->ips.cur_delay;
I915_WRITE16(MEMINTRSTS, MEMINT_EVAL_CHG);
busy_up = I915_READ(RCPREVBSYTUPAVG);
busy_down = I915_READ(RCPREVBSYTDNAVG);
max_avg = I915_READ(RCBMAXAVG);
min_avg = I915_READ(RCBMINAVG);
/* Handle RCS change request from hw */
if (busy_up > max_avg) {
if (dev_priv->ips.cur_delay != dev_priv->ips.max_delay)
new_delay = dev_priv->ips.cur_delay - 1;
if (new_delay < dev_priv->ips.max_delay)
new_delay = dev_priv->ips.max_delay;
} else if (busy_down < min_avg) {
if (dev_priv->ips.cur_delay != dev_priv->ips.min_delay)
new_delay = dev_priv->ips.cur_delay + 1;
if (new_delay > dev_priv->ips.min_delay)
new_delay = dev_priv->ips.min_delay;
}
if (ironlake_set_drps(dev, new_delay))
dev_priv->ips.cur_delay = new_delay;
spin_unlock(&mchdev_lock);
return;
}
static void notify_ring(struct intel_engine_cs *engine)
{
if (!intel_engine_initialized(engine))
return;
trace_i915_gem_request_notify(engine);
engine->user_interrupts++;
wake_up_all(&engine->irq_queue);
}
static void vlv_c0_read(struct drm_i915_private *dev_priv,
struct intel_rps_ei *ei)
{
ei->cz_clock = vlv_punit_read(dev_priv, PUNIT_REG_CZ_TIMESTAMP);
ei->render_c0 = I915_READ(VLV_RENDER_C0_COUNT);
ei->media_c0 = I915_READ(VLV_MEDIA_C0_COUNT);
}
static bool vlv_c0_above(struct drm_i915_private *dev_priv,
const struct intel_rps_ei *old,
const struct intel_rps_ei *now,
int threshold)
{
u64 time, c0;
unsigned int mul = 100;
if (old->cz_clock == 0)
return false;
if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH)
mul <<= 8;
time = now->cz_clock - old->cz_clock;
time *= threshold * dev_priv->czclk_freq;
/* Workload can be split between render + media, e.g. SwapBuffers
* being blitted in X after being rendered in mesa. To account for
* this we need to combine both engines into our activity counter.
*/
c0 = now->render_c0 - old->render_c0;
c0 += now->media_c0 - old->media_c0;
c0 *= mul * VLV_CZ_CLOCK_TO_MILLI_SEC;
return c0 >= time;
}
void gen6_rps_reset_ei(struct drm_i915_private *dev_priv)
{
vlv_c0_read(dev_priv, &dev_priv->rps.down_ei);
dev_priv->rps.up_ei = dev_priv->rps.down_ei;
}
static u32 vlv_wa_c0_ei(struct drm_i915_private *dev_priv, u32 pm_iir)
{
struct intel_rps_ei now;
u32 events = 0;
if ((pm_iir & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED)) == 0)
return 0;
vlv_c0_read(dev_priv, &now);
if (now.cz_clock == 0)
return 0;
if (pm_iir & GEN6_PM_RP_DOWN_EI_EXPIRED) {
if (!vlv_c0_above(dev_priv,
&dev_priv->rps.down_ei, &now,
dev_priv->rps.down_threshold))
events |= GEN6_PM_RP_DOWN_THRESHOLD;
dev_priv->rps.down_ei = now;
}
if (pm_iir & GEN6_PM_RP_UP_EI_EXPIRED) {
if (vlv_c0_above(dev_priv,
&dev_priv->rps.up_ei, &now,
dev_priv->rps.up_threshold))
events |= GEN6_PM_RP_UP_THRESHOLD;
dev_priv->rps.up_ei = now;
}
return events;
}
static bool any_waiters(struct drm_i915_private *dev_priv)
{
struct intel_engine_cs *engine;
for_each_engine(engine, dev_priv)
if (engine->irq_refcount)
return true;
return false;
}
static void gen6_pm_rps_work(struct work_struct *work)
{
struct drm_i915_private *dev_priv =
container_of(work, struct drm_i915_private, rps.work);
bool client_boost;
int new_delay, adj, min, max;
u32 pm_iir;
spin_lock_irq(&dev_priv->irq_lock);
drm/i915: sanitize rps irq disabling When disabling the RPS interrupts there is a tricky dependency between the thread disabling the interrupts, the RPS interrupt handler and the corresponding RPS work. The RPS work can reenable the interrupts, so there is no straightforward order in the disabling thread to (1) make sure that any RPS work is flushed and to (2) disable all RPS interrupts. Currently this is solved by masking the interrupts using two separate mask registers (first level display IMR and PM IMR) and doing the disabling when all first level interrupts are disabled. This works, but the requirement to run with all first level interrupts disabled is unnecessary making the suspend / unload time ordering of RPS disabling wrt. other unitialization steps difficult and error prone. Removing this restriction allows us to disable RPS early during suspend / unload and forget about it for the rest of the sequence. By adding a more explicit method for avoiding the above race, it also becomes easier to prove its correctness. Finally currently we can hit the WARN in snb_update_pm_irq(), when a final RPS work runs with the first level interrupts already disabled. This won't lead to any problem (due to the separate interrupt masks), but with the change in this and the next patch we can get rid of the WARN, while leaving it in place for other scenarios. To address the above points, add a new RPS interrupts_enabled flag and use this during RPS disabling to avoid requeuing the RPS work and reenabling of the RPS interrupts. Since the interrupt disabling happens now in intel_suspend_gt_powersave(), we will disable RPS interrupts explicitly during suspend (and not just through the first level mask), but there is no problem doing so, it's also more consistent and allows us to unify more of the RPS disabling during suspend and unload time in the next patch. v2/v3: - rebase on patch "drm/i915: move rps irq disable one level up" in the patchset Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-19 21:30:04 +08:00
/* Speed up work cancelation during disabling rps interrupts. */
if (!dev_priv->rps.interrupts_enabled) {
spin_unlock_irq(&dev_priv->irq_lock);
return;
}
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/*
* The RPS work is synced during runtime suspend, we don't require a
* wakeref. TODO: instead of disabling the asserts make sure that we
* always hold an RPM reference while the work is running.
*/
DISABLE_RPM_WAKEREF_ASSERTS(dev_priv);
pm_iir = dev_priv->rps.pm_iir;
dev_priv->rps.pm_iir = 0;
/* Make sure not to corrupt PMIMR state used by ringbuffer on GEN6 */
gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events);
client_boost = dev_priv->rps.client_boost;
dev_priv->rps.client_boost = false;
spin_unlock_irq(&dev_priv->irq_lock);
/* Make sure we didn't queue anything we're not going to process. */
WARN_ON(pm_iir & ~dev_priv->pm_rps_events);
if ((pm_iir & dev_priv->pm_rps_events) == 0 && !client_boost)
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
goto out;
mutex_lock(&dev_priv->rps.hw_lock);
pm_iir |= vlv_wa_c0_ei(dev_priv, pm_iir);
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
adj = dev_priv->rps.last_adj;
new_delay = dev_priv->rps.cur_freq;
min = dev_priv->rps.min_freq_softlimit;
max = dev_priv->rps.max_freq_softlimit;
if (client_boost) {
new_delay = dev_priv->rps.max_freq_softlimit;
adj = 0;
} else if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) {
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
if (adj > 0)
adj *= 2;
else /* CHV needs even encode values */
adj = IS_CHERRYVIEW(dev_priv) ? 2 : 1;
/*
* For better performance, jump directly
* to RPe if we're below it.
*/
if (new_delay < dev_priv->rps.efficient_freq - adj) {
new_delay = dev_priv->rps.efficient_freq;
adj = 0;
}
} else if (any_waiters(dev_priv)) {
adj = 0;
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
} else if (pm_iir & GEN6_PM_RP_DOWN_TIMEOUT) {
if (dev_priv->rps.cur_freq > dev_priv->rps.efficient_freq)
new_delay = dev_priv->rps.efficient_freq;
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
else
new_delay = dev_priv->rps.min_freq_softlimit;
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
adj = 0;
} else if (pm_iir & GEN6_PM_RP_DOWN_THRESHOLD) {
if (adj < 0)
adj *= 2;
else /* CHV needs even encode values */
adj = IS_CHERRYVIEW(dev_priv) ? -2 : -1;
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
} else { /* unknown event */
adj = 0;
drm/i915: Tweak RPS thresholds to more aggressively downclock After applying wait-boost we often find ourselves stuck at higher clocks than required. The current threshold value requires the GPU to be continuously and completely idle for 313ms before it is dropped by one bin. Conversely, we require the GPU to be busy for an average of 90% over a 84ms period before we upclock. So the current thresholds almost never downclock the GPU, and respond very slowly to sudden demands for more power. It is easy to observe that we currently lock into the wrong bin and both underperform in benchmarks and consume more power than optimal (just by repeating the task and measuring the different results). An alternative approach, as discussed in the bspec, is to use a continuous threshold for upclocking, and an average value for downclocking. This is good for quickly detecting and reacting to state changes within a frame, however it fails with the common throttling method of waiting upon the outstanding frame - at least it is difficult to choose a threshold that works well at 15,000fps and at 60fps. So continue to use average busy/idle loads to determine frequency change. v2: Use 3 power zones to keep frequencies low in steady-state mostly idle (e.g. scrolling, interactive 2D drawing), and frequencies high for demanding games. In between those end-states, we use a fast-reclocking algorithm to converge more quickly on the desired bin. v3: Bug fixes - make sure we reset adj after switching power zones. v4: Tune - drop the continuous busy thresholds as it prevents us from choosing the right frequency for glxgears style swap benchmarks. Instead the goal is to be able to find the right clocks irrespective of the wait-boost. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Stéphane Marchesin <stephane.marchesin@gmail.com> Cc: Owen Taylor <otaylor@redhat.com> Cc: "Meng, Mengmeng" <mengmeng.meng@intel.com> Cc: "Zhuang, Lena" <lena.zhuang@intel.com> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-26 00:34:57 +08:00
}
dev_priv->rps.last_adj = adj;
/* sysfs frequency interfaces may have snuck in while servicing the
* interrupt
*/
new_delay += adj;
new_delay = clamp_t(int, new_delay, min, max);
intel_set_rps(dev_priv->dev, new_delay);
mutex_unlock(&dev_priv->rps.hw_lock);
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
out:
ENABLE_RPM_WAKEREF_ASSERTS(dev_priv);
}
/**
* ivybridge_parity_work - Workqueue called when a parity error interrupt
* occurred.
* @work: workqueue struct
*
* Doesn't actually do anything except notify userspace. As a consequence of
* this event, userspace should try to remap the bad rows since statistically
* it is likely the same row is more likely to go bad again.
*/
static void ivybridge_parity_work(struct work_struct *work)
{
struct drm_i915_private *dev_priv =
container_of(work, struct drm_i915_private, l3_parity.error_work);
u32 error_status, row, bank, subbank;
char *parity_event[6];
uint32_t misccpctl;
uint8_t slice = 0;
/* We must turn off DOP level clock gating to access the L3 registers.
* In order to prevent a get/put style interface, acquire struct mutex
* any time we access those registers.
*/
mutex_lock(&dev_priv->dev->struct_mutex);
/* If we've screwed up tracking, just let the interrupt fire again */
if (WARN_ON(!dev_priv->l3_parity.which_slice))
goto out;
misccpctl = I915_READ(GEN7_MISCCPCTL);
I915_WRITE(GEN7_MISCCPCTL, misccpctl & ~GEN7_DOP_CLOCK_GATE_ENABLE);
POSTING_READ(GEN7_MISCCPCTL);
while ((slice = ffs(dev_priv->l3_parity.which_slice)) != 0) {
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg;
slice--;
if (WARN_ON_ONCE(slice >= NUM_L3_SLICES(dev_priv)))
break;
dev_priv->l3_parity.which_slice &= ~(1<<slice);
reg = GEN7_L3CDERRST1(slice);
error_status = I915_READ(reg);
row = GEN7_PARITY_ERROR_ROW(error_status);
bank = GEN7_PARITY_ERROR_BANK(error_status);
subbank = GEN7_PARITY_ERROR_SUBBANK(error_status);
I915_WRITE(reg, GEN7_PARITY_ERROR_VALID | GEN7_L3CDERRST1_ENABLE);
POSTING_READ(reg);
parity_event[0] = I915_L3_PARITY_UEVENT "=1";
parity_event[1] = kasprintf(GFP_KERNEL, "ROW=%d", row);
parity_event[2] = kasprintf(GFP_KERNEL, "BANK=%d", bank);
parity_event[3] = kasprintf(GFP_KERNEL, "SUBBANK=%d", subbank);
parity_event[4] = kasprintf(GFP_KERNEL, "SLICE=%d", slice);
parity_event[5] = NULL;
kobject_uevent_env(&dev_priv->dev->primary->kdev->kobj,
KOBJ_CHANGE, parity_event);
DRM_DEBUG("Parity error: Slice = %d, Row = %d, Bank = %d, Sub bank = %d.\n",
slice, row, bank, subbank);
kfree(parity_event[4]);
kfree(parity_event[3]);
kfree(parity_event[2]);
kfree(parity_event[1]);
}
I915_WRITE(GEN7_MISCCPCTL, misccpctl);
out:
WARN_ON(dev_priv->l3_parity.which_slice);
spin_lock_irq(&dev_priv->irq_lock);
gen5_enable_gt_irq(dev_priv, GT_PARITY_ERROR(dev_priv));
spin_unlock_irq(&dev_priv->irq_lock);
mutex_unlock(&dev_priv->dev->struct_mutex);
}
static void ivybridge_parity_error_irq_handler(struct drm_device *dev, u32 iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!HAS_L3_DPF(dev))
return;
spin_lock(&dev_priv->irq_lock);
gen5_disable_gt_irq(dev_priv, GT_PARITY_ERROR(dev));
spin_unlock(&dev_priv->irq_lock);
iir &= GT_PARITY_ERROR(dev);
if (iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT_S1)
dev_priv->l3_parity.which_slice |= 1 << 1;
if (iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT)
dev_priv->l3_parity.which_slice |= 1 << 0;
queue_work(dev_priv->wq, &dev_priv->l3_parity.error_work);
}
static void ilk_gt_irq_handler(struct drm_device *dev,
struct drm_i915_private *dev_priv,
u32 gt_iir)
{
if (gt_iir &
(GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
notify_ring(&dev_priv->engine[RCS]);
if (gt_iir & ILK_BSD_USER_INTERRUPT)
notify_ring(&dev_priv->engine[VCS]);
}
static void snb_gt_irq_handler(struct drm_device *dev,
struct drm_i915_private *dev_priv,
u32 gt_iir)
{
if (gt_iir &
(GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
notify_ring(&dev_priv->engine[RCS]);
if (gt_iir & GT_BSD_USER_INTERRUPT)
notify_ring(&dev_priv->engine[VCS]);
if (gt_iir & GT_BLT_USER_INTERRUPT)
notify_ring(&dev_priv->engine[BCS]);
if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT |
GT_BSD_CS_ERROR_INTERRUPT |
GT_RENDER_CS_MASTER_ERROR_INTERRUPT))
DRM_DEBUG("Command parser error, gt_iir 0x%08x\n", gt_iir);
if (gt_iir & GT_PARITY_ERROR(dev))
ivybridge_parity_error_irq_handler(dev, gt_iir);
}
static __always_inline void
gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift)
{
if (iir & (GT_RENDER_USER_INTERRUPT << test_shift))
notify_ring(engine);
if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift))
drm/i915: Move execlists irq handler to a bottom half Doing a lot of work in the interrupt handler introduces huge latencies to the system as a whole. Most dramatic effect can be seen by running an all engine stress test like igt/gem_exec_nop/all where, when the kernel config is lean enough, the whole system can be brought into multi-second periods of complete non-interactivty. That can look for example like this: NMI watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [kworker/u8:3:143] Modules linked in: [redacted for brevity] CPU: 0 PID: 143 Comm: kworker/u8:3 Tainted: G U L 4.5.0-160321+ #183 Hardware name: Intel Corporation Broadwell Client platform/WhiteTip Mountain 1 Workqueue: i915 gen6_pm_rps_work [i915] task: ffff8800aae88000 ti: ffff8800aae90000 task.ti: ffff8800aae90000 RIP: 0010:[<ffffffff8104a3c2>] [<ffffffff8104a3c2>] __do_softirq+0x72/0x1d0 RSP: 0000:ffff88014f403f38 EFLAGS: 00000206 RAX: ffff8800aae94000 RBX: 0000000000000000 RCX: 00000000000006e0 RDX: 0000000000000020 RSI: 0000000004208060 RDI: 0000000000215d80 RBP: ffff88014f403f80 R08: 0000000b1b42c180 R09: 0000000000000022 R10: 0000000000000004 R11: 00000000ffffffff R12: 000000000000a030 R13: 0000000000000082 R14: ffff8800aa4d0080 R15: 0000000000000082 FS: 0000000000000000(0000) GS:ffff88014f400000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fa53b90c000 CR3: 0000000001a0a000 CR4: 00000000001406f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Stack: 042080601b33869f ffff8800aae94000 00000000fffc2678 ffff88010000000a 0000000000000000 000000000000a030 0000000000005302 ffff8800aa4d0080 0000000000000206 ffff88014f403f90 ffffffff8104a716 ffff88014f403fa8 Call Trace: <IRQ> [<ffffffff8104a716>] irq_exit+0x86/0x90 [<ffffffff81031e7d>] smp_apic_timer_interrupt+0x3d/0x50 [<ffffffff814f3eac>] apic_timer_interrupt+0x7c/0x90 <EOI> [<ffffffffa01c5b40>] ? gen8_write64+0x1a0/0x1a0 [i915] [<ffffffff814f2b39>] ? _raw_spin_unlock_irqrestore+0x9/0x20 [<ffffffffa01c5c44>] gen8_write32+0x104/0x1a0 [i915] [<ffffffff8132c6a2>] ? n_tty_receive_buf_common+0x372/0xae0 [<ffffffffa017cc9e>] gen6_set_rps_thresholds+0x1be/0x330 [i915] [<ffffffffa017eaf0>] gen6_set_rps+0x70/0x200 [i915] [<ffffffffa0185375>] intel_set_rps+0x25/0x30 [i915] [<ffffffffa01768fd>] gen6_pm_rps_work+0x10d/0x2e0 [i915] [<ffffffff81063852>] ? finish_task_switch+0x72/0x1c0 [<ffffffff8105ab29>] process_one_work+0x139/0x350 [<ffffffff8105b186>] worker_thread+0x126/0x490 [<ffffffff8105b060>] ? rescuer_thread+0x320/0x320 [<ffffffff8105fa64>] kthread+0xc4/0xe0 [<ffffffff8105f9a0>] ? kthread_create_on_node+0x170/0x170 [<ffffffff814f351f>] ret_from_fork+0x3f/0x70 [<ffffffff8105f9a0>] ? kthread_create_on_node+0x170/0x170 I could not explain, or find a code path, which would explain a +20 second lockup, but from some instrumentation it was apparent the interrupts off proportion of time was between 10-25% under heavy load which is quite bad. When a interrupt "cliff" is reached, which was >~320k irq/s on my machine, the whole system goes into a terrible state of the above described multi-second lockups. By moving the GT interrupt handling to a tasklet in a most simple way, the problem above disappears completely. Testing the effect on sytem-wide latencies using igt/gem_syslatency shows the following before this patch: gem_syslatency: cycles=1532739, latency mean=416531.829us max=2499237us gem_syslatency: cycles=1839434, latency mean=1458099.157us max=4998944us gem_syslatency: cycles=1432570, latency mean=2688.451us max=1201185us gem_syslatency: cycles=1533543, latency mean=416520.499us max=2498886us This shows that the unrelated process is experiencing huge delays in its wake-up latency. After the patch the results look like this: gem_syslatency: cycles=808907, latency mean=53.133us max=1640us gem_syslatency: cycles=862154, latency mean=62.778us max=2117us gem_syslatency: cycles=856039, latency mean=58.079us max=2123us gem_syslatency: cycles=841683, latency mean=56.914us max=1667us Showing a huge improvement in the unrelated process wake-up latency. It also shows an approximate halving in the number of total empty batches submitted during the test. This may not be worrying since the test puts the driver under a very unrealistic load with ncpu threads doing empty batch submission to all GPU engines each. Another benefit compared to the hard-irq handling is that now work on all engines can be dispatched in parallel since we can have up to number of CPUs active tasklets. (While previously a single hard-irq would serially dispatch on one engine after another.) More interesting scenario with regards to throughput is "gem_latency -n 100" which shows 25% better throughput and CPU usage, and 14% better dispatch latencies. I did not find any gains or regressions with Synmark2 or GLbench under light testing. More benchmarking is certainly required. v2: * execlists_lock should be taken as spin_lock_bh when queuing work from userspace now. (Chris Wilson) * uncore.lock must be taken with spin_lock_irq when submitting requests since that now runs from either softirq or process context. v3: * Expanded commit message with more testing data; * converted missed locking sites to _bh; * added execlist_lock comment. (Chris Wilson) v4: * Mention dispatch parallelism in commit. (Chris Wilson) * Do not hold uncore.lock over MMIO reads since the block is already serialised per-engine via the tasklet itself. (Chris Wilson) * intel_lrc_irq_handler should be static. (Chris Wilson) * Cancel/sync the tasklet on GPU reset. (Chris Wilson) * Document and WARN that tasklet cannot be active/pending on engine cleanup. (Chris Wilson/Imre Deak) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Imre Deak <imre.deak@intel.com> Testcase: igt/gem_exec_nop/all Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94350 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1459768316-6670-1-git-send-email-tvrtko.ursulin@linux.intel.com
2016-04-04 19:11:56 +08:00
tasklet_schedule(&engine->irq_tasklet);
}
static irqreturn_t gen8_gt_irq_handler(struct drm_i915_private *dev_priv,
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
u32 master_ctl)
{
irqreturn_t ret = IRQ_NONE;
if (master_ctl & (GEN8_GT_RCS_IRQ | GEN8_GT_BCS_IRQ)) {
u32 iir = I915_READ_FW(GEN8_GT_IIR(0));
if (iir) {
I915_WRITE_FW(GEN8_GT_IIR(0), iir);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
ret = IRQ_HANDLED;
2014-07-25 00:04:39 +08:00
gen8_cs_irq_handler(&dev_priv->engine[RCS],
iir, GEN8_RCS_IRQ_SHIFT);
gen8_cs_irq_handler(&dev_priv->engine[BCS],
iir, GEN8_BCS_IRQ_SHIFT);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
} else
DRM_ERROR("The master control interrupt lied (GT0)!\n");
}
if (master_ctl & (GEN8_GT_VCS1_IRQ | GEN8_GT_VCS2_IRQ)) {
u32 iir = I915_READ_FW(GEN8_GT_IIR(1));
if (iir) {
I915_WRITE_FW(GEN8_GT_IIR(1), iir);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
ret = IRQ_HANDLED;
2014-07-25 00:04:39 +08:00
gen8_cs_irq_handler(&dev_priv->engine[VCS],
iir, GEN8_VCS1_IRQ_SHIFT);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
gen8_cs_irq_handler(&dev_priv->engine[VCS2],
iir, GEN8_VCS2_IRQ_SHIFT);
drm/i915/bdw: Implement a basic PM interrupt handler Almost all of it is reusable from the existing code. The primary difference is we need to do even less in the interrupt handler, since interrupts are not shared in the same way. The patch is mostly a copy-paste of the existing snb+ code, with updates to the relevant parts requiring changes to the interrupt handling. As such it /should/ be relatively trivial. It's highly likely that I missed some places where I need a gen8 version of the PM interrupts, but it has become invisible to me by now. This patch could probably be split into adding the new functions, followed by actually handling the interrupts. Since the code is currently disabled (and broken) I think the patch stands better by itself. v2: Move the commit about not touching the ringbuffer interrupt to the snb_* function where it belongs (Rodrigo) v3: Rebased on Paulo's runtime PM changes v4: Not well validated, but rebase on commit 730488b2eddded4497f63f70867b1256cd9e117c Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Mar 7 20:12:32 2014 -0300 drm/i915: kill dev_priv->pm.regsave v5: Rebased on latest code base. (Deepak) v6: Remove conflict markers, Unnecessary empty line and use right IIR interrupt (Ville) v7: mask modified without rmw (Ville Syrjälä) Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-05-16 01:58:08 +08:00
} else
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
DRM_ERROR("The master control interrupt lied (GT1)!\n");
drm/i915/bdw: Implement a basic PM interrupt handler Almost all of it is reusable from the existing code. The primary difference is we need to do even less in the interrupt handler, since interrupts are not shared in the same way. The patch is mostly a copy-paste of the existing snb+ code, with updates to the relevant parts requiring changes to the interrupt handling. As such it /should/ be relatively trivial. It's highly likely that I missed some places where I need a gen8 version of the PM interrupts, but it has become invisible to me by now. This patch could probably be split into adding the new functions, followed by actually handling the interrupts. Since the code is currently disabled (and broken) I think the patch stands better by itself. v2: Move the commit about not touching the ringbuffer interrupt to the snb_* function where it belongs (Rodrigo) v3: Rebased on Paulo's runtime PM changes v4: Not well validated, but rebase on commit 730488b2eddded4497f63f70867b1256cd9e117c Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Mar 7 20:12:32 2014 -0300 drm/i915: kill dev_priv->pm.regsave v5: Rebased on latest code base. (Deepak) v6: Remove conflict markers, Unnecessary empty line and use right IIR interrupt (Ville) v7: mask modified without rmw (Ville Syrjälä) Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-05-16 01:58:08 +08:00
}
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
if (master_ctl & GEN8_GT_VECS_IRQ) {
u32 iir = I915_READ_FW(GEN8_GT_IIR(3));
if (iir) {
I915_WRITE_FW(GEN8_GT_IIR(3), iir);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
ret = IRQ_HANDLED;
2014-07-25 00:04:39 +08:00
gen8_cs_irq_handler(&dev_priv->engine[VECS],
iir, GEN8_VECS_IRQ_SHIFT);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
} else
DRM_ERROR("The master control interrupt lied (GT3)!\n");
}
drm/i915/bdw: Implement a basic PM interrupt handler Almost all of it is reusable from the existing code. The primary difference is we need to do even less in the interrupt handler, since interrupts are not shared in the same way. The patch is mostly a copy-paste of the existing snb+ code, with updates to the relevant parts requiring changes to the interrupt handling. As such it /should/ be relatively trivial. It's highly likely that I missed some places where I need a gen8 version of the PM interrupts, but it has become invisible to me by now. This patch could probably be split into adding the new functions, followed by actually handling the interrupts. Since the code is currently disabled (and broken) I think the patch stands better by itself. v2: Move the commit about not touching the ringbuffer interrupt to the snb_* function where it belongs (Rodrigo) v3: Rebased on Paulo's runtime PM changes v4: Not well validated, but rebase on commit 730488b2eddded4497f63f70867b1256cd9e117c Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Mar 7 20:12:32 2014 -0300 drm/i915: kill dev_priv->pm.regsave v5: Rebased on latest code base. (Deepak) v6: Remove conflict markers, Unnecessary empty line and use right IIR interrupt (Ville) v7: mask modified without rmw (Ville Syrjälä) Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-05-16 01:58:08 +08:00
if (master_ctl & GEN8_GT_PM_IRQ) {
u32 iir = I915_READ_FW(GEN8_GT_IIR(2));
if (iir & dev_priv->pm_rps_events) {
I915_WRITE_FW(GEN8_GT_IIR(2),
iir & dev_priv->pm_rps_events);
ret = IRQ_HANDLED;
gen6_rps_irq_handler(dev_priv, iir);
drm/i915/bdw: Implement a basic PM interrupt handler Almost all of it is reusable from the existing code. The primary difference is we need to do even less in the interrupt handler, since interrupts are not shared in the same way. The patch is mostly a copy-paste of the existing snb+ code, with updates to the relevant parts requiring changes to the interrupt handling. As such it /should/ be relatively trivial. It's highly likely that I missed some places where I need a gen8 version of the PM interrupts, but it has become invisible to me by now. This patch could probably be split into adding the new functions, followed by actually handling the interrupts. Since the code is currently disabled (and broken) I think the patch stands better by itself. v2: Move the commit about not touching the ringbuffer interrupt to the snb_* function where it belongs (Rodrigo) v3: Rebased on Paulo's runtime PM changes v4: Not well validated, but rebase on commit 730488b2eddded4497f63f70867b1256cd9e117c Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Mar 7 20:12:32 2014 -0300 drm/i915: kill dev_priv->pm.regsave v5: Rebased on latest code base. (Deepak) v6: Remove conflict markers, Unnecessary empty line and use right IIR interrupt (Ville) v7: mask modified without rmw (Ville Syrjälä) Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-05-16 01:58:08 +08:00
} else
DRM_ERROR("The master control interrupt lied (PM)!\n");
}
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
return ret;
}
static bool bxt_port_hotplug_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_A:
return val & PORTA_HOTPLUG_LONG_DETECT;
case PORT_B:
return val & PORTB_HOTPLUG_LONG_DETECT;
case PORT_C:
return val & PORTC_HOTPLUG_LONG_DETECT;
default:
return false;
}
}
static bool spt_port_hotplug2_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_E:
return val & PORTE_HOTPLUG_LONG_DETECT;
default:
return false;
}
}
static bool spt_port_hotplug_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_A:
return val & PORTA_HOTPLUG_LONG_DETECT;
case PORT_B:
return val & PORTB_HOTPLUG_LONG_DETECT;
case PORT_C:
return val & PORTC_HOTPLUG_LONG_DETECT;
case PORT_D:
return val & PORTD_HOTPLUG_LONG_DETECT;
default:
return false;
}
}
static bool ilk_port_hotplug_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_A:
return val & DIGITAL_PORTA_HOTPLUG_LONG_DETECT;
default:
return false;
}
}
static bool pch_port_hotplug_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_B:
return val & PORTB_HOTPLUG_LONG_DETECT;
case PORT_C:
return val & PORTC_HOTPLUG_LONG_DETECT;
case PORT_D:
return val & PORTD_HOTPLUG_LONG_DETECT;
default:
return false;
}
}
static bool i9xx_port_hotplug_long_detect(enum port port, u32 val)
{
switch (port) {
case PORT_B:
return val & PORTB_HOTPLUG_INT_LONG_PULSE;
case PORT_C:
return val & PORTC_HOTPLUG_INT_LONG_PULSE;
case PORT_D:
return val & PORTD_HOTPLUG_INT_LONG_PULSE;
default:
return false;
}
}
/*
* Get a bit mask of pins that have triggered, and which ones may be long.
* This can be called multiple times with the same masks to accumulate
* hotplug detection results from several registers.
*
* Note that the caller is expected to zero out the masks initially.
*/
static void intel_get_hpd_pins(u32 *pin_mask, u32 *long_mask,
u32 hotplug_trigger, u32 dig_hotplug_reg,
const u32 hpd[HPD_NUM_PINS],
bool long_pulse_detect(enum port port, u32 val))
{
enum port port;
int i;
for_each_hpd_pin(i) {
if ((hpd[i] & hotplug_trigger) == 0)
continue;
*pin_mask |= BIT(i);
if (!intel_hpd_pin_to_port(i, &port))
continue;
if (long_pulse_detect(port, dig_hotplug_reg))
*long_mask |= BIT(i);
}
DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x, dig 0x%08x, pins 0x%08x\n",
hotplug_trigger, dig_hotplug_reg, *pin_mask);
}
static void gmbus_irq_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: use the gmbus irq for waits We need two special things to properly wire this up: - Add another argument to gmbus_wait_hw_status to pass in the correct interrupt bit in gmbus4. - Since we can only get an irq for one of the two events we want, hand-roll the wait_event_timeout code so that we wake up every jiffie and can check for NAKs. This way we also subsume gmbus support for platforms without interrupts (or where those are not yet enabled). The important bit really is to only enable one gmbus interrupt source at the same time - with that piece of lore figured out, this seems to work flawlessly. Ben Widawsky rightfully complained the lack of measurements for the claimed benefits (especially since the first version was actually broken and fell back to bit-banging). Previously reading the 256 byte hdmi EDID takes about 72 ms here. With this patch it's down to 33 ms. Given that transfering the 256 bytes over i2c at wire speed takes 20.5ms alone, the reduction in additional overhead is rather nice. v2: Chris Wilson wondered whether GMBUS4 might contain some set bits when booting up an hence result in some spurious interrupts. Since we clear GMBUS4 after every wait and we do gmbus transfer really early in the setup sequence to detect displays the window is small, but still be paranoid and clear it properly. v3: Clarify the comment that gmbus irq generation can only support one kind of event, why it bothers us and how we work around that limit. Cc: Daniel Kurtz <djkurtz@chromium.org> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-01 20:53:45 +08:00
wake_up_all(&dev_priv->gmbus_wait_queue);
}
static void dp_aux_irq_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: irq-drive the dp aux communication At least on the platforms that have a dp aux irq and also have it enabled - vlvhsw should have one, too. But I don't have a machine to test this on. Judging from docs there's no dp aux interrupt for gm45. Also, I only have an ivb cpu edp machine, so the dp aux A code for snb/ilk is untested. For dpcd probing when nothing is connected it slashes about 5ms of cpu time (cpu time is now negligible), which agrees with 3 * 5 400 usec timeouts. A previous version of this patch increases the time required to go through the dp_detect cycle (which includes reading the edid) from around 33 ms to around 40 ms. Experiments indicated that this is purely due to the irq latency - the hw doesn't allow us to queue up dp aux transactions and hence irq latency directly affects throughput. gmbus is much better, there we have a 8 byte buffer, and we get the irq once another 4 bytes can be queued up. But by using the pm_qos interface to request the lowest possible cpu wake-up latency this slowdown completely disappeared. Since all our output detection logic is single-threaded with the mode_config mutex right now anyway, I've decide not ot play fancy and to just reuse the gmbus wait queue. But this would definitely prep the way to run dp detection on different ports in parallel v2: Add a timeout for dp aux transfers when using interrupts - the hw _does_ prevent this with the hw-based 400 usec timeout, but if the irq somehow doesn't arrive we're screwed. Lesson learned while developing this ;-) v3: While at it also convert the busy-loop to wait_for_atomic, so that we don't run the risk of an infinite loop any more. v4: Ensure we have the smallest possible irq latency by using the pm_qos interface. v5: Add a comment to the code to explain why we frob pm_qos. Suggested by Chris Wilson. v6: Disable dp irq for vlv, that's easier than trying to get at docs and hw. v7: Squash in a fix for Haswell that Paulo Zanoni tracked down - the dp aux registers aren't at a fixed offset any more, but can be on the PCH while the DP port is on the cpu die. Reviewed-by: Imre Deak <imre.deak@intel.com> (v6) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-01 20:53:48 +08:00
wake_up_all(&dev_priv->gmbus_wait_queue);
}
#if defined(CONFIG_DEBUG_FS)
static void display_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe,
uint32_t crc0, uint32_t crc1,
uint32_t crc2, uint32_t crc3,
uint32_t crc4)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe];
struct intel_pipe_crc_entry *entry;
int head, tail;
spin_lock(&pipe_crc->lock);
if (!pipe_crc->entries) {
spin_unlock(&pipe_crc->lock);
DRM_DEBUG_KMS("spurious interrupt\n");
return;
}
head = pipe_crc->head;
tail = pipe_crc->tail;
if (CIRC_SPACE(head, tail, INTEL_PIPE_CRC_ENTRIES_NR) < 1) {
spin_unlock(&pipe_crc->lock);
DRM_ERROR("CRC buffer overflowing\n");
return;
}
entry = &pipe_crc->entries[head];
entry->frame = dev->driver->get_vblank_counter(dev, pipe);
entry->crc[0] = crc0;
entry->crc[1] = crc1;
entry->crc[2] = crc2;
entry->crc[3] = crc3;
entry->crc[4] = crc4;
head = (head + 1) & (INTEL_PIPE_CRC_ENTRIES_NR - 1);
pipe_crc->head = head;
spin_unlock(&pipe_crc->lock);
wake_up_interruptible(&pipe_crc->wq);
}
#else
static inline void
display_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe,
uint32_t crc0, uint32_t crc1,
uint32_t crc2, uint32_t crc3,
uint32_t crc4) {}
#endif
static void hsw_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
display_pipe_crc_irq_handler(dev, pipe,
I915_READ(PIPE_CRC_RES_1_IVB(pipe)),
0, 0, 0, 0);
}
static void ivb_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
display_pipe_crc_irq_handler(dev, pipe,
I915_READ(PIPE_CRC_RES_1_IVB(pipe)),
I915_READ(PIPE_CRC_RES_2_IVB(pipe)),
I915_READ(PIPE_CRC_RES_3_IVB(pipe)),
I915_READ(PIPE_CRC_RES_4_IVB(pipe)),
I915_READ(PIPE_CRC_RES_5_IVB(pipe)));
}
static void i9xx_pipe_crc_irq_handler(struct drm_device *dev, enum pipe pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
uint32_t res1, res2;
if (INTEL_INFO(dev)->gen >= 3)
res1 = I915_READ(PIPE_CRC_RES_RES1_I915(pipe));
else
res1 = 0;
if (INTEL_INFO(dev)->gen >= 5 || IS_G4X(dev))
res2 = I915_READ(PIPE_CRC_RES_RES2_G4X(pipe));
else
res2 = 0;
display_pipe_crc_irq_handler(dev, pipe,
I915_READ(PIPE_CRC_RES_RED(pipe)),
I915_READ(PIPE_CRC_RES_GREEN(pipe)),
I915_READ(PIPE_CRC_RES_BLUE(pipe)),
res1, res2);
}
/* The RPS events need forcewake, so we add them to a work queue and mask their
* IMR bits until the work is done. Other interrupts can be processed without
* the work queue. */
static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
{
if (pm_iir & dev_priv->pm_rps_events) {
spin_lock(&dev_priv->irq_lock);
gen6_disable_pm_irq(dev_priv, pm_iir & dev_priv->pm_rps_events);
drm/i915: sanitize rps irq disabling When disabling the RPS interrupts there is a tricky dependency between the thread disabling the interrupts, the RPS interrupt handler and the corresponding RPS work. The RPS work can reenable the interrupts, so there is no straightforward order in the disabling thread to (1) make sure that any RPS work is flushed and to (2) disable all RPS interrupts. Currently this is solved by masking the interrupts using two separate mask registers (first level display IMR and PM IMR) and doing the disabling when all first level interrupts are disabled. This works, but the requirement to run with all first level interrupts disabled is unnecessary making the suspend / unload time ordering of RPS disabling wrt. other unitialization steps difficult and error prone. Removing this restriction allows us to disable RPS early during suspend / unload and forget about it for the rest of the sequence. By adding a more explicit method for avoiding the above race, it also becomes easier to prove its correctness. Finally currently we can hit the WARN in snb_update_pm_irq(), when a final RPS work runs with the first level interrupts already disabled. This won't lead to any problem (due to the separate interrupt masks), but with the change in this and the next patch we can get rid of the WARN, while leaving it in place for other scenarios. To address the above points, add a new RPS interrupts_enabled flag and use this during RPS disabling to avoid requeuing the RPS work and reenabling of the RPS interrupts. Since the interrupt disabling happens now in intel_suspend_gt_powersave(), we will disable RPS interrupts explicitly during suspend (and not just through the first level mask), but there is no problem doing so, it's also more consistent and allows us to unify more of the RPS disabling during suspend and unload time in the next patch. v2/v3: - rebase on patch "drm/i915: move rps irq disable one level up" in the patchset Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-11-19 21:30:04 +08:00
if (dev_priv->rps.interrupts_enabled) {
dev_priv->rps.pm_iir |= pm_iir & dev_priv->pm_rps_events;
queue_work(dev_priv->wq, &dev_priv->rps.work);
}
spin_unlock(&dev_priv->irq_lock);
}
if (INTEL_INFO(dev_priv)->gen >= 8)
return;
if (HAS_VEBOX(dev_priv)) {
if (pm_iir & PM_VEBOX_USER_INTERRUPT)
notify_ring(&dev_priv->engine[VECS]);
if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT)
DRM_DEBUG("Command parser error, pm_iir 0x%08x\n", pm_iir);
}
}
drm/i915: Make sprite updates atomic Add a mechanism by which we can evade the leading edge of vblank. This guarantees that no two sprite register writes will straddle on either side of the vblank start, and that means all the writes will be latched together in one atomic operation. We do the vblank evade by checking the scanline counter, and if it's too close to the start of vblank (too close has been hardcoded to 100usec for now), we will wait for the vblank start to pass. In order to eliminate random delayes from the rest of the system, we operate with interrupts disabled, except when waiting for the vblank obviously. Note that we now go digging through pipe_to_crtc_mapping[] in the vblank interrupt handler, which is a bit dangerous since we set up interrupts before the crtcs. However in this case since it's the vblank interrupt, we don't actually unmask it until some piece of code requests it. v2: preempt_check_resched() calls after local_irq_enable() (Jesse) Hook up the vblank irq stuff on BDW as well v3: Pass intel_crtc instead of drm_crtc (Daniel) Warn if crtc.mutex isn't locked (Daniel) Add an explicit compiler barrier and document the barriers (Daniel) Note the irq vs. modeset setup madness in the commit message (Daniel) v4: Use prepare_to_wait() & co. directly and eliminate vbl_received v5: Refactor intel_pipe_handle_vblank() vs. drm_handle_vblank() (Chris) Check for min/max scanline <= 0 (Chris) Don't call intel_pipe_update_end() if start failed totally (Chris) Check that the vblank counters match on both sides of the critical section (Chris) v6: Fix atomic update for interlaced modes v7: Reorder code for better readability (Chris) v8: Drop preempt_check_resched(). It's not available to modules anymore and isn't even needed unless we ourselves cause a wakeup needing reschedule while interrupts are off Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:46 +08:00
static bool intel_pipe_handle_vblank(struct drm_device *dev, enum pipe pipe)
{
if (!drm_handle_vblank(dev, pipe))
return false;
return true;
}
static void valleyview_pipestat_irq_handler(struct drm_device *dev, u32 iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 pipe_stats[I915_MAX_PIPES] = { };
int pipe;
spin_lock(&dev_priv->irq_lock);
if (!dev_priv->display_irqs_enabled) {
spin_unlock(&dev_priv->irq_lock);
return;
}
for_each_pipe(dev_priv, pipe) {
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg;
u32 mask, iir_bit = 0;
/*
* PIPESTAT bits get signalled even when the interrupt is
* disabled with the mask bits, and some of the status bits do
* not generate interrupts at all (like the underrun bit). Hence
* we need to be careful that we only handle what we want to
* handle.
*/
/* fifo underruns are filterered in the underrun handler. */
mask = PIPE_FIFO_UNDERRUN_STATUS;
switch (pipe) {
case PIPE_A:
iir_bit = I915_DISPLAY_PIPE_A_EVENT_INTERRUPT;
break;
case PIPE_B:
iir_bit = I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
break;
case PIPE_C:
iir_bit = I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
break;
}
if (iir & iir_bit)
mask |= dev_priv->pipestat_irq_mask[pipe];
if (!mask)
continue;
reg = PIPESTAT(pipe);
mask |= PIPESTAT_INT_ENABLE_MASK;
pipe_stats[pipe] = I915_READ(reg) & mask;
/*
* Clear the PIPE*STAT regs before the IIR
*/
if (pipe_stats[pipe] & (PIPE_FIFO_UNDERRUN_STATUS |
PIPESTAT_INT_STATUS_MASK))
I915_WRITE(reg, pipe_stats[pipe]);
}
spin_unlock(&dev_priv->irq_lock);
for_each_pipe(dev_priv, pipe) {
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
intel_pipe_handle_vblank(dev, pipe))
intel_check_page_flip(dev, pipe);
if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV) {
intel_prepare_page_flip(dev, pipe);
intel_finish_page_flip(dev, pipe);
}
if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
i9xx_pipe_crc_irq_handler(dev, pipe);
if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
}
if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
gmbus_irq_handler(dev);
}
static void i9xx_hpd_irq_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
u32 pin_mask = 0, long_mask = 0;
if (!hotplug_status)
return;
I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
/*
* Make sure hotplug status is cleared before we clear IIR, or else we
* may miss hotplug events.
*/
POSTING_READ(PORT_HOTPLUG_STAT);
if (IS_G4X(dev) || IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)) {
u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_G4X;
if (hotplug_trigger) {
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
hotplug_trigger, hpd_status_g4x,
i9xx_port_hotplug_long_detect);
intel_hpd_irq_handler(dev, pin_mask, long_mask);
}
if (hotplug_status & DP_AUX_CHANNEL_MASK_INT_STATUS_G4X)
dp_aux_irq_handler(dev);
} else {
u32 hotplug_trigger = hotplug_status & HOTPLUG_INT_STATUS_I915;
if (hotplug_trigger) {
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
hotplug_trigger, hpd_status_i915,
i9xx_port_hotplug_long_detect);
intel_hpd_irq_handler(dev, pin_mask, long_mask);
}
}
}
static irqreturn_t valleyview_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 iir, gt_iir, pm_iir;
irqreturn_t ret = IRQ_NONE;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
while (true) {
/* Find, clear, then process each source of interrupt */
gt_iir = I915_READ(GTIIR);
if (gt_iir)
I915_WRITE(GTIIR, gt_iir);
pm_iir = I915_READ(GEN6_PMIIR);
if (pm_iir)
I915_WRITE(GEN6_PMIIR, pm_iir);
iir = I915_READ(VLV_IIR);
if (iir) {
/* Consume port before clearing IIR or we'll miss events */
if (iir & I915_DISPLAY_PORT_INTERRUPT)
i9xx_hpd_irq_handler(dev);
I915_WRITE(VLV_IIR, iir);
}
if (gt_iir == 0 && pm_iir == 0 && iir == 0)
goto out;
ret = IRQ_HANDLED;
if (gt_iir)
snb_gt_irq_handler(dev, dev_priv, gt_iir);
if (pm_iir)
gen6_rps_irq_handler(dev_priv, pm_iir);
/* Call regardless, as some status bits might not be
* signalled in iir */
valleyview_pipestat_irq_handler(dev, iir);
}
out:
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
enable_rpm_wakeref_asserts(dev_priv);
return ret;
}
static irqreturn_t cherryview_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 master_ctl, iir;
irqreturn_t ret = IRQ_NONE;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
drm/i915: Exit cherryview_irq_handler() after one pass This effectively reverts commit 8e5fd599eb219f1054e39b40d18b217af669eea9 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Wed Apr 9 13:28:50 2014 +0300 drm/i915/chv: Make CHV irq handler loop until all interrupts are consumed as under continuous execlists load we can saturate the IRQ handler, destablising the tsc clock and triggering the NMI watchdog to declare a hung CPU. [ 552.756051] clocksource: timekeeping watchdog on CPU0: Marking clocksource 'tsc' as unstable because the skew is too large: [ 552.756080] clocksource: 'refined-jiffies' wd_now: 10003b480 wd_last: 10003b28c mask: ffffffff [ 552.756091] clocksource: 'tsc' cs_now: d55d31aa50 cs_last: d17446166c mask: ffffffffffffffff [ 552.756210] clocksource: Switched to clocksource refined-jiffies [ 575.217870] NMI watchdog: Watchdog detected hard LOCKUP on cpu 1 [ 575.217893] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.5.0-rc7+ #18 [ 575.217905] Hardware name: /NUC5CPYB, BIOS PYBSWCEL.86A.0027.2015.0507.1758 05/07/2015 [ 575.217915] 0000000000000000 ffff88027fd05bc0 ffffffff81288c6d 0000000000000000 [ 575.217935] 0000000000000001 ffff88027fd05be0 ffffffff810e72d1 0000000000000000 [ 575.217951] ffff88027fd05c80 ffff88027fd05c20 ffffffff81114b60 0000000181015f1e [ 575.217967] Call Trace: [ 575.217973] <NMI> [<ffffffff81288c6d>] dump_stack+0x4f/0x72 [ 575.217994] [<ffffffff810e72d1>] watchdog_overflow_callback+0x151/0x160 [ 575.218003] [<ffffffff81114b60>] __perf_event_overflow+0xa0/0x1e0 [ 575.218016] [<ffffffff811154c4>] perf_event_overflow+0x14/0x20 [ 575.218028] [<ffffffff8101d2ca>] intel_pmu_handle_irq+0x1da/0x460 [ 575.218042] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218052] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218064] [<ffffffff81014ae8>] perf_event_nmi_handler+0x28/0x50 [ 575.218075] [<ffffffff81007540>] nmi_handle+0x60/0x130 [ 575.218086] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218096] [<ffffffff810079c0>] do_nmi+0x140/0x470 [ 575.218108] [<ffffffff81559ec7>] end_repeat_nmi+0x1a/0x1e [ 575.218119] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218129] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218139] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218148] <<EOE>> [<ffffffff814a8353>] cpuidle_enter_state+0xf3/0x2f0 [ 575.218164] [<ffffffff814a8587>] cpuidle_enter+0x17/0x20 [ 575.218175] [<ffffffff810aaa3a>] call_cpuidle+0x2a/0x40 [ 575.218185] [<ffffffff810aade3>] cpu_startup_entry+0x273/0x330 [ 575.218196] [<ffffffff81033a1e>] start_secondary+0x10e/0x130 However, not servicing all available IIR within the handler does hurt the throughput of pathological nop execbuf by about 20%, with a similar effect upon the dispatch latency of a series of execbuf. v2: use do {} while(0) for a smaller patch, and easier to revert again I have reasonable confidence that we do not miss GT interrupts (as execlists provides a stress case with a failure mechanism easily detected by igt), however I have less confidence about all the other sources of interrupts and worry that may lose a display hotplug interrupt, for example. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93467 Testcase: igt/gem_exec_nop/basic # requires NMI watchdog Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Antti Koskipää <antti.koskipaa@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: stable@vger.kernel.org Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1457946117-6714-1-git-send-email-chris@chris-wilson.co.uk
2016-03-14 17:01:57 +08:00
do {
master_ctl = I915_READ(GEN8_MASTER_IRQ) & ~GEN8_MASTER_IRQ_CONTROL;
iir = I915_READ(VLV_IIR);
if (master_ctl == 0 && iir == 0)
break;
ret = IRQ_HANDLED;
I915_WRITE(GEN8_MASTER_IRQ, 0);
/* Find, clear, then process each source of interrupt */
if (iir) {
/* Consume port before clearing IIR or we'll miss events */
if (iir & I915_DISPLAY_PORT_INTERRUPT)
i9xx_hpd_irq_handler(dev);
I915_WRITE(VLV_IIR, iir);
}
gen8_gt_irq_handler(dev_priv, master_ctl);
/* Call regardless, as some status bits might not be
* signalled in iir */
valleyview_pipestat_irq_handler(dev, iir);
I915_WRITE(GEN8_MASTER_IRQ, DE_MASTER_IRQ_CONTROL);
POSTING_READ(GEN8_MASTER_IRQ);
drm/i915: Exit cherryview_irq_handler() after one pass This effectively reverts commit 8e5fd599eb219f1054e39b40d18b217af669eea9 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Wed Apr 9 13:28:50 2014 +0300 drm/i915/chv: Make CHV irq handler loop until all interrupts are consumed as under continuous execlists load we can saturate the IRQ handler, destablising the tsc clock and triggering the NMI watchdog to declare a hung CPU. [ 552.756051] clocksource: timekeeping watchdog on CPU0: Marking clocksource 'tsc' as unstable because the skew is too large: [ 552.756080] clocksource: 'refined-jiffies' wd_now: 10003b480 wd_last: 10003b28c mask: ffffffff [ 552.756091] clocksource: 'tsc' cs_now: d55d31aa50 cs_last: d17446166c mask: ffffffffffffffff [ 552.756210] clocksource: Switched to clocksource refined-jiffies [ 575.217870] NMI watchdog: Watchdog detected hard LOCKUP on cpu 1 [ 575.217893] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.5.0-rc7+ #18 [ 575.217905] Hardware name: /NUC5CPYB, BIOS PYBSWCEL.86A.0027.2015.0507.1758 05/07/2015 [ 575.217915] 0000000000000000 ffff88027fd05bc0 ffffffff81288c6d 0000000000000000 [ 575.217935] 0000000000000001 ffff88027fd05be0 ffffffff810e72d1 0000000000000000 [ 575.217951] ffff88027fd05c80 ffff88027fd05c20 ffffffff81114b60 0000000181015f1e [ 575.217967] Call Trace: [ 575.217973] <NMI> [<ffffffff81288c6d>] dump_stack+0x4f/0x72 [ 575.217994] [<ffffffff810e72d1>] watchdog_overflow_callback+0x151/0x160 [ 575.218003] [<ffffffff81114b60>] __perf_event_overflow+0xa0/0x1e0 [ 575.218016] [<ffffffff811154c4>] perf_event_overflow+0x14/0x20 [ 575.218028] [<ffffffff8101d2ca>] intel_pmu_handle_irq+0x1da/0x460 [ 575.218042] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218052] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218064] [<ffffffff81014ae8>] perf_event_nmi_handler+0x28/0x50 [ 575.218075] [<ffffffff81007540>] nmi_handle+0x60/0x130 [ 575.218086] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218096] [<ffffffff810079c0>] do_nmi+0x140/0x470 [ 575.218108] [<ffffffff81559ec7>] end_repeat_nmi+0x1a/0x1e [ 575.218119] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218129] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218139] [<ffffffff814a8aae>] ? poll_idle+0x3e/0x70 [ 575.218148] <<EOE>> [<ffffffff814a8353>] cpuidle_enter_state+0xf3/0x2f0 [ 575.218164] [<ffffffff814a8587>] cpuidle_enter+0x17/0x20 [ 575.218175] [<ffffffff810aaa3a>] call_cpuidle+0x2a/0x40 [ 575.218185] [<ffffffff810aade3>] cpu_startup_entry+0x273/0x330 [ 575.218196] [<ffffffff81033a1e>] start_secondary+0x10e/0x130 However, not servicing all available IIR within the handler does hurt the throughput of pathological nop execbuf by about 20%, with a similar effect upon the dispatch latency of a series of execbuf. v2: use do {} while(0) for a smaller patch, and easier to revert again I have reasonable confidence that we do not miss GT interrupts (as execlists provides a stress case with a failure mechanism easily detected by igt), however I have less confidence about all the other sources of interrupts and worry that may lose a display hotplug interrupt, for example. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93467 Testcase: igt/gem_exec_nop/basic # requires NMI watchdog Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Antti Koskipää <antti.koskipaa@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: stable@vger.kernel.org Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1457946117-6714-1-git-send-email-chris@chris-wilson.co.uk
2016-03-14 17:01:57 +08:00
} while (0);
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
enable_rpm_wakeref_asserts(dev_priv);
return ret;
}
static void ibx_hpd_irq_handler(struct drm_device *dev, u32 hotplug_trigger,
const u32 hpd[HPD_NUM_PINS])
{
struct drm_i915_private *dev_priv = to_i915(dev);
u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
drm/i915: fix the SDE irq dmesg warnings properly We had the "The master control interrupt lied (SDE)!" check and error message in place for a long time without any problems, until commit aaf5ec2e51ab1d9c5e962b4728a1107ed3ff7a3e Author: Sonika Jindal <sonika.jindal@intel.com> Date: Wed Jul 8 17:07:47 2015 +0530 drm/i915: Handle HPD when it has actually occurred caused the errors to start happening. This was bisected and reported, but the error message was silenced in commit 97e5ed1111dcc5300a0f59a55248cd243937a8ab Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Oct 23 10:56:12 2015 +0200 drm/i915: shut up gen8+ SDE irq dmesg noise shooting the messenger while the debugging for why Sonika's commit triggered the errors was still in progress. It looks like we need to read and acknowledge the PCH_PORT_HOTPLUG register even though the hotplug trigger indicates there isn't a hotplug irq to handle. The PCH doesn't seem to really ack the the interrupt to the CPU unless we touch the hotplug register. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sonika Jindal <sonika.jindal@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92084 Fixes: aaf5ec2e51ab ("drm/i915: Handle HPD when it has actually occurred") [Jani: added a comment and amended the commit message while applying] Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448462843-32739-1-git-send-email-jani.nikula@intel.com
2015-11-25 22:47:22 +08:00
/*
* Somehow the PCH doesn't seem to really ack the interrupt to the CPU
* unless we touch the hotplug register, even if hotplug_trigger is
* zero. Not acking leads to "The master control interrupt lied (SDE)!"
* errors.
*/
dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG);
drm/i915: fix the SDE irq dmesg warnings properly We had the "The master control interrupt lied (SDE)!" check and error message in place for a long time without any problems, until commit aaf5ec2e51ab1d9c5e962b4728a1107ed3ff7a3e Author: Sonika Jindal <sonika.jindal@intel.com> Date: Wed Jul 8 17:07:47 2015 +0530 drm/i915: Handle HPD when it has actually occurred caused the errors to start happening. This was bisected and reported, but the error message was silenced in commit 97e5ed1111dcc5300a0f59a55248cd243937a8ab Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Oct 23 10:56:12 2015 +0200 drm/i915: shut up gen8+ SDE irq dmesg noise shooting the messenger while the debugging for why Sonika's commit triggered the errors was still in progress. It looks like we need to read and acknowledge the PCH_PORT_HOTPLUG register even though the hotplug trigger indicates there isn't a hotplug irq to handle. The PCH doesn't seem to really ack the the interrupt to the CPU unless we touch the hotplug register. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sonika Jindal <sonika.jindal@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92084 Fixes: aaf5ec2e51ab ("drm/i915: Handle HPD when it has actually occurred") [Jani: added a comment and amended the commit message while applying] Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448462843-32739-1-git-send-email-jani.nikula@intel.com
2015-11-25 22:47:22 +08:00
if (!hotplug_trigger) {
u32 mask = PORTA_HOTPLUG_STATUS_MASK |
PORTD_HOTPLUG_STATUS_MASK |
PORTC_HOTPLUG_STATUS_MASK |
PORTB_HOTPLUG_STATUS_MASK;
dig_hotplug_reg &= ~mask;
}
I915_WRITE(PCH_PORT_HOTPLUG, dig_hotplug_reg);
drm/i915: fix the SDE irq dmesg warnings properly We had the "The master control interrupt lied (SDE)!" check and error message in place for a long time without any problems, until commit aaf5ec2e51ab1d9c5e962b4728a1107ed3ff7a3e Author: Sonika Jindal <sonika.jindal@intel.com> Date: Wed Jul 8 17:07:47 2015 +0530 drm/i915: Handle HPD when it has actually occurred caused the errors to start happening. This was bisected and reported, but the error message was silenced in commit 97e5ed1111dcc5300a0f59a55248cd243937a8ab Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Oct 23 10:56:12 2015 +0200 drm/i915: shut up gen8+ SDE irq dmesg noise shooting the messenger while the debugging for why Sonika's commit triggered the errors was still in progress. It looks like we need to read and acknowledge the PCH_PORT_HOTPLUG register even though the hotplug trigger indicates there isn't a hotplug irq to handle. The PCH doesn't seem to really ack the the interrupt to the CPU unless we touch the hotplug register. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sonika Jindal <sonika.jindal@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92084 Fixes: aaf5ec2e51ab ("drm/i915: Handle HPD when it has actually occurred") [Jani: added a comment and amended the commit message while applying] Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448462843-32739-1-git-send-email-jani.nikula@intel.com
2015-11-25 22:47:22 +08:00
if (!hotplug_trigger)
return;
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
dig_hotplug_reg, hpd,
pch_port_hotplug_long_detect);
intel_hpd_irq_handler(dev, pin_mask, long_mask);
}
static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
2013-04-16 19:36:54 +08:00
u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
drm/i915: fix the SDE irq dmesg warnings properly We had the "The master control interrupt lied (SDE)!" check and error message in place for a long time without any problems, until commit aaf5ec2e51ab1d9c5e962b4728a1107ed3ff7a3e Author: Sonika Jindal <sonika.jindal@intel.com> Date: Wed Jul 8 17:07:47 2015 +0530 drm/i915: Handle HPD when it has actually occurred caused the errors to start happening. This was bisected and reported, but the error message was silenced in commit 97e5ed1111dcc5300a0f59a55248cd243937a8ab Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Oct 23 10:56:12 2015 +0200 drm/i915: shut up gen8+ SDE irq dmesg noise shooting the messenger while the debugging for why Sonika's commit triggered the errors was still in progress. It looks like we need to read and acknowledge the PCH_PORT_HOTPLUG register even though the hotplug trigger indicates there isn't a hotplug irq to handle. The PCH doesn't seem to really ack the the interrupt to the CPU unless we touch the hotplug register. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sonika Jindal <sonika.jindal@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92084 Fixes: aaf5ec2e51ab ("drm/i915: Handle HPD when it has actually occurred") [Jani: added a comment and amended the commit message while applying] Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448462843-32739-1-git-send-email-jani.nikula@intel.com
2015-11-25 22:47:22 +08:00
ibx_hpd_irq_handler(dev, hotplug_trigger, hpd_ibx);
if (pch_iir & SDE_AUDIO_POWER_MASK) {
int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK) >>
SDE_AUDIO_POWER_SHIFT);
DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
port_name(port));
}
if (pch_iir & SDE_AUX_MASK)
dp_aux_irq_handler(dev);
if (pch_iir & SDE_GMBUS)
gmbus_irq_handler(dev);
if (pch_iir & SDE_AUDIO_HDCP_MASK)
DRM_DEBUG_DRIVER("PCH HDCP audio interrupt\n");
if (pch_iir & SDE_AUDIO_TRANS_MASK)
DRM_DEBUG_DRIVER("PCH transcoder audio interrupt\n");
if (pch_iir & SDE_POISON)
DRM_ERROR("PCH poison interrupt\n");
if (pch_iir & SDE_FDI_MASK)
for_each_pipe(dev_priv, pipe)
DRM_DEBUG_DRIVER(" pipe %c FDI IIR: 0x%08x\n",
pipe_name(pipe),
I915_READ(FDI_RX_IIR(pipe)));
if (pch_iir & (SDE_TRANSB_CRC_DONE | SDE_TRANSA_CRC_DONE))
DRM_DEBUG_DRIVER("PCH transcoder CRC done interrupt\n");
if (pch_iir & (SDE_TRANSB_CRC_ERR | SDE_TRANSA_CRC_ERR))
DRM_DEBUG_DRIVER("PCH transcoder CRC error interrupt\n");
if (pch_iir & SDE_TRANSA_FIFO_UNDER)
intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_A);
if (pch_iir & SDE_TRANSB_FIFO_UNDER)
intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_B);
}
static void ivb_err_int_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 err_int = I915_READ(GEN7_ERR_INT);
enum pipe pipe;
if (err_int & ERR_INT_POISON)
DRM_ERROR("Poison interrupt\n");
for_each_pipe(dev_priv, pipe) {
if (err_int & ERR_INT_FIFO_UNDERRUN(pipe))
intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
if (err_int & ERR_INT_PIPE_CRC_DONE(pipe)) {
if (IS_IVYBRIDGE(dev))
ivb_pipe_crc_irq_handler(dev, pipe);
else
hsw_pipe_crc_irq_handler(dev, pipe);
}
}
I915_WRITE(GEN7_ERR_INT, err_int);
}
static void cpt_serr_int_handler(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 serr_int = I915_READ(SERR_INT);
if (serr_int & SERR_INT_POISON)
DRM_ERROR("PCH poison interrupt\n");
if (serr_int & SERR_INT_TRANS_A_FIFO_UNDERRUN)
intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_A);
if (serr_int & SERR_INT_TRANS_B_FIFO_UNDERRUN)
intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_B);
if (serr_int & SERR_INT_TRANS_C_FIFO_UNDERRUN)
intel_pch_fifo_underrun_irq_handler(dev_priv, TRANSCODER_C);
I915_WRITE(SERR_INT, serr_int);
}
static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
drm/i915: fix the SDE irq dmesg warnings properly We had the "The master control interrupt lied (SDE)!" check and error message in place for a long time without any problems, until commit aaf5ec2e51ab1d9c5e962b4728a1107ed3ff7a3e Author: Sonika Jindal <sonika.jindal@intel.com> Date: Wed Jul 8 17:07:47 2015 +0530 drm/i915: Handle HPD when it has actually occurred caused the errors to start happening. This was bisected and reported, but the error message was silenced in commit 97e5ed1111dcc5300a0f59a55248cd243937a8ab Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Fri Oct 23 10:56:12 2015 +0200 drm/i915: shut up gen8+ SDE irq dmesg noise shooting the messenger while the debugging for why Sonika's commit triggered the errors was still in progress. It looks like we need to read and acknowledge the PCH_PORT_HOTPLUG register even though the hotplug trigger indicates there isn't a hotplug irq to handle. The PCH doesn't seem to really ack the the interrupt to the CPU unless we touch the hotplug register. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Sonika Jindal <sonika.jindal@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=92084 Fixes: aaf5ec2e51ab ("drm/i915: Handle HPD when it has actually occurred") [Jani: added a comment and amended the commit message while applying] Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448462843-32739-1-git-send-email-jani.nikula@intel.com
2015-11-25 22:47:22 +08:00
ibx_hpd_irq_handler(dev, hotplug_trigger, hpd_cpt);
if (pch_iir & SDE_AUDIO_POWER_MASK_CPT) {
int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
SDE_AUDIO_POWER_SHIFT_CPT);
DRM_DEBUG_DRIVER("PCH audio power change on port %c\n",
port_name(port));
}
if (pch_iir & SDE_AUX_MASK_CPT)
dp_aux_irq_handler(dev);
if (pch_iir & SDE_GMBUS_CPT)
gmbus_irq_handler(dev);
if (pch_iir & SDE_AUDIO_CP_REQ_CPT)
DRM_DEBUG_DRIVER("Audio CP request interrupt\n");
if (pch_iir & SDE_AUDIO_CP_CHG_CPT)
DRM_DEBUG_DRIVER("Audio CP change interrupt\n");
if (pch_iir & SDE_FDI_MASK_CPT)
for_each_pipe(dev_priv, pipe)
DRM_DEBUG_DRIVER(" pipe %c FDI IIR: 0x%08x\n",
pipe_name(pipe),
I915_READ(FDI_RX_IIR(pipe)));
if (pch_iir & SDE_ERROR_CPT)
cpt_serr_int_handler(dev);
}
static void spt_irq_handler(struct drm_device *dev, u32 pch_iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_SPT &
~SDE_PORTE_HOTPLUG_SPT;
u32 hotplug2_trigger = pch_iir & SDE_PORTE_HOTPLUG_SPT;
u32 pin_mask = 0, long_mask = 0;
if (hotplug_trigger) {
u32 dig_hotplug_reg;
dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG);
I915_WRITE(PCH_PORT_HOTPLUG, dig_hotplug_reg);
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
dig_hotplug_reg, hpd_spt,
spt_port_hotplug_long_detect);
}
if (hotplug2_trigger) {
u32 dig_hotplug_reg;
dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG2);
I915_WRITE(PCH_PORT_HOTPLUG2, dig_hotplug_reg);
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug2_trigger,
dig_hotplug_reg, hpd_spt,
spt_port_hotplug2_long_detect);
}
if (pin_mask)
intel_hpd_irq_handler(dev, pin_mask, long_mask);
if (pch_iir & SDE_GMBUS_CPT)
gmbus_irq_handler(dev);
}
static void ilk_hpd_irq_handler(struct drm_device *dev, u32 hotplug_trigger,
const u32 hpd[HPD_NUM_PINS])
{
struct drm_i915_private *dev_priv = to_i915(dev);
u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
dig_hotplug_reg = I915_READ(DIGITAL_PORT_HOTPLUG_CNTRL);
I915_WRITE(DIGITAL_PORT_HOTPLUG_CNTRL, dig_hotplug_reg);
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
dig_hotplug_reg, hpd,
ilk_port_hotplug_long_detect);
intel_hpd_irq_handler(dev, pin_mask, long_mask);
}
static void ilk_display_irq_handler(struct drm_device *dev, u32 de_iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
enum pipe pipe;
u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG;
if (hotplug_trigger)
ilk_hpd_irq_handler(dev, hotplug_trigger, hpd_ilk);
if (de_iir & DE_AUX_CHANNEL_A)
dp_aux_irq_handler(dev);
if (de_iir & DE_GSE)
intel_opregion_asle_intr(dev);
if (de_iir & DE_POISON)
DRM_ERROR("Poison interrupt\n");
for_each_pipe(dev_priv, pipe) {
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
if (de_iir & DE_PIPE_VBLANK(pipe) &&
intel_pipe_handle_vblank(dev, pipe))
intel_check_page_flip(dev, pipe);
if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
if (de_iir & DE_PIPE_CRC_DONE(pipe))
i9xx_pipe_crc_irq_handler(dev, pipe);
/* plane/pipes map 1:1 on ilk+ */
if (de_iir & DE_PLANE_FLIP_DONE(pipe)) {
intel_prepare_page_flip(dev, pipe);
intel_finish_page_flip_plane(dev, pipe);
}
}
/* check event from PCH */
if (de_iir & DE_PCH_EVENT) {
u32 pch_iir = I915_READ(SDEIIR);
if (HAS_PCH_CPT(dev))
cpt_irq_handler(dev, pch_iir);
else
ibx_irq_handler(dev, pch_iir);
/* should clear PCH hotplug event before clear CPU irq */
I915_WRITE(SDEIIR, pch_iir);
}
if (IS_GEN5(dev) && de_iir & DE_PCU_EVENT)
ironlake_rps_change_irq_handler(dev);
}
static void ivb_display_irq_handler(struct drm_device *dev, u32 de_iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
enum pipe pipe;
u32 hotplug_trigger = de_iir & DE_DP_A_HOTPLUG_IVB;
if (hotplug_trigger)
ilk_hpd_irq_handler(dev, hotplug_trigger, hpd_ivb);
if (de_iir & DE_ERR_INT_IVB)
ivb_err_int_handler(dev);
if (de_iir & DE_AUX_CHANNEL_A_IVB)
dp_aux_irq_handler(dev);
if (de_iir & DE_GSE_IVB)
intel_opregion_asle_intr(dev);
for_each_pipe(dev_priv, pipe) {
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)) &&
intel_pipe_handle_vblank(dev, pipe))
intel_check_page_flip(dev, pipe);
/* plane/pipes map 1:1 on ilk+ */
if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe)) {
intel_prepare_page_flip(dev, pipe);
intel_finish_page_flip_plane(dev, pipe);
}
}
/* check event from PCH */
if (!HAS_PCH_NOP(dev) && (de_iir & DE_PCH_EVENT_IVB)) {
u32 pch_iir = I915_READ(SDEIIR);
cpt_irq_handler(dev, pch_iir);
/* clear PCH hotplug event before clear CPU irq */
I915_WRITE(SDEIIR, pch_iir);
}
}
/*
* To handle irqs with the minimum potential races with fresh interrupts, we:
* 1 - Disable Master Interrupt Control.
* 2 - Find the source(s) of the interrupt.
* 3 - Clear the Interrupt Identity bits (IIR).
* 4 - Process the interrupt(s) that had bits set in the IIRs.
* 5 - Re-enable Master Interrupt Control.
*/
static irqreturn_t ironlake_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 de_iir, gt_iir, de_ier, sde_ier = 0;
irqreturn_t ret = IRQ_NONE;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
/* disable master interrupt before clearing iir */
de_ier = I915_READ(DEIER);
I915_WRITE(DEIER, de_ier & ~DE_MASTER_IRQ_CONTROL);
POSTING_READ(DEIER);
drm/i915: also disable south interrupts when handling them From the docs: "IIR can queue up to two interrupt events. When the IIR is cleared, it will set itself again after one clock if a second event was stored." "Only the rising edge of the PCH Display interrupt will cause the North Display IIR (DEIIR) PCH Display Interrupt even bit to be set, so all PCH Display Interrupts, including back to back interrupts, must be cleared before a new PCH Display interrupt can cause DEIIR to be set". The current code works fine because we don't get many interrupts, but if we enable the PCH FIFO underrun interrupts we'll start getting so many interrupts that at some point new PCH interrupts won't cause DEIIR to be set. The initial implementation I tried was to turn the code that checks SDEIIR into a loop, but we can still get interrupts even after the loop is done (and before the irq handler finishes), so we have to either disable the interrupts or mask them. In the end I concluded that just disabling the PCH interrupts is enough, you don't even need the loop, so this is what this patch implements. I've tested it and it passes the 2 "PCH FIFO underrun interrupt storms" I can reproduce: the "ironlake_crtc_disable" case and the "wrong watermarks" case. In other words, here's how to reproduce the problem fixed by this patch: 1 - Enable PCH FIFO underrun interrupts (SERR_INT on SNB+) 2 - Boot the machine 3 - While booting we'll get tons of PCH FIFO underrun interrupts 4 - Plug a new monitor 5 - Run xrandr, notice it won't detect the new monitor 6 - Read SDEIIR and notice it's not 0 while DEIIR is 0 Q: Can't we just clear DEIIR before SDEIIR? A: It doesn't work. SDEIIR has to be completely cleared (including the interrupts stored on its back queue) before it can flip DEIIR's bit to 1 again, and even while you're clearing it you'll be getting more and more interrupts. Q: Why does it work by just disabling+enabling the south interrupts? A: Because when we re-enable them, if there's something on the SDEIIR register (maybe an interrupt stored on the queue), the re-enabling will make DEIIR's bit flip to 1, and since we'll already have interrupts enabled we'll get another interrupt, then run our irq handler again to process the "back" interrupts. v2: Even bigger commit message, added code comments. Note that this fixes missed dp aux irqs which have been reported for 3.9-rc1. This regression has been introduced by switching to irq-driven dp aux transactions with commit 9ee32fea5fe810ec06af3a15e4c65478de56d4f5 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 13:53:48 2012 +0100 drm/i915: irq-drive the dp aux communication References: http://www.mail-archive.com/intel-gfx@lists.freedesktop.org/msg18588.html References: https://lkml.org/lkml/2013/2/26/769 Tested-by: Imre Deak <imre.deak@intel.com> Reported-by: Sedat Dilek <sedat.dilek@gmail.com> Reported-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> [danvet: Pimp commit message with references for the dp aux irq timeout regression this fixes.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-02-23 04:05:28 +08:00
/* Disable south interrupts. We'll only write to SDEIIR once, so further
* interrupts will will be stored on its back queue, and then we'll be
* able to process them after we restore SDEIER (as soon as we restore
* it, we'll get an interrupt if SDEIIR still has something to process
* due to its back queue). */
if (!HAS_PCH_NOP(dev)) {
sde_ier = I915_READ(SDEIER);
I915_WRITE(SDEIER, 0);
POSTING_READ(SDEIER);
}
drm/i915: also disable south interrupts when handling them From the docs: "IIR can queue up to two interrupt events. When the IIR is cleared, it will set itself again after one clock if a second event was stored." "Only the rising edge of the PCH Display interrupt will cause the North Display IIR (DEIIR) PCH Display Interrupt even bit to be set, so all PCH Display Interrupts, including back to back interrupts, must be cleared before a new PCH Display interrupt can cause DEIIR to be set". The current code works fine because we don't get many interrupts, but if we enable the PCH FIFO underrun interrupts we'll start getting so many interrupts that at some point new PCH interrupts won't cause DEIIR to be set. The initial implementation I tried was to turn the code that checks SDEIIR into a loop, but we can still get interrupts even after the loop is done (and before the irq handler finishes), so we have to either disable the interrupts or mask them. In the end I concluded that just disabling the PCH interrupts is enough, you don't even need the loop, so this is what this patch implements. I've tested it and it passes the 2 "PCH FIFO underrun interrupt storms" I can reproduce: the "ironlake_crtc_disable" case and the "wrong watermarks" case. In other words, here's how to reproduce the problem fixed by this patch: 1 - Enable PCH FIFO underrun interrupts (SERR_INT on SNB+) 2 - Boot the machine 3 - While booting we'll get tons of PCH FIFO underrun interrupts 4 - Plug a new monitor 5 - Run xrandr, notice it won't detect the new monitor 6 - Read SDEIIR and notice it's not 0 while DEIIR is 0 Q: Can't we just clear DEIIR before SDEIIR? A: It doesn't work. SDEIIR has to be completely cleared (including the interrupts stored on its back queue) before it can flip DEIIR's bit to 1 again, and even while you're clearing it you'll be getting more and more interrupts. Q: Why does it work by just disabling+enabling the south interrupts? A: Because when we re-enable them, if there's something on the SDEIIR register (maybe an interrupt stored on the queue), the re-enabling will make DEIIR's bit flip to 1, and since we'll already have interrupts enabled we'll get another interrupt, then run our irq handler again to process the "back" interrupts. v2: Even bigger commit message, added code comments. Note that this fixes missed dp aux irqs which have been reported for 3.9-rc1. This regression has been introduced by switching to irq-driven dp aux transactions with commit 9ee32fea5fe810ec06af3a15e4c65478de56d4f5 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 13:53:48 2012 +0100 drm/i915: irq-drive the dp aux communication References: http://www.mail-archive.com/intel-gfx@lists.freedesktop.org/msg18588.html References: https://lkml.org/lkml/2013/2/26/769 Tested-by: Imre Deak <imre.deak@intel.com> Reported-by: Sedat Dilek <sedat.dilek@gmail.com> Reported-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> [danvet: Pimp commit message with references for the dp aux irq timeout regression this fixes.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-02-23 04:05:28 +08:00
/* Find, clear, then process each source of interrupt */
gt_iir = I915_READ(GTIIR);
if (gt_iir) {
I915_WRITE(GTIIR, gt_iir);
ret = IRQ_HANDLED;
if (INTEL_INFO(dev)->gen >= 6)
snb_gt_irq_handler(dev, dev_priv, gt_iir);
else
ilk_gt_irq_handler(dev, dev_priv, gt_iir);
}
de_iir = I915_READ(DEIIR);
if (de_iir) {
I915_WRITE(DEIIR, de_iir);
ret = IRQ_HANDLED;
if (INTEL_INFO(dev)->gen >= 7)
ivb_display_irq_handler(dev, de_iir);
else
ilk_display_irq_handler(dev, de_iir);
}
if (INTEL_INFO(dev)->gen >= 6) {
u32 pm_iir = I915_READ(GEN6_PMIIR);
if (pm_iir) {
I915_WRITE(GEN6_PMIIR, pm_iir);
ret = IRQ_HANDLED;
gen6_rps_irq_handler(dev_priv, pm_iir);
}
}
I915_WRITE(DEIER, de_ier);
POSTING_READ(DEIER);
if (!HAS_PCH_NOP(dev)) {
I915_WRITE(SDEIER, sde_ier);
POSTING_READ(SDEIER);
}
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
enable_rpm_wakeref_asserts(dev_priv);
return ret;
}
static void bxt_hpd_irq_handler(struct drm_device *dev, u32 hotplug_trigger,
const u32 hpd[HPD_NUM_PINS])
{
struct drm_i915_private *dev_priv = to_i915(dev);
u32 dig_hotplug_reg, pin_mask = 0, long_mask = 0;
dig_hotplug_reg = I915_READ(PCH_PORT_HOTPLUG);
I915_WRITE(PCH_PORT_HOTPLUG, dig_hotplug_reg);
intel_get_hpd_pins(&pin_mask, &long_mask, hotplug_trigger,
dig_hotplug_reg, hpd,
bxt_port_hotplug_long_detect);
intel_hpd_irq_handler(dev, pin_mask, long_mask);
}
static irqreturn_t
gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
{
struct drm_device *dev = dev_priv->dev;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
irqreturn_t ret = IRQ_NONE;
u32 iir;
enum pipe pipe;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
if (master_ctl & GEN8_DE_MISC_IRQ) {
iir = I915_READ(GEN8_DE_MISC_IIR);
if (iir) {
I915_WRITE(GEN8_DE_MISC_IIR, iir);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
ret = IRQ_HANDLED;
if (iir & GEN8_DE_MISC_GSE)
intel_opregion_asle_intr(dev);
else
DRM_ERROR("Unexpected DE Misc interrupt\n");
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
else
DRM_ERROR("The master control interrupt lied (DE MISC)!\n");
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
if (master_ctl & GEN8_DE_PORT_IRQ) {
iir = I915_READ(GEN8_DE_PORT_IIR);
if (iir) {
u32 tmp_mask;
bool found = false;
I915_WRITE(GEN8_DE_PORT_IIR, iir);
ret = IRQ_HANDLED;
tmp_mask = GEN8_AUX_CHANNEL_A;
if (INTEL_INFO(dev_priv)->gen >= 9)
tmp_mask |= GEN9_AUX_CHANNEL_B |
GEN9_AUX_CHANNEL_C |
GEN9_AUX_CHANNEL_D;
if (iir & tmp_mask) {
dp_aux_irq_handler(dev);
found = true;
}
if (IS_BROXTON(dev_priv)) {
tmp_mask = iir & BXT_DE_PORT_HOTPLUG_MASK;
if (tmp_mask) {
bxt_hpd_irq_handler(dev, tmp_mask, hpd_bxt);
found = true;
}
} else if (IS_BROADWELL(dev_priv)) {
tmp_mask = iir & GEN8_PORT_DP_A_HOTPLUG;
if (tmp_mask) {
ilk_hpd_irq_handler(dev, tmp_mask, hpd_bdw);
found = true;
}
}
if (IS_BROXTON(dev) && (iir & BXT_DE_PORT_GMBUS)) {
gmbus_irq_handler(dev);
found = true;
}
if (!found)
DRM_ERROR("Unexpected DE Port interrupt\n");
}
else
DRM_ERROR("The master control interrupt lied (DE PORT)!\n");
}
for_each_pipe(dev_priv, pipe) {
u32 flip_done, fault_errors;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
if (!(master_ctl & GEN8_DE_PIPE_IRQ(pipe)))
continue;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
iir = I915_READ(GEN8_DE_PIPE_IIR(pipe));
if (!iir) {
DRM_ERROR("The master control interrupt lied (DE PIPE)!\n");
continue;
}
ret = IRQ_HANDLED;
I915_WRITE(GEN8_DE_PIPE_IIR(pipe), iir);
if (iir & GEN8_PIPE_VBLANK &&
intel_pipe_handle_vblank(dev, pipe))
intel_check_page_flip(dev, pipe);
flip_done = iir;
if (INTEL_INFO(dev_priv)->gen >= 9)
flip_done &= GEN9_PIPE_PLANE1_FLIP_DONE;
else
flip_done &= GEN8_PIPE_PRIMARY_FLIP_DONE;
if (flip_done) {
intel_prepare_page_flip(dev, pipe);
intel_finish_page_flip_plane(dev, pipe);
}
if (iir & GEN8_PIPE_CDCLK_CRC_DONE)
hsw_pipe_crc_irq_handler(dev, pipe);
if (iir & GEN8_PIPE_FIFO_UNDERRUN)
intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
fault_errors = iir;
if (INTEL_INFO(dev_priv)->gen >= 9)
fault_errors &= GEN9_DE_PIPE_IRQ_FAULT_ERRORS;
else
fault_errors &= GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
if (fault_errors)
DRM_ERROR("Fault errors on pipe %c\n: 0x%08x",
pipe_name(pipe),
fault_errors);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
if (HAS_PCH_SPLIT(dev) && !HAS_PCH_NOP(dev) &&
master_ctl & GEN8_DE_PCH_IRQ) {
/*
* FIXME(BDW): Assume for now that the new interrupt handling
* scheme also closed the SDE interrupt handling race we've seen
* on older pch-split platforms. But this needs testing.
*/
iir = I915_READ(SDEIIR);
if (iir) {
I915_WRITE(SDEIIR, iir);
ret = IRQ_HANDLED;
if (HAS_PCH_SPT(dev_priv))
spt_irq_handler(dev, iir);
else
cpt_irq_handler(dev, iir);
} else {
/*
* Like on previous PCH there seems to be something
* fishy going on with forwarding PCH interrupts.
*/
DRM_DEBUG_DRIVER("The master control interrupt lied (SDE)!\n");
}
}
return ret;
}
static irqreturn_t gen8_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 master_ctl;
irqreturn_t ret;
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
master_ctl = I915_READ_FW(GEN8_MASTER_IRQ);
master_ctl &= ~GEN8_MASTER_IRQ_CONTROL;
if (!master_ctl)
return IRQ_NONE;
I915_WRITE_FW(GEN8_MASTER_IRQ, 0);
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
/* Find, clear, then process each source of interrupt */
ret = gen8_gt_irq_handler(dev_priv, master_ctl);
ret |= gen8_de_irq_handler(dev_priv, master_ctl);
I915_WRITE_FW(GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL);
POSTING_READ_FW(GEN8_MASTER_IRQ);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
enable_rpm_wakeref_asserts(dev_priv);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
return ret;
}
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
static void i915_error_wake_up(struct drm_i915_private *dev_priv,
bool reset_completed)
{
struct intel_engine_cs *engine;
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
/*
* Notify all waiters for GPU completion events that reset state has
* been changed, and that they need to restart their wait after
* checking for potential errors (and bail out to drop locks if there is
* a gpu reset pending so that i915_error_work_func can acquire them).
*/
/* Wake up __wait_seqno, potentially holding dev->struct_mutex. */
for_each_engine(engine, dev_priv)
wake_up_all(&engine->irq_queue);
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
/* Wake up intel_crtc_wait_for_pending_flips, holding crtc->mutex. */
wake_up_all(&dev_priv->pending_flip_queue);
/*
* Signal tasks blocked in i915_gem_wait_for_error that the pending
* reset state is cleared.
*/
if (reset_completed)
wake_up_all(&dev_priv->gpu_error.reset_queue);
}
/**
* i915_reset_and_wakeup - do process context error handling work
* @dev: drm device
*
* Fire an error uevent so userspace can see that a hang or error
* was detected.
*/
static void i915_reset_and_wakeup(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = to_i915(dev);
struct i915_gpu_error *error = &dev_priv->gpu_error;
char *error_event[] = { I915_ERROR_UEVENT "=1", NULL };
char *reset_event[] = { I915_RESET_UEVENT "=1", NULL };
char *reset_done_event[] = { I915_ERROR_UEVENT "=0", NULL };
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
int ret;
kobject_uevent_env(&dev->primary->kdev->kobj, KOBJ_CHANGE, error_event);
/*
* Note that there's only one work item which does gpu resets, so we
* need not worry about concurrent gpu resets potentially incrementing
* error->reset_counter twice. We only need to take care of another
* racing irq/hangcheck declaring the gpu dead for a second time. A
* quick check for that is good enough: schedule_work ensures the
* correct ordering between hang detection and this work item, and since
* the reset in-progress bit is only ever set by code outside of this
* work we don't need to worry about any other races.
*/
if (i915_reset_in_progress(error) && !i915_terminally_wedged(error)) {
DRM_DEBUG_DRIVER("resetting chip\n");
kobject_uevent_env(&dev->primary->kdev->kobj, KOBJ_CHANGE,
reset_event);
drm/i915: clear up wedged transitions We have two important transitions of the wedged state in the current code: - 0 -> 1: This means a hang has been detected, and signals to everyone that they please get of any locks, so that the reset work item can do its job. - 1 -> 0: The reset handler has completed. Now the last transition mixes up two states: "Reset completed and successful" and "Reset failed". To distinguish these two we do some tricks with the reset completion, but I simply could not convince myself that this doesn't race under odd circumstances. Hence split this up, and add a new terminal state indicating that the hw is gone for good. Also add explicit #defines for both states, update comments. v2: Split out the reset handling bugfix for the throttle ioctl. v3: s/tmp/wedged/ sugested by Chris Wilson. Also fixup up a rebase error which prevented this patch from actually compiling. v4: To unify the wedged state with the reset counter, keep the reset-in-progress state just as a flag. The terminally-wedged state is now denoted with a big number. v5: Add a comment to the reset_counter special values explaining that WEDGED & RESET_IN_PROGRESS needs to be true for the code to be correct. v6: Fixup logic errors introduced with the wedged+reset_counter unification. Since WEDGED implies reset-in-progress (in a way we're terminally stuck in the dead-but-reset-not-completed state), we need ensure that we check for this everywhere. The specific bug was in wait_for_error, which would simply have timed out. v7: Extract an inline i915_reset_in_progress helper to make the code more readable. Also annote the reset-in-progress case with an unlikely, to help the compiler optimize the fastpath. Do the same for the terminally wedged case with i915_terminally_wedged. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-11-16 00:17:22 +08:00
/*
* In most cases it's guaranteed that we get here with an RPM
* reference held, for example because there is a pending GPU
* request that won't finish until the reset is done. This
* isn't the case at least when we get here by doing a
* simulated reset via debugs, so get an RPM reference.
*/
intel_runtime_pm_get(dev_priv);
intel_prepare_reset(dev);
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
/*
* All state reset _must_ be completed before we update the
* reset counter, for otherwise waiters might miss the reset
* pending state and not properly drop locks, resulting in
* deadlocks with the reset work.
*/
drm/i915: create a race-free reset detection With the previous patch the state transition handling of the reset code itself is now (hopefully) race free and solid. But that still leaves out everyone else - with the various lock-free wait paths we have there's the possibility that the reset happens between the point where we read the seqno we should wait on and the actual wait. And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll happily wait for a seqno which will in all likelyhood never signal. In practice this is not a big problem since the X server gets constantly interrupted, and can then submit more work (hopefully) to unblock everyone else: As soon as a new seqno write lands, all waiters will unblock. But running the i-g-t reset testcase ZZ_hangman can expose this race, especially on slower hw with fewer cpu cores. Now looking forward to ARB_robustness and friends that's not the best possible behaviour, hence this patch adds a reset_counter to be able to detect any reset, even if a given thread never observed the in-progress state. The important part is to correctly order things: - The write side needs to increment the counter after any seqno gets reset. Hence we need to do that at the end of the reset work, and again wake everyone up. We also need to place a barrier in between any possible seqno changes and the counter increment, since any unlock operations only guarantee that nothing leaks out, but not that at later load operation gets moved ahead. - On the read side we need to ensure that no reset can sneak in and invalidate the seqno. In all cases we can use the one-sided barrier that unlock operations guarantee (of the lock protecting the respective seqno/ring pair) to ensure correct ordering. Hence it is sufficient to place the atomic read before the mutex/spin_unlock and no additional barriers are required. The end-result of all this is that we need to wake up everyone twice in a reset operation: - First, before the reset starts, to get any lockholders of the locks, so that the reset can proceed. - Second, after the reset is completed, to allow waiters to properly and reliably detect the reset condition and bail out. I admit that this entire reset_counter thing smells a bit like overkill, but I think it's justified since it makes it really explicit what the bail-out condition is. And we need a reset counter anyway to implement ARB_robustness, and imo with finer-grained locking on the horizont this is the most resilient scheme I could think of. v2: Drop spurious change in the wait_for_error EXIT_COND - we only need to wait until we leave the reset-in-progress wedged state. v3: Don't play tricks with barriers in the throttle ioctl, the spin_unlock is barrier enough. I've also considered using a little helper to grab the current reset_counter, but then decided that hiding the atomic_read isn't a great idea, since having it explicitly show up in the code is a nice remainder to reviews to check the memory barriers. v4: Add a comment to explain why we need to fall through in __wait_seqno in the end variable assignments. v5: Review from Damien: - s/smb/smp/ in a comment - don't increment the reset counter after we've set it to WEDGED. Now we (again) properly wedge the gpu when the reset fails. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-06 16:01:42 +08:00
ret = i915_reset(dev);
intel_finish_reset(dev);
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
intel_runtime_pm_put(dev_priv);
drm/i915: create a race-free reset detection With the previous patch the state transition handling of the reset code itself is now (hopefully) race free and solid. But that still leaves out everyone else - with the various lock-free wait paths we have there's the possibility that the reset happens between the point where we read the seqno we should wait on and the actual wait. And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll happily wait for a seqno which will in all likelyhood never signal. In practice this is not a big problem since the X server gets constantly interrupted, and can then submit more work (hopefully) to unblock everyone else: As soon as a new seqno write lands, all waiters will unblock. But running the i-g-t reset testcase ZZ_hangman can expose this race, especially on slower hw with fewer cpu cores. Now looking forward to ARB_robustness and friends that's not the best possible behaviour, hence this patch adds a reset_counter to be able to detect any reset, even if a given thread never observed the in-progress state. The important part is to correctly order things: - The write side needs to increment the counter after any seqno gets reset. Hence we need to do that at the end of the reset work, and again wake everyone up. We also need to place a barrier in between any possible seqno changes and the counter increment, since any unlock operations only guarantee that nothing leaks out, but not that at later load operation gets moved ahead. - On the read side we need to ensure that no reset can sneak in and invalidate the seqno. In all cases we can use the one-sided barrier that unlock operations guarantee (of the lock protecting the respective seqno/ring pair) to ensure correct ordering. Hence it is sufficient to place the atomic read before the mutex/spin_unlock and no additional barriers are required. The end-result of all this is that we need to wake up everyone twice in a reset operation: - First, before the reset starts, to get any lockholders of the locks, so that the reset can proceed. - Second, after the reset is completed, to allow waiters to properly and reliably detect the reset condition and bail out. I admit that this entire reset_counter thing smells a bit like overkill, but I think it's justified since it makes it really explicit what the bail-out condition is. And we need a reset counter anyway to implement ARB_robustness, and imo with finer-grained locking on the horizont this is the most resilient scheme I could think of. v2: Drop spurious change in the wait_for_error EXIT_COND - we only need to wait until we leave the reset-in-progress wedged state. v3: Don't play tricks with barriers in the throttle ioctl, the spin_unlock is barrier enough. I've also considered using a little helper to grab the current reset_counter, but then decided that hiding the atomic_read isn't a great idea, since having it explicitly show up in the code is a nice remainder to reviews to check the memory barriers. v4: Add a comment to explain why we need to fall through in __wait_seqno in the end variable assignments. v5: Review from Damien: - s/smb/smp/ in a comment - don't increment the reset counter after we've set it to WEDGED. Now we (again) properly wedge the gpu when the reset fails. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-06 16:01:42 +08:00
if (ret == 0) {
/*
* After all the gem state is reset, increment the reset
* counter and wake up everyone waiting for the reset to
* complete.
*
* Since unlock operations are a one-sided barrier only,
* we need to insert a barrier here to order any seqno
* updates before
* the counter increment.
*/
smp_mb__before_atomic();
drm/i915: create a race-free reset detection With the previous patch the state transition handling of the reset code itself is now (hopefully) race free and solid. But that still leaves out everyone else - with the various lock-free wait paths we have there's the possibility that the reset happens between the point where we read the seqno we should wait on and the actual wait. And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll happily wait for a seqno which will in all likelyhood never signal. In practice this is not a big problem since the X server gets constantly interrupted, and can then submit more work (hopefully) to unblock everyone else: As soon as a new seqno write lands, all waiters will unblock. But running the i-g-t reset testcase ZZ_hangman can expose this race, especially on slower hw with fewer cpu cores. Now looking forward to ARB_robustness and friends that's not the best possible behaviour, hence this patch adds a reset_counter to be able to detect any reset, even if a given thread never observed the in-progress state. The important part is to correctly order things: - The write side needs to increment the counter after any seqno gets reset. Hence we need to do that at the end of the reset work, and again wake everyone up. We also need to place a barrier in between any possible seqno changes and the counter increment, since any unlock operations only guarantee that nothing leaks out, but not that at later load operation gets moved ahead. - On the read side we need to ensure that no reset can sneak in and invalidate the seqno. In all cases we can use the one-sided barrier that unlock operations guarantee (of the lock protecting the respective seqno/ring pair) to ensure correct ordering. Hence it is sufficient to place the atomic read before the mutex/spin_unlock and no additional barriers are required. The end-result of all this is that we need to wake up everyone twice in a reset operation: - First, before the reset starts, to get any lockholders of the locks, so that the reset can proceed. - Second, after the reset is completed, to allow waiters to properly and reliably detect the reset condition and bail out. I admit that this entire reset_counter thing smells a bit like overkill, but I think it's justified since it makes it really explicit what the bail-out condition is. And we need a reset counter anyway to implement ARB_robustness, and imo with finer-grained locking on the horizont this is the most resilient scheme I could think of. v2: Drop spurious change in the wait_for_error EXIT_COND - we only need to wait until we leave the reset-in-progress wedged state. v3: Don't play tricks with barriers in the throttle ioctl, the spin_unlock is barrier enough. I've also considered using a little helper to grab the current reset_counter, but then decided that hiding the atomic_read isn't a great idea, since having it explicitly show up in the code is a nice remainder to reviews to check the memory barriers. v4: Add a comment to explain why we need to fall through in __wait_seqno in the end variable assignments. v5: Review from Damien: - s/smb/smp/ in a comment - don't increment the reset counter after we've set it to WEDGED. Now we (again) properly wedge the gpu when the reset fails. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-06 16:01:42 +08:00
atomic_inc(&dev_priv->gpu_error.reset_counter);
kobject_uevent_env(&dev->primary->kdev->kobj,
drm/i915: create a race-free reset detection With the previous patch the state transition handling of the reset code itself is now (hopefully) race free and solid. But that still leaves out everyone else - with the various lock-free wait paths we have there's the possibility that the reset happens between the point where we read the seqno we should wait on and the actual wait. And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll happily wait for a seqno which will in all likelyhood never signal. In practice this is not a big problem since the X server gets constantly interrupted, and can then submit more work (hopefully) to unblock everyone else: As soon as a new seqno write lands, all waiters will unblock. But running the i-g-t reset testcase ZZ_hangman can expose this race, especially on slower hw with fewer cpu cores. Now looking forward to ARB_robustness and friends that's not the best possible behaviour, hence this patch adds a reset_counter to be able to detect any reset, even if a given thread never observed the in-progress state. The important part is to correctly order things: - The write side needs to increment the counter after any seqno gets reset. Hence we need to do that at the end of the reset work, and again wake everyone up. We also need to place a barrier in between any possible seqno changes and the counter increment, since any unlock operations only guarantee that nothing leaks out, but not that at later load operation gets moved ahead. - On the read side we need to ensure that no reset can sneak in and invalidate the seqno. In all cases we can use the one-sided barrier that unlock operations guarantee (of the lock protecting the respective seqno/ring pair) to ensure correct ordering. Hence it is sufficient to place the atomic read before the mutex/spin_unlock and no additional barriers are required. The end-result of all this is that we need to wake up everyone twice in a reset operation: - First, before the reset starts, to get any lockholders of the locks, so that the reset can proceed. - Second, after the reset is completed, to allow waiters to properly and reliably detect the reset condition and bail out. I admit that this entire reset_counter thing smells a bit like overkill, but I think it's justified since it makes it really explicit what the bail-out condition is. And we need a reset counter anyway to implement ARB_robustness, and imo with finer-grained locking on the horizont this is the most resilient scheme I could think of. v2: Drop spurious change in the wait_for_error EXIT_COND - we only need to wait until we leave the reset-in-progress wedged state. v3: Don't play tricks with barriers in the throttle ioctl, the spin_unlock is barrier enough. I've also considered using a little helper to grab the current reset_counter, but then decided that hiding the atomic_read isn't a great idea, since having it explicitly show up in the code is a nice remainder to reviews to check the memory barriers. v4: Add a comment to explain why we need to fall through in __wait_seqno in the end variable assignments. v5: Review from Damien: - s/smb/smp/ in a comment - don't increment the reset counter after we've set it to WEDGED. Now we (again) properly wedge the gpu when the reset fails. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-06 16:01:42 +08:00
KOBJ_CHANGE, reset_done_event);
drm/i915: clear up wedged transitions We have two important transitions of the wedged state in the current code: - 0 -> 1: This means a hang has been detected, and signals to everyone that they please get of any locks, so that the reset work item can do its job. - 1 -> 0: The reset handler has completed. Now the last transition mixes up two states: "Reset completed and successful" and "Reset failed". To distinguish these two we do some tricks with the reset completion, but I simply could not convince myself that this doesn't race under odd circumstances. Hence split this up, and add a new terminal state indicating that the hw is gone for good. Also add explicit #defines for both states, update comments. v2: Split out the reset handling bugfix for the throttle ioctl. v3: s/tmp/wedged/ sugested by Chris Wilson. Also fixup up a rebase error which prevented this patch from actually compiling. v4: To unify the wedged state with the reset counter, keep the reset-in-progress state just as a flag. The terminally-wedged state is now denoted with a big number. v5: Add a comment to the reset_counter special values explaining that WEDGED & RESET_IN_PROGRESS needs to be true for the code to be correct. v6: Fixup logic errors introduced with the wedged+reset_counter unification. Since WEDGED implies reset-in-progress (in a way we're terminally stuck in the dead-but-reset-not-completed state), we need ensure that we check for this everywhere. The specific bug was in wait_for_error, which would simply have timed out. v7: Extract an inline i915_reset_in_progress helper to make the code more readable. Also annote the reset-in-progress case with an unlikely, to help the compiler optimize the fastpath. Do the same for the terminally wedged case with i915_terminally_wedged. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-11-16 00:17:22 +08:00
} else {
atomic_or(I915_WEDGED, &error->reset_counter);
}
drm/i915: clear up wedged transitions We have two important transitions of the wedged state in the current code: - 0 -> 1: This means a hang has been detected, and signals to everyone that they please get of any locks, so that the reset work item can do its job. - 1 -> 0: The reset handler has completed. Now the last transition mixes up two states: "Reset completed and successful" and "Reset failed". To distinguish these two we do some tricks with the reset completion, but I simply could not convince myself that this doesn't race under odd circumstances. Hence split this up, and add a new terminal state indicating that the hw is gone for good. Also add explicit #defines for both states, update comments. v2: Split out the reset handling bugfix for the throttle ioctl. v3: s/tmp/wedged/ sugested by Chris Wilson. Also fixup up a rebase error which prevented this patch from actually compiling. v4: To unify the wedged state with the reset counter, keep the reset-in-progress state just as a flag. The terminally-wedged state is now denoted with a big number. v5: Add a comment to the reset_counter special values explaining that WEDGED & RESET_IN_PROGRESS needs to be true for the code to be correct. v6: Fixup logic errors introduced with the wedged+reset_counter unification. Since WEDGED implies reset-in-progress (in a way we're terminally stuck in the dead-but-reset-not-completed state), we need ensure that we check for this everywhere. The specific bug was in wait_for_error, which would simply have timed out. v7: Extract an inline i915_reset_in_progress helper to make the code more readable. Also annote the reset-in-progress case with an unlikely, to help the compiler optimize the fastpath. Do the same for the terminally wedged case with i915_terminally_wedged. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-11-16 00:17:22 +08:00
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
/*
* Note: The wake_up also serves as a memory barrier so that
* waiters see the update value of the reset counter atomic_t.
*/
i915_error_wake_up(dev_priv, true);
}
}
static void i915_report_and_clear_eir(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
uint32_t instdone[I915_NUM_INSTDONE_REG];
u32 eir = I915_READ(EIR);
int pipe, i;
if (!eir)
return;
pr_err("render error detected, EIR: 0x%08x\n", eir);
i915_get_extra_instdone(dev, instdone);
if (IS_G4X(dev)) {
if (eir & (GM45_ERROR_MEM_PRIV | GM45_ERROR_CP_PRIV)) {
u32 ipeir = I915_READ(IPEIR_I965);
pr_err(" IPEIR: 0x%08x\n", I915_READ(IPEIR_I965));
pr_err(" IPEHR: 0x%08x\n", I915_READ(IPEHR_I965));
for (i = 0; i < ARRAY_SIZE(instdone); i++)
pr_err(" INSTDONE_%d: 0x%08x\n", i, instdone[i]);
pr_err(" INSTPS: 0x%08x\n", I915_READ(INSTPS));
pr_err(" ACTHD: 0x%08x\n", I915_READ(ACTHD_I965));
I915_WRITE(IPEIR_I965, ipeir);
POSTING_READ(IPEIR_I965);
}
if (eir & GM45_ERROR_PAGE_TABLE) {
u32 pgtbl_err = I915_READ(PGTBL_ER);
pr_err("page table error\n");
pr_err(" PGTBL_ER: 0x%08x\n", pgtbl_err);
I915_WRITE(PGTBL_ER, pgtbl_err);
POSTING_READ(PGTBL_ER);
}
}
if (!IS_GEN2(dev)) {
if (eir & I915_ERROR_PAGE_TABLE) {
u32 pgtbl_err = I915_READ(PGTBL_ER);
pr_err("page table error\n");
pr_err(" PGTBL_ER: 0x%08x\n", pgtbl_err);
I915_WRITE(PGTBL_ER, pgtbl_err);
POSTING_READ(PGTBL_ER);
}
}
if (eir & I915_ERROR_MEMORY_REFRESH) {
pr_err("memory refresh error:\n");
for_each_pipe(dev_priv, pipe)
pr_err("pipe %c stat: 0x%08x\n",
pipe_name(pipe), I915_READ(PIPESTAT(pipe)));
/* pipestat has already been acked */
}
if (eir & I915_ERROR_INSTRUCTION) {
pr_err("instruction error\n");
pr_err(" INSTPM: 0x%08x\n", I915_READ(INSTPM));
for (i = 0; i < ARRAY_SIZE(instdone); i++)
pr_err(" INSTDONE_%d: 0x%08x\n", i, instdone[i]);
if (INTEL_INFO(dev)->gen < 4) {
u32 ipeir = I915_READ(IPEIR);
pr_err(" IPEIR: 0x%08x\n", I915_READ(IPEIR));
pr_err(" IPEHR: 0x%08x\n", I915_READ(IPEHR));
pr_err(" ACTHD: 0x%08x\n", I915_READ(ACTHD));
I915_WRITE(IPEIR, ipeir);
POSTING_READ(IPEIR);
} else {
u32 ipeir = I915_READ(IPEIR_I965);
pr_err(" IPEIR: 0x%08x\n", I915_READ(IPEIR_I965));
pr_err(" IPEHR: 0x%08x\n", I915_READ(IPEHR_I965));
pr_err(" INSTPS: 0x%08x\n", I915_READ(INSTPS));
pr_err(" ACTHD: 0x%08x\n", I915_READ(ACTHD_I965));
I915_WRITE(IPEIR_I965, ipeir);
POSTING_READ(IPEIR_I965);
}
}
I915_WRITE(EIR, eir);
POSTING_READ(EIR);
eir = I915_READ(EIR);
if (eir) {
/*
* some errors might have become stuck,
* mask them.
*/
DRM_ERROR("EIR stuck: 0x%08x, masking\n", eir);
I915_WRITE(EMR, I915_READ(EMR) | eir);
I915_WRITE(IIR, I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT);
}
}
/**
* i915_handle_error - handle a gpu error
* @dev: drm device
* @engine_mask: mask representing engines that are hung
* Do some basic checking of register state at error time and
* dump it to the syslog. Also call i915_capture_error_state() to make
* sure we get a record and make it available in debugfs. Fire a uevent
* so userspace knows something bad happened (should trigger collection
* of a ring dump etc.).
*/
void i915_handle_error(struct drm_device *dev, u32 engine_mask,
const char *fmt, ...)
{
struct drm_i915_private *dev_priv = dev->dev_private;
va_list args;
char error_msg[80];
va_start(args, fmt);
vscnprintf(error_msg, sizeof(error_msg), fmt, args);
va_end(args);
i915_capture_error_state(dev, engine_mask, error_msg);
i915_report_and_clear_eir(dev);
if (engine_mask) {
atomic_or(I915_RESET_IN_PROGRESS_FLAG,
drm/i915: create a race-free reset detection With the previous patch the state transition handling of the reset code itself is now (hopefully) race free and solid. But that still leaves out everyone else - with the various lock-free wait paths we have there's the possibility that the reset happens between the point where we read the seqno we should wait on and the actual wait. And if __wait_seqno then never sees the RESET_IN_PROGRESS state, we'll happily wait for a seqno which will in all likelyhood never signal. In practice this is not a big problem since the X server gets constantly interrupted, and can then submit more work (hopefully) to unblock everyone else: As soon as a new seqno write lands, all waiters will unblock. But running the i-g-t reset testcase ZZ_hangman can expose this race, especially on slower hw with fewer cpu cores. Now looking forward to ARB_robustness and friends that's not the best possible behaviour, hence this patch adds a reset_counter to be able to detect any reset, even if a given thread never observed the in-progress state. The important part is to correctly order things: - The write side needs to increment the counter after any seqno gets reset. Hence we need to do that at the end of the reset work, and again wake everyone up. We also need to place a barrier in between any possible seqno changes and the counter increment, since any unlock operations only guarantee that nothing leaks out, but not that at later load operation gets moved ahead. - On the read side we need to ensure that no reset can sneak in and invalidate the seqno. In all cases we can use the one-sided barrier that unlock operations guarantee (of the lock protecting the respective seqno/ring pair) to ensure correct ordering. Hence it is sufficient to place the atomic read before the mutex/spin_unlock and no additional barriers are required. The end-result of all this is that we need to wake up everyone twice in a reset operation: - First, before the reset starts, to get any lockholders of the locks, so that the reset can proceed. - Second, after the reset is completed, to allow waiters to properly and reliably detect the reset condition and bail out. I admit that this entire reset_counter thing smells a bit like overkill, but I think it's justified since it makes it really explicit what the bail-out condition is. And we need a reset counter anyway to implement ARB_robustness, and imo with finer-grained locking on the horizont this is the most resilient scheme I could think of. v2: Drop spurious change in the wait_for_error EXIT_COND - we only need to wait until we leave the reset-in-progress wedged state. v3: Don't play tricks with barriers in the throttle ioctl, the spin_unlock is barrier enough. I've also considered using a little helper to grab the current reset_counter, but then decided that hiding the atomic_read isn't a great idea, since having it explicitly show up in the code is a nice remainder to reviews to check the memory barriers. v4: Add a comment to explain why we need to fall through in __wait_seqno in the end variable assignments. v5: Review from Damien: - s/smb/smp/ in a comment - don't increment the reset counter after we've set it to WEDGED. Now we (again) properly wedge the gpu when the reset fails. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-06 16:01:42 +08:00
&dev_priv->gpu_error.reset_counter);
/*
* Wakeup waiting processes so that the reset function
* i915_reset_and_wakeup doesn't deadlock trying to grab
* various locks. By bumping the reset counter first, the woken
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
* processes will see a reset in progress and back off,
* releasing their locks and then wait for the reset completion.
* We must do this for _all_ gpu waiters that might hold locks
* that the reset work needs to acquire.
*
* Note: The wake_up serves as the required memory barrier to
* ensure that the waiters see the updated value of the reset
* counter atomic_t.
*/
drm/i915: fix wait_for_pending_flips vs gpu hang deadlock My g33 here seems to be shockingly good at hitting them all. This time around kms_flip/flip-vs-panning-vs-hang blows up: intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and if a gpu hang is pending aborts the wait for outstanding flips so that the setcrtc call will succeed and release the crtc mutex. And the gpu hang handler needs that lock in intel_display_handle_reset to be able to complete outstanding flips. The problem is that we can race in two ways: - Waiters on the dev_priv->pending_flip_queue aren't woken up after we've the reset as pending, but before we actually start the reset work. This means that the waiter doesn't notice the pending reset and hence will keep on hogging the locks. Like with dev->struct_mutex and the ring->irq_queue wait queues we there need to wake up everyone that potentially holds a lock which the reset handler needs. - intel_display_handle_reset was called _after_ we've already signalled the completion of the reset work. Which means a waiter could sneak in, grab the lock and never release it (since the pageflips won't ever get released). Similar to resetting the gem state all the reset work must complete before we update the reset counter. Contrary to the gem reset we don't need to have a second explicit wake up call since that will have happened already when completing the pageflips. We also don't have any issues that the completion happens while the reset state is still pending - wait_for_pending_flips is only there to ensure we display the right frame. After a gpu hang&reset events such guarantees are out the window anyway. This is in contrast to the gem code where too-early wake-up would result in unnecessary restarting of ioctls. Also, since we've gotten these various deadlocks and ordering constraints wrong so often throw copious amounts of comments at the code. This deadlock regression has been introduced in the commit which added the pageflip reset logic to the gpu hang work: commit 96a02917a0131e52efefde49c2784c0421d6c439 Author: Ville Syrjälä <ville.syrjala@linux.intel.com> Date: Mon Feb 18 19:08:49 2013 +0200 drm/i915: Finish page flips and update primary planes after a GPU reset v2: - Add comments to explain how the wake_up serves as memory barriers for the atomic_t reset counter. - Improve the comments a bit as suggested by Chris Wilson. - Extract the wake_up calls before/after the reset into a little i915_error_wake_up and unconditionally wake up the pending_flip_queue waiters, again as suggested by Chris Wilson. v3: Throw copious amounts of comments at i915_error_wake_up as suggested by Chris Wilson. Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-09-09 03:57:13 +08:00
i915_error_wake_up(dev_priv, false);
}
i915_reset_and_wakeup(dev);
}
/* Called from drm generic code, passed 'crtc' which
* we use as a pipe index
*/
static int i915_enable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
if (INTEL_INFO(dev)->gen >= 4)
i915_enable_pipestat(dev_priv, pipe,
PIPE_START_VBLANK_INTERRUPT_STATUS);
else
i915_enable_pipestat(dev_priv, pipe,
PIPE_VBLANK_INTERRUPT_STATUS);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
return 0;
}
static int ironlake_enable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
uint32_t bit = (INTEL_INFO(dev)->gen >= 7) ? DE_PIPE_VBLANK_IVB(pipe) :
DE_PIPE_VBLANK(pipe);
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
ilk_enable_display_irq(dev_priv, bit);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
return 0;
}
static int valleyview_enable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
i915_enable_pipestat(dev_priv, pipe,
PIPE_START_VBLANK_INTERRUPT_STATUS);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
return 0;
}
static int gen8_enable_vblank(struct drm_device *dev, unsigned int pipe)
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
bdw_enable_pipe_irq(dev_priv, pipe, GEN8_PIPE_VBLANK);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
return 0;
}
/* Called from drm generic code, passed 'crtc' which
* we use as a pipe index
*/
static void i915_disable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
i915_disable_pipestat(dev_priv, pipe,
PIPE_VBLANK_INTERRUPT_STATUS |
PIPE_START_VBLANK_INTERRUPT_STATUS);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
}
static void ironlake_disable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
uint32_t bit = (INTEL_INFO(dev)->gen >= 7) ? DE_PIPE_VBLANK_IVB(pipe) :
DE_PIPE_VBLANK(pipe);
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
ilk_disable_display_irq(dev_priv, bit);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
}
static void valleyview_disable_vblank(struct drm_device *dev, unsigned int pipe)
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
i915_disable_pipestat(dev_priv, pipe,
PIPE_START_VBLANK_INTERRUPT_STATUS);
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
}
static void gen8_disable_vblank(struct drm_device *dev, unsigned int pipe)
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
{
struct drm_i915_private *dev_priv = dev->dev_private;
unsigned long irqflags;
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
bdw_disable_pipe_irq(dev_priv, pipe, GEN8_PIPE_VBLANK);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
}
static bool
ring_idle(struct intel_engine_cs *engine, u32 seqno)
{
return i915_seqno_passed(seqno,
READ_ONCE(engine->last_submitted_seqno));
}
static bool
ipehr_is_semaphore_wait(struct drm_device *dev, u32 ipehr)
{
if (INTEL_INFO(dev)->gen >= 8) {
return (ipehr >> 23) == 0x1c;
} else {
ipehr &= ~MI_SEMAPHORE_SYNC_MASK;
return ipehr == (MI_SEMAPHORE_MBOX | MI_SEMAPHORE_COMPARE |
MI_SEMAPHORE_REGISTER);
}
}
static struct intel_engine_cs *
semaphore_wait_to_signaller_ring(struct intel_engine_cs *engine, u32 ipehr,
u64 offset)
{
struct drm_i915_private *dev_priv = engine->dev->dev_private;
struct intel_engine_cs *signaller;
if (INTEL_INFO(dev_priv)->gen >= 8) {
for_each_engine(signaller, dev_priv) {
if (engine == signaller)
continue;
if (offset == signaller->semaphore.signal_ggtt[engine->id])
return signaller;
}
} else {
u32 sync_bits = ipehr & MI_SEMAPHORE_SYNC_MASK;
for_each_engine(signaller, dev_priv) {
if(engine == signaller)
continue;
if (sync_bits == signaller->semaphore.mbox.wait[engine->id])
return signaller;
}
}
DRM_ERROR("No signaller ring found for ring %i, ipehr 0x%08x, offset 0x%016llx\n",
engine->id, ipehr, offset);
return NULL;
}
static struct intel_engine_cs *
semaphore_waits_for(struct intel_engine_cs *engine, u32 *seqno)
{
struct drm_i915_private *dev_priv = engine->dev->dev_private;
u32 cmd, ipehr, head;
u64 offset = 0;
int i, backwards;
drm/i915: Early exit from semaphore_waits_for for execlist mode. When submitting semaphores in execlist mode the hang checker crashes in this function because it is only runnable in ring submission mode. The reason this is of particular interest to the TDR patch series is because we use semaphores as a mean to induce hangs during testing (which is the recommended way to induce hangs for gen8+). It's not clear how this is supposed to work in execlist mode since: 1. This function requires a ring buffer. 2. Retrieving a ring buffer in execlist mode requires us to retrieve the corresponding context, which we get from a request. 3. Retieving a request from the hang checker is not straight-forward since that requires us to grab the struct_mutex in order to synchronize against the request retirement thread. 4. Grabbing the struct_mutex from the hang checker is nothing that we will do since that puts us at risk of deadlock since a hung thread might be holding the struct_mutex already. Therefore it's not obvious how we're supposed to deal with this. For now, we're doing an early exit from this function, which avoids any kernel panic situation when running our own internal TDR ULT. * v2: (Chris Wilson) Turned the execlist mode check into a ringbuffer NULL check to make it more submission mode agnostic and less of a layering violation. Signed-off-by: Tomas Elf <tomas.elf@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-10-09 02:31:33 +08:00
/*
* This function does not support execlist mode - any attempt to
* proceed further into this function will result in a kernel panic
* when dereferencing ring->buffer, which is not set up in execlist
* mode.
*
* The correct way of doing it would be to derive the currently
* executing ring buffer from the current context, which is derived
* from the currently running request. Unfortunately, to get the
* current request we would have to grab the struct_mutex before doing
* anything else, which would be ill-advised since some other thread
* might have grabbed it already and managed to hang itself, causing
* the hang checker to deadlock.
*
* Therefore, this function does not support execlist mode in its
* current form. Just return NULL and move on.
*/
if (engine->buffer == NULL)
drm/i915: Early exit from semaphore_waits_for for execlist mode. When submitting semaphores in execlist mode the hang checker crashes in this function because it is only runnable in ring submission mode. The reason this is of particular interest to the TDR patch series is because we use semaphores as a mean to induce hangs during testing (which is the recommended way to induce hangs for gen8+). It's not clear how this is supposed to work in execlist mode since: 1. This function requires a ring buffer. 2. Retrieving a ring buffer in execlist mode requires us to retrieve the corresponding context, which we get from a request. 3. Retieving a request from the hang checker is not straight-forward since that requires us to grab the struct_mutex in order to synchronize against the request retirement thread. 4. Grabbing the struct_mutex from the hang checker is nothing that we will do since that puts us at risk of deadlock since a hung thread might be holding the struct_mutex already. Therefore it's not obvious how we're supposed to deal with this. For now, we're doing an early exit from this function, which avoids any kernel panic situation when running our own internal TDR ULT. * v2: (Chris Wilson) Turned the execlist mode check into a ringbuffer NULL check to make it more submission mode agnostic and less of a layering violation. Signed-off-by: Tomas Elf <tomas.elf@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-10-09 02:31:33 +08:00
return NULL;
ipehr = I915_READ(RING_IPEHR(engine->mmio_base));
if (!ipehr_is_semaphore_wait(engine->dev, ipehr))
return NULL;
/*
* HEAD is likely pointing to the dword after the actual command,
* so scan backwards until we find the MBOX. But limit it to just 3
* or 4 dwords depending on the semaphore wait command size.
* Note that we don't care about ACTHD here since that might
* point at at batch, and semaphores are always emitted into the
* ringbuffer itself.
*/
head = I915_READ_HEAD(engine) & HEAD_ADDR;
backwards = (INTEL_INFO(engine->dev)->gen >= 8) ? 5 : 4;
for (i = backwards; i; --i) {
/*
* Be paranoid and presume the hw has gone off into the wild -
* our ring is smaller than what the hardware (and hence
* HEAD_ADDR) allows. Also handles wrap-around.
*/
head &= engine->buffer->size - 1;
/* This here seems to blow up */
cmd = ioread32(engine->buffer->virtual_start + head);
if (cmd == ipehr)
break;
head -= 4;
}
if (!i)
return NULL;
*seqno = ioread32(engine->buffer->virtual_start + head + 4) + 1;
if (INTEL_INFO(engine->dev)->gen >= 8) {
offset = ioread32(engine->buffer->virtual_start + head + 12);
offset <<= 32;
offset = ioread32(engine->buffer->virtual_start + head + 8);
}
return semaphore_wait_to_signaller_ring(engine, ipehr, offset);
}
static int semaphore_passed(struct intel_engine_cs *engine)
{
struct drm_i915_private *dev_priv = engine->dev->dev_private;
struct intel_engine_cs *signaller;
u32 seqno;
engine->hangcheck.deadlock++;
signaller = semaphore_waits_for(engine, &seqno);
if (signaller == NULL)
return -1;
/* Prevent pathological recursion due to driver bugs */
if (signaller->hangcheck.deadlock >= I915_NUM_ENGINES)
return -1;
if (i915_seqno_passed(signaller->get_seqno(signaller), seqno))
return 1;
/* cursory check for an unkickable deadlock */
if (I915_READ_CTL(signaller) & RING_WAIT_SEMAPHORE &&
semaphore_passed(signaller) < 0)
return -1;
return 0;
}
static void semaphore_clear_deadlocks(struct drm_i915_private *dev_priv)
{
struct intel_engine_cs *engine;
for_each_engine(engine, dev_priv)
engine->hangcheck.deadlock = 0;
}
static bool subunits_stuck(struct intel_engine_cs *engine)
{
u32 instdone[I915_NUM_INSTDONE_REG];
bool stuck;
int i;
if (engine->id != RCS)
return true;
i915_get_extra_instdone(engine->dev, instdone);
/* There might be unstable subunit states even when
* actual head is not moving. Filter out the unstable ones by
* accumulating the undone -> done transitions and only
* consider those as progress.
*/
stuck = true;
for (i = 0; i < I915_NUM_INSTDONE_REG; i++) {
const u32 tmp = instdone[i] | engine->hangcheck.instdone[i];
if (tmp != engine->hangcheck.instdone[i])
stuck = false;
engine->hangcheck.instdone[i] |= tmp;
}
return stuck;
}
static enum intel_ring_hangcheck_action
head_stuck(struct intel_engine_cs *engine, u64 acthd)
{
if (acthd != engine->hangcheck.acthd) {
/* Clear subunit states on head movement */
memset(engine->hangcheck.instdone, 0,
sizeof(engine->hangcheck.instdone));
return HANGCHECK_ACTIVE;
}
if (!subunits_stuck(engine))
return HANGCHECK_ACTIVE;
return HANGCHECK_HUNG;
}
static enum intel_ring_hangcheck_action
ring_stuck(struct intel_engine_cs *engine, u64 acthd)
{
struct drm_device *dev = engine->dev;
struct drm_i915_private *dev_priv = dev->dev_private;
enum intel_ring_hangcheck_action ha;
u32 tmp;
ha = head_stuck(engine, acthd);
if (ha != HANGCHECK_HUNG)
return ha;
if (IS_GEN2(dev))
return HANGCHECK_HUNG;
/* Is the chip hanging on a WAIT_FOR_EVENT?
* If so we can simply poke the RB_WAIT bit
* and break the hang. This should work on
* all but the second generation chipsets.
*/
tmp = I915_READ_CTL(engine);
if (tmp & RING_WAIT) {
i915_handle_error(dev, 0,
"Kicking stuck wait on %s",
engine->name);
I915_WRITE_CTL(engine, tmp);
return HANGCHECK_KICK;
}
if (INTEL_INFO(dev)->gen >= 6 && tmp & RING_WAIT_SEMAPHORE) {
switch (semaphore_passed(engine)) {
default:
return HANGCHECK_HUNG;
case 1:
i915_handle_error(dev, 0,
"Kicking stuck semaphore on %s",
engine->name);
I915_WRITE_CTL(engine, tmp);
return HANGCHECK_KICK;
case 0:
return HANGCHECK_WAIT;
}
}
return HANGCHECK_HUNG;
}
static unsigned kick_waiters(struct intel_engine_cs *engine)
{
struct drm_i915_private *i915 = to_i915(engine->dev);
unsigned user_interrupts = READ_ONCE(engine->user_interrupts);
if (engine->hangcheck.user_interrupts == user_interrupts &&
!test_and_set_bit(engine->id, &i915->gpu_error.missed_irq_rings)) {
if (!(i915->gpu_error.test_irq_rings & intel_engine_flag(engine)))
DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
engine->name);
else
DRM_INFO("Fake missed irq on %s\n",
engine->name);
wake_up_all(&engine->irq_queue);
}
return user_interrupts;
}
/*
* This is called when the chip hasn't reported back with completed
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
* batchbuffers in a long time. We keep track per ring seqno progress and
* if there are no progress, hangcheck score for that ring is increased.
* Further, acthd is inspected to see if the ring is stuck. On stuck case
* we kick the ring. If we see no progress on three subsequent calls
* we assume chip is wedged and try to fix it by resetting the chip.
*/
static void i915_hangcheck_elapsed(struct work_struct *work)
{
struct drm_i915_private *dev_priv =
container_of(work, typeof(*dev_priv),
gpu_error.hangcheck_work.work);
struct drm_device *dev = dev_priv->dev;
struct intel_engine_cs *engine;
enum intel_engine_id id;
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
int busy_count = 0, rings_hung = 0;
bool stuck[I915_NUM_ENGINES] = { 0 };
#define BUSY 1
#define KICK 5
#define HUNG 20
#define ACTIVE_DECAY 15
if (!i915.enable_hangcheck)
return;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/*
* The hangcheck work is synced during runtime suspend, we don't
* require a wakeref. TODO: instead of disabling the asserts make
* sure that we hold a reference when this work is running.
*/
DISABLE_RPM_WAKEREF_ASSERTS(dev_priv);
/* As enabling the GPU requires fairly extensive mmio access,
* periodically arm the mmio checker to see if we are triggering
* any invalid access.
*/
intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
for_each_engine_id(engine, dev_priv, id) {
u64 acthd;
u32 seqno;
unsigned user_interrupts;
bool busy = true;
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
semaphore_clear_deadlocks(dev_priv);
/* We don't strictly need an irq-barrier here, as we are not
* serving an interrupt request, be paranoid in case the
* barrier has side-effects (such as preventing a broken
* cacheline snoop) and so be sure that we can see the seqno
* advance. If the seqno should stick, due to a stale
* cacheline, we would erroneously declare the GPU hung.
*/
if (engine->irq_seqno_barrier)
engine->irq_seqno_barrier(engine);
acthd = intel_ring_get_active_head(engine);
seqno = engine->get_seqno(engine);
/* Reset stuck interrupts between batch advances */
user_interrupts = 0;
if (engine->hangcheck.seqno == seqno) {
if (ring_idle(engine, seqno)) {
engine->hangcheck.action = HANGCHECK_IDLE;
if (waitqueue_active(&engine->irq_queue)) {
/* Safeguard against driver failure */
user_interrupts = kick_waiters(engine);
engine->hangcheck.score += BUSY;
} else
busy = false;
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
} else {
/* We always increment the hangcheck score
* if the ring is busy and still processing
* the same request, so that no single request
* can run indefinitely (such as a chain of
* batches). The only time we do not increment
* the hangcheck score on this ring, if this
* ring is in a legitimate wait for another
* ring. In that case the waiting ring is a
* victim and we want to be sure we catch the
* right culprit. Then every time we do kick
* the ring, add a small increment to the
* score so that we can catch a batch that is
* being repeatedly kicked and so responsible
* for stalling the machine.
*/
engine->hangcheck.action = ring_stuck(engine,
acthd);
switch (engine->hangcheck.action) {
case HANGCHECK_IDLE:
case HANGCHECK_WAIT:
break;
case HANGCHECK_ACTIVE:
engine->hangcheck.score += BUSY;
break;
case HANGCHECK_KICK:
engine->hangcheck.score += KICK;
break;
case HANGCHECK_HUNG:
engine->hangcheck.score += HUNG;
stuck[id] = true;
break;
}
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
}
} else {
engine->hangcheck.action = HANGCHECK_ACTIVE;
/* Gradually reduce the count so that we catch DoS
* attempts across multiple batches.
*/
if (engine->hangcheck.score > 0)
engine->hangcheck.score -= ACTIVE_DECAY;
if (engine->hangcheck.score < 0)
engine->hangcheck.score = 0;
/* Clear head and subunit states on seqno movement */
acthd = 0;
memset(engine->hangcheck.instdone, 0,
sizeof(engine->hangcheck.instdone));
}
engine->hangcheck.seqno = seqno;
engine->hangcheck.acthd = acthd;
engine->hangcheck.user_interrupts = user_interrupts;
busy_count += busy;
}
for_each_engine_id(engine, dev_priv, id) {
if (engine->hangcheck.score >= HANGCHECK_SCORE_RING_HUNG) {
DRM_INFO("%s on %s\n",
stuck[id] ? "stuck" : "no progress",
engine->name);
rings_hung |= intel_engine_flag(engine);
}
}
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
if (rings_hung) {
i915_handle_error(dev, rings_hung, "Engine(s) hung");
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
goto out;
}
drm/i915: detect hang using per ring hangcheck_score Keep track of ring seqno progress and if there are no progress detected, declare hang. Use actual head (acthd) to distinguish between ring stuck and batchbuffer looping situation. Stuck ring will be kicked to trigger progress. This commit adds a hard limit for batchbuffer completion time. If batchbuffer completion time is more than 4.5 seconds, the gpu will be declared hung. Review comment from Ben which nicely clarifies the semantic change: "Maybe I'm just stating the functional changes of the patch, but in case they were unintended here is what I see as potential issues: 1. "If ring B is waiting on ring A via semaphore, and ring A is making progress, albeit slowly - the hangcheck will fire. The check will determine that A is moving, however ring B will appear hung because the ACTHD doesn't move. I honestly can't say if that's actually a realistic problem to hit it probably implies the timeout value is too low. 2. "There's also another corner case on the kick. If the seqno = 2 (though not stuck), and on the 3rd hangcheck, the ring is stuck, and we try to kick it... we don't actually try to find out if the kick helped" v2: use atchd to detect stuck ring from loop (Ben Widawsky) v3: Use acthd to check when ring needs kicking. Declare hang on third time in order to give time for kick_ring to take effect. v4: Update commit msg Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Paste in Ben's review comment.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-05-30 14:04:29 +08:00
if (busy_count)
/* Reset timer case chip hangs without another request
* being added */
i915_queue_hangcheck(dev);
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
out:
ENABLE_RPM_WAKEREF_ASSERTS(dev_priv);
}
void i915_queue_hangcheck(struct drm_device *dev)
{
struct i915_gpu_error *e = &to_i915(dev)->gpu_error;
if (!i915.enable_hangcheck)
return;
/* Don't continually defer the hangcheck so that it is always run at
* least once after work has been scheduled on any ring. Otherwise,
* we will ignore a hung ring if a second ring is kept busy.
*/
queue_delayed_work(e->hangcheck_wq, &e->hangcheck_work,
round_jiffies_up_relative(DRM_I915_HANGCHECK_JIFFIES));
}
static void ibx_irq_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (HAS_PCH_NOP(dev))
return;
GEN5_IRQ_RESET(SDE);
if (HAS_PCH_CPT(dev) || HAS_PCH_LPT(dev))
I915_WRITE(SERR_INT, 0xffffffff);
}
/*
* SDEIER is also touched by the interrupt handler to work around missed PCH
* interrupts. Hence we can't update it after the interrupt handler is enabled -
* instead we unconditionally enable all PCH interrupt sources here, but then
* only unmask them as needed with SDEIMR.
*
* This function needs to be called before interrupts are enabled.
*/
static void ibx_irq_pre_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (HAS_PCH_NOP(dev))
return;
WARN_ON(I915_READ(SDEIER) != 0);
I915_WRITE(SDEIER, 0xffffffff);
POSTING_READ(SDEIER);
}
static void gen5_gt_irq_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
GEN5_IRQ_RESET(GT);
if (INTEL_INFO(dev)->gen >= 6)
GEN5_IRQ_RESET(GEN6_PM);
}
/* drm_dma.h hooks
*/
static void ironlake_irq_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
I915_WRITE(HWSTAM, 0xffffffff);
GEN5_IRQ_RESET(DE);
if (IS_GEN7(dev))
I915_WRITE(GEN7_ERR_INT, 0xffffffff);
gen5_gt_irq_reset(dev);
ibx_irq_reset(dev);
}
static void vlv_display_irq_reset(struct drm_i915_private *dev_priv)
{
enum pipe pipe;
i915_hotplug_interrupt_update_locked(dev_priv, 0xffffffff, 0);
I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
for_each_pipe(dev_priv, pipe) {
I915_WRITE(PIPESTAT(pipe),
PIPE_FIFO_UNDERRUN_STATUS |
PIPESTAT_INT_STATUS_MASK);
dev_priv->pipestat_irq_mask[pipe] = 0;
}
GEN5_IRQ_RESET(VLV_);
dev_priv->irq_mask = ~0;
}
static void valleyview_irq_preinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
gen5_gt_irq_reset(dev);
I915_WRITE(DPINVGTT, DPINVGTT_STATUS_MASK);
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_reset(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
}
static void gen8_gt_irq_reset(struct drm_i915_private *dev_priv)
{
GEN8_IRQ_RESET_NDX(GT, 0);
GEN8_IRQ_RESET_NDX(GT, 1);
GEN8_IRQ_RESET_NDX(GT, 2);
GEN8_IRQ_RESET_NDX(GT, 3);
}
static void gen8_irq_reset(struct drm_device *dev)
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
I915_WRITE(GEN8_MASTER_IRQ, 0);
POSTING_READ(GEN8_MASTER_IRQ);
gen8_gt_irq_reset(dev_priv);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
for_each_pipe(dev_priv, pipe)
if (intel_display_power_is_enabled(dev_priv,
POWER_DOMAIN_PIPE(pipe)))
GEN8_IRQ_RESET_NDX(DE_PIPE, pipe);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
GEN5_IRQ_RESET(GEN8_DE_PORT_);
GEN5_IRQ_RESET(GEN8_DE_MISC_);
GEN5_IRQ_RESET(GEN8_PCU_);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
if (HAS_PCH_SPLIT(dev))
ibx_irq_reset(dev);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv,
unsigned int pipe_mask)
{
uint32_t extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN;
enum pipe pipe;
spin_lock_irq(&dev_priv->irq_lock);
for_each_pipe_masked(dev_priv, pipe, pipe_mask)
GEN8_IRQ_INIT_NDX(DE_PIPE, pipe,
dev_priv->de_irq_mask[pipe],
~dev_priv->de_irq_mask[pipe] | extra_ier);
spin_unlock_irq(&dev_priv->irq_lock);
}
void gen8_irq_power_well_pre_disable(struct drm_i915_private *dev_priv,
unsigned int pipe_mask)
{
enum pipe pipe;
spin_lock_irq(&dev_priv->irq_lock);
for_each_pipe_masked(dev_priv, pipe, pipe_mask)
GEN8_IRQ_RESET_NDX(DE_PIPE, pipe);
spin_unlock_irq(&dev_priv->irq_lock);
/* make sure we're done processing display irqs */
synchronize_irq(dev_priv->dev->irq);
}
static void cherryview_irq_preinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
I915_WRITE(GEN8_MASTER_IRQ, 0);
POSTING_READ(GEN8_MASTER_IRQ);
gen8_gt_irq_reset(dev_priv);
GEN5_IRQ_RESET(GEN8_PCU_);
I915_WRITE(DPINVGTT, DPINVGTT_STATUS_MASK_CHV);
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_reset(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
}
static u32 intel_hpd_enabled_irqs(struct drm_device *dev,
const u32 hpd[HPD_NUM_PINS])
{
struct drm_i915_private *dev_priv = to_i915(dev);
struct intel_encoder *encoder;
u32 enabled_irqs = 0;
for_each_intel_encoder(dev, encoder)
if (dev_priv->hotplug.stats[encoder->hpd_pin].state == HPD_ENABLED)
enabled_irqs |= hpd[encoder->hpd_pin];
return enabled_irqs;
}
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
static void ibx_hpd_irq_setup(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_irqs, hotplug, enabled_irqs;
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
if (HAS_PCH_IBX(dev)) {
hotplug_irqs = SDE_HOTPLUG_MASK;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_ibx);
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
} else {
hotplug_irqs = SDE_HOTPLUG_MASK_CPT;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_cpt);
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
}
ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
/*
* Enable digital hotplug on the PCH, and configure the DP short pulse
* duration to 2ms (which is the minimum in the Display Port spec).
* The pulse duration bits are reserved on LPT+.
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
*/
hotplug = I915_READ(PCH_PORT_HOTPLUG);
hotplug &= ~(PORTD_PULSE_DURATION_MASK|PORTC_PULSE_DURATION_MASK|PORTB_PULSE_DURATION_MASK);
hotplug |= PORTD_HOTPLUG_ENABLE | PORTD_PULSE_DURATION_2ms;
hotplug |= PORTC_HOTPLUG_ENABLE | PORTC_PULSE_DURATION_2ms;
hotplug |= PORTB_HOTPLUG_ENABLE | PORTB_PULSE_DURATION_2ms;
/*
* When CPU and PCH are on the same package, port A
* HPD must be enabled in both north and south.
*/
if (HAS_PCH_LPT_LP(dev))
hotplug |= PORTA_HOTPLUG_ENABLE;
I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
}
static void spt_hpd_irq_setup(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_irqs, hotplug, enabled_irqs;
hotplug_irqs = SDE_HOTPLUG_MASK_SPT;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_spt);
ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
/* Enable digital hotplug on the PCH */
hotplug = I915_READ(PCH_PORT_HOTPLUG);
hotplug |= PORTD_HOTPLUG_ENABLE | PORTC_HOTPLUG_ENABLE |
PORTB_HOTPLUG_ENABLE | PORTA_HOTPLUG_ENABLE;
I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
hotplug = I915_READ(PCH_PORT_HOTPLUG2);
hotplug |= PORTE_HOTPLUG_ENABLE;
I915_WRITE(PCH_PORT_HOTPLUG2, hotplug);
}
static void ilk_hpd_irq_setup(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_irqs, hotplug, enabled_irqs;
if (INTEL_INFO(dev)->gen >= 8) {
hotplug_irqs = GEN8_PORT_DP_A_HOTPLUG;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_bdw);
bdw_update_port_irq(dev_priv, hotplug_irqs, enabled_irqs);
} else if (INTEL_INFO(dev)->gen >= 7) {
hotplug_irqs = DE_DP_A_HOTPLUG_IVB;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_ivb);
ilk_update_display_irq(dev_priv, hotplug_irqs, enabled_irqs);
} else {
hotplug_irqs = DE_DP_A_HOTPLUG;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_ilk);
ilk_update_display_irq(dev_priv, hotplug_irqs, enabled_irqs);
}
/*
* Enable digital hotplug on the CPU, and configure the DP short pulse
* duration to 2ms (which is the minimum in the Display Port spec)
* The pulse duration bits are reserved on HSW+.
*/
hotplug = I915_READ(DIGITAL_PORT_HOTPLUG_CNTRL);
hotplug &= ~DIGITAL_PORTA_PULSE_DURATION_MASK;
hotplug |= DIGITAL_PORTA_HOTPLUG_ENABLE | DIGITAL_PORTA_PULSE_DURATION_2ms;
I915_WRITE(DIGITAL_PORT_HOTPLUG_CNTRL, hotplug);
ibx_hpd_irq_setup(dev);
}
static void bxt_hpd_irq_setup(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 hotplug_irqs, hotplug, enabled_irqs;
enabled_irqs = intel_hpd_enabled_irqs(dev, hpd_bxt);
hotplug_irqs = BXT_DE_PORT_HOTPLUG_MASK;
bdw_update_port_irq(dev_priv, hotplug_irqs, enabled_irqs);
hotplug = I915_READ(PCH_PORT_HOTPLUG);
hotplug |= PORTC_HOTPLUG_ENABLE | PORTB_HOTPLUG_ENABLE |
PORTA_HOTPLUG_ENABLE;
DRM_DEBUG_KMS("Invert bit setting: hp_ctl:%x hp_port:%x\n",
hotplug, enabled_irqs);
hotplug &= ~BXT_DDI_HPD_INVERT_MASK;
/*
* For BXT invert bit has to be set based on AOB design
* for HPD detection logic, update it based on VBT fields.
*/
if ((enabled_irqs & BXT_DE_PORT_HP_DDIA) &&
intel_bios_is_port_hpd_inverted(dev_priv, PORT_A))
hotplug |= BXT_DDIA_HPD_INVERT;
if ((enabled_irqs & BXT_DE_PORT_HP_DDIB) &&
intel_bios_is_port_hpd_inverted(dev_priv, PORT_B))
hotplug |= BXT_DDIB_HPD_INVERT;
if ((enabled_irqs & BXT_DE_PORT_HP_DDIC) &&
intel_bios_is_port_hpd_inverted(dev_priv, PORT_C))
hotplug |= BXT_DDIC_HPD_INVERT;
I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
}
static void ibx_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: implement ibx_hpd_irq_setup This fixes a regression introduced in commit e5868a318d1ae28f760f77bb91ce5deb751733fd Author: Egbert Eich <eich@suse.de> Date: Thu Feb 28 04:17:12 2013 -0500 DRM/i915: Convert HPD interrupts to make use of HPD pin assignment in encode Due to the irq setup rework in 3.9, see commit 20afbda209d708be66944907966486d0c1331cb8 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Tue Dec 11 14:05:07 2012 +0100 drm/i915: Fixup hpd irq register setup ordering Egbert Eich's hpd rework blows up on pch-split platforms - it walks the encoder list before that has been set up completely. The new init sequence is: 1. irq enabling 2. modeset init 3. hpd setup We need to move around the ibx setup a bit to fix this. Ville Syrjälä pointed out in his review that we can't touch SDEIER after the interrupt handler is set up, since that'll race with Paulo Zanoni's PCH interrupt race fix: commit 44498aea293b37af1d463acd9658cdce1ecdf427 Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Feb 22 17:05:28 2013 -0300 drm/i915: also disable south interrupts when handling them We fix that by unconditionally enabling all interrupts in SDEIER, but masking them as-needed in SDEIMR. Since only the single-threaded setup/teardown (or suspend/resume) code touches that, no further locking is required. While at it also simplify the mask handling - we start out with all interrupts cleared in the postinstall hook, and never enable a hpd interrupt before hpd_irq_setup is called. And finally, for consistency rename the ibx hpd setup function to ibx_hpd_irq_setup. v2: Fix race around SDEIER writes (Ville). v3: Remove the superflous posting read for SDEIER, spotted by Ville. Ville also wondered whether we shouldn't clear SDEIIR, since now SDE interrupts are enabled before we have an irq handler installed. But the master interrupt control bit in DEIER is still cleared, so we should be fine. Cc: Egbert Eich <eich@suse.de> Cc: Jesse Barnes <jbarnes@virtuousgeek.org> Cc: Paulo Zanoni <paulo.r.zanoni@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62798 Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-03-27 22:55:01 +08:00
u32 mask;
if (HAS_PCH_NOP(dev))
return;
if (HAS_PCH_IBX(dev))
drm/i915: Don't enable display error interrupts from the start We need to enable interrupt processing before all the modeset state is set up. But that means we can fall over when we get a pipe underrun. This shouldn't happen as long as the bios works correctly but as usual this turns out to be wishful thinking. So disable error interrupts at irq install time and rely on the re-enabling code in the modeset functions to take care of this. Note that due to the SDE interrupt handling race we must uncondtionally enable all interrupt sources in SDEIER, hence no need to enable the SERR bit specifically. On gmch platforms we don't have an explicit enable/mask bit for fifo underruns. Fixing this up would require a bit of software tracking, hence is material for a separate patch. To make this possible we need to switch all gmch platforms to the new pipestat interrupt handling scheme Imre implemented for vlv, and then also add a safe form of sw state checking to __cpu_fifo_underrun_reporting_enabled a bit. v2: Also handle the ilk/snb cpu fifo underrun bits accordingly. Spotted by Ville. v3: Also handle the south interrupt underrun bits on ibx. Again spotted by Ville. Reported-by: Rob Clark <robdclark@gmail.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: stable@vger.kernel.org Tested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2014-03-08 03:34:46 +08:00
mask = SDE_GMBUS | SDE_AUX_MASK | SDE_POISON;
else
drm/i915: Don't enable display error interrupts from the start We need to enable interrupt processing before all the modeset state is set up. But that means we can fall over when we get a pipe underrun. This shouldn't happen as long as the bios works correctly but as usual this turns out to be wishful thinking. So disable error interrupts at irq install time and rely on the re-enabling code in the modeset functions to take care of this. Note that due to the SDE interrupt handling race we must uncondtionally enable all interrupt sources in SDEIER, hence no need to enable the SERR bit specifically. On gmch platforms we don't have an explicit enable/mask bit for fifo underruns. Fixing this up would require a bit of software tracking, hence is material for a separate patch. To make this possible we need to switch all gmch platforms to the new pipestat interrupt handling scheme Imre implemented for vlv, and then also add a safe form of sw state checking to __cpu_fifo_underrun_reporting_enabled a bit. v2: Also handle the ilk/snb cpu fifo underrun bits accordingly. Spotted by Ville. v3: Also handle the south interrupt underrun bits on ibx. Again spotted by Ville. Reported-by: Rob Clark <robdclark@gmail.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: stable@vger.kernel.org Tested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2014-03-08 03:34:46 +08:00
mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT;
gen5_assert_iir_is_zero(dev_priv, SDEIIR);
I915_WRITE(SDEIMR, ~mask);
}
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
static void gen5_gt_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 pm_irqs, gt_irqs;
pm_irqs = gt_irqs = 0;
dev_priv->gt_irq_mask = ~0;
if (HAS_L3_DPF(dev)) {
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
/* L3 parity interrupt is always unmasked. */
dev_priv->gt_irq_mask = ~GT_PARITY_ERROR(dev);
gt_irqs |= GT_PARITY_ERROR(dev);
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
}
gt_irqs |= GT_RENDER_USER_INTERRUPT;
if (IS_GEN5(dev)) {
gt_irqs |= GT_RENDER_PIPECTL_NOTIFY_INTERRUPT |
ILK_BSD_USER_INTERRUPT;
} else {
gt_irqs |= GT_BLT_USER_INTERRUPT | GT_BSD_USER_INTERRUPT;
}
GEN5_IRQ_INIT(GT, dev_priv->gt_irq_mask, gt_irqs);
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
if (INTEL_INFO(dev)->gen >= 6) {
/*
* RPS interrupts will get enabled/disabled on demand when RPS
* itself is enabled/disabled.
*/
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
if (HAS_VEBOX(dev))
pm_irqs |= PM_VEBOX_USER_INTERRUPT;
dev_priv->pm_irq_mask = 0xffffffff;
GEN5_IRQ_INIT(GEN6_PM, dev_priv->pm_irq_mask, pm_irqs);
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
}
}
static int ironlake_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 display_mask, extra_mask;
if (INTEL_INFO(dev)->gen >= 7) {
display_mask = (DE_MASTER_IRQ_CONTROL | DE_GSE_IVB |
DE_PCH_EVENT_IVB | DE_PLANEC_FLIP_DONE_IVB |
DE_PLANEB_FLIP_DONE_IVB |
drm/i915: Don't enable display error interrupts from the start We need to enable interrupt processing before all the modeset state is set up. But that means we can fall over when we get a pipe underrun. This shouldn't happen as long as the bios works correctly but as usual this turns out to be wishful thinking. So disable error interrupts at irq install time and rely on the re-enabling code in the modeset functions to take care of this. Note that due to the SDE interrupt handling race we must uncondtionally enable all interrupt sources in SDEIER, hence no need to enable the SERR bit specifically. On gmch platforms we don't have an explicit enable/mask bit for fifo underruns. Fixing this up would require a bit of software tracking, hence is material for a separate patch. To make this possible we need to switch all gmch platforms to the new pipestat interrupt handling scheme Imre implemented for vlv, and then also add a safe form of sw state checking to __cpu_fifo_underrun_reporting_enabled a bit. v2: Also handle the ilk/snb cpu fifo underrun bits accordingly. Spotted by Ville. v3: Also handle the south interrupt underrun bits on ibx. Again spotted by Ville. Reported-by: Rob Clark <robdclark@gmail.com> Cc: Rob Clark <robdclark@gmail.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: stable@vger.kernel.org Tested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2014-03-08 03:34:46 +08:00
DE_PLANEA_FLIP_DONE_IVB | DE_AUX_CHANNEL_A_IVB);
extra_mask = (DE_PIPEC_VBLANK_IVB | DE_PIPEB_VBLANK_IVB |
DE_PIPEA_VBLANK_IVB | DE_ERR_INT_IVB |
DE_DP_A_HOTPLUG_IVB);
} else {
display_mask = (DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE |
DE_AUX_CHANNEL_A |
DE_PIPEB_CRC_DONE | DE_PIPEA_CRC_DONE |
DE_POISON);
extra_mask = (DE_PIPEA_VBLANK | DE_PIPEB_VBLANK | DE_PCU_EVENT |
DE_PIPEB_FIFO_UNDERRUN | DE_PIPEA_FIFO_UNDERRUN |
DE_DP_A_HOTPLUG);
}
dev_priv->irq_mask = ~display_mask;
I915_WRITE(HWSTAM, 0xeffe);
ibx_irq_pre_postinstall(dev);
GEN5_IRQ_INIT(DE, dev_priv->irq_mask, display_mask | extra_mask);
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
gen5_gt_irq_postinstall(dev);
ibx_irq_postinstall(dev);
if (IS_IRONLAKE_M(dev)) {
/* Enable PCU event interrupts
*
* spinlocking not required here for correctness since interrupt
* setup is guaranteed to run in single-threaded context. But we
* need it to make the assert_spin_locked happy. */
spin_lock_irq(&dev_priv->irq_lock);
ilk_enable_display_irq(dev_priv, DE_PCU_EVENT);
spin_unlock_irq(&dev_priv->irq_lock);
}
return 0;
}
static void vlv_display_irq_postinstall(struct drm_i915_private *dev_priv)
{
u32 pipestat_mask;
u32 iir_mask;
enum pipe pipe;
pipestat_mask = PIPESTAT_INT_STATUS_MASK |
PIPE_FIFO_UNDERRUN_STATUS;
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe), pipestat_mask);
POSTING_READ(PIPESTAT(PIPE_A));
pipestat_mask = PLANE_FLIP_DONE_INT_STATUS_VLV |
PIPE_CRC_DONE_INTERRUPT_STATUS;
i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
for_each_pipe(dev_priv, pipe)
i915_enable_pipestat(dev_priv, pipe, pipestat_mask);
iir_mask = I915_DISPLAY_PORT_INTERRUPT |
I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
if (IS_CHERRYVIEW(dev_priv))
iir_mask |= I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
dev_priv->irq_mask &= ~iir_mask;
I915_WRITE(VLV_IIR, iir_mask);
I915_WRITE(VLV_IIR, iir_mask);
I915_WRITE(VLV_IER, ~dev_priv->irq_mask);
I915_WRITE(VLV_IMR, dev_priv->irq_mask);
POSTING_READ(VLV_IMR);
}
void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv)
{
assert_spin_locked(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
return;
dev_priv->display_irqs_enabled = true;
if (intel_irqs_enabled(dev_priv))
vlv_display_irq_postinstall(dev_priv);
}
void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv)
{
assert_spin_locked(&dev_priv->irq_lock);
if (!dev_priv->display_irqs_enabled)
return;
dev_priv->display_irqs_enabled = false;
if (intel_irqs_enabled(dev_priv))
vlv_display_irq_reset(dev_priv);
}
static int valleyview_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: unify GT/PM irq postinstall code Again extract a common helper. For the postinstall hook things are a bit more complicated since we have more cases on ilk-hsw/vlv here. But since vlv was clearly broken by failing to initialize dev_priv->gt_irq_mask correctly the shared code is clearly justified. Also kill the PMIER setting in the async rps enable work. I should have been save, but also clearly looked rather fragile. PMIER setup is now all down in the irq pre/postinstall hooks. With this we now have the usual interrupt register sequence for GT/PM irq registers: - IER is setup once with all the interrupts we ever need in the postinstall hook and never touched again. Exceptions are SDEIER, which is touched in the preinstall hook (when the irq handler isn't enabled) and then only from the irq handler. And DEIER/VLV_IER with is used in the irq handler but also written to once in the postinstall hook. But since that write is essentially what enables the interrupt and we should always have MSI interrupts we should be save. In case we ever have non-MSI interrupts we'd be screwed. - IIR is cleared in the postinstall hook before we enable/unmask the respective interrupt sources. Hence we can't steal an interrupt event an accidentally trigger the spurious interrupt logic in the core kernel. Note that after some discussion with Ben Widawsky we think that we actually should clear the IIR registers in the preinstall hook. But doing that is a much larger patch series. - IMR regs are (usually) all masked off. Those are the only regs changed at runtime, which is all protected by dev_priv->irq_lock. This unification also kills the cargo-culted read-modify-write PM register setup for VECS. Interrupt setup is done without userspace being able to interfere, so we better know what values we want to put into those registers. RMW cycles otoh are really good at papering over races, until stuff magically blows up and no one has a clue why. v2: Touch the gen6+ PM interrupt registers only on gen6+. v3: Improve the commit message to more clearly spell out why we want to unify the code and what exactly changes. Cc: Ben Widawsky <ben@bwidawsk.net> Cc: Paulo Zanoni <przanoni@gmail.com> Reviewed-by: Ben Widawsky <ben@bwidawsk.net> [danvet: Add a comment to explain why the l3 parity interrupt is special.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-07-13 04:43:26 +08:00
gen5_gt_irq_postinstall(dev);
/* ack & enable invalid PTE error interrupts */
#if 0 /* FIXME: add support to irq handler for checking these bits */
I915_WRITE(DPINVGTT, DPINVGTT_STATUS_MASK);
I915_WRITE(DPINVGTT, DPINVGTT_EN_MASK);
#endif
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_postinstall(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
I915_WRITE(VLV_MASTER_IER, MASTER_INTERRUPT_ENABLE);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
return 0;
}
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
static void gen8_gt_irq_postinstall(struct drm_i915_private *dev_priv)
{
/* These are interrupts we'll toggle with the ring mask register */
uint32_t gt_interrupts[] = {
GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT |
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
GT_RENDER_L3_PARITY_ERROR_INTERRUPT |
GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT |
GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT,
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT |
GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT |
GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT,
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
0,
GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT |
GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
};
drm/i915/bdw: Implement a basic PM interrupt handler Almost all of it is reusable from the existing code. The primary difference is we need to do even less in the interrupt handler, since interrupts are not shared in the same way. The patch is mostly a copy-paste of the existing snb+ code, with updates to the relevant parts requiring changes to the interrupt handling. As such it /should/ be relatively trivial. It's highly likely that I missed some places where I need a gen8 version of the PM interrupts, but it has become invisible to me by now. This patch could probably be split into adding the new functions, followed by actually handling the interrupts. Since the code is currently disabled (and broken) I think the patch stands better by itself. v2: Move the commit about not touching the ringbuffer interrupt to the snb_* function where it belongs (Rodrigo) v3: Rebased on Paulo's runtime PM changes v4: Not well validated, but rebase on commit 730488b2eddded4497f63f70867b1256cd9e117c Author: Paulo Zanoni <paulo.r.zanoni@intel.com> Date: Fri Mar 7 20:12:32 2014 -0300 drm/i915: kill dev_priv->pm.regsave v5: Rebased on latest code base. (Deepak) v6: Remove conflict markers, Unnecessary empty line and use right IIR interrupt (Ville) v7: mask modified without rmw (Ville Syrjälä) Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Ben Widawsky <ben@bwidawsk.net> Signed-off-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-05-16 01:58:08 +08:00
dev_priv->pm_irq_mask = 0xffffffff;
GEN8_IRQ_INIT_NDX(GT, 0, ~gt_interrupts[0], gt_interrupts[0]);
GEN8_IRQ_INIT_NDX(GT, 1, ~gt_interrupts[1], gt_interrupts[1]);
/*
* RPS interrupts will get enabled/disabled on demand when RPS itself
* is enabled/disabled.
*/
GEN8_IRQ_INIT_NDX(GT, 2, dev_priv->pm_irq_mask, 0);
GEN8_IRQ_INIT_NDX(GT, 3, ~gt_interrupts[3], gt_interrupts[3]);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
{
uint32_t de_pipe_masked = GEN8_PIPE_CDCLK_CRC_DONE;
uint32_t de_pipe_enables;
u32 de_port_masked = GEN8_AUX_CHANNEL_A;
u32 de_port_enables;
enum pipe pipe;
if (INTEL_INFO(dev_priv)->gen >= 9) {
de_pipe_masked |= GEN9_PIPE_PLANE1_FLIP_DONE |
GEN9_DE_PIPE_IRQ_FAULT_ERRORS;
de_port_masked |= GEN9_AUX_CHANNEL_B | GEN9_AUX_CHANNEL_C |
GEN9_AUX_CHANNEL_D;
if (IS_BROXTON(dev_priv))
de_port_masked |= BXT_DE_PORT_GMBUS;
} else {
de_pipe_masked |= GEN8_PIPE_PRIMARY_FLIP_DONE |
GEN8_DE_PIPE_IRQ_FAULT_ERRORS;
}
de_pipe_enables = de_pipe_masked | GEN8_PIPE_VBLANK |
GEN8_PIPE_FIFO_UNDERRUN;
de_port_enables = de_port_masked;
if (IS_BROXTON(dev_priv))
de_port_enables |= BXT_DE_PORT_HOTPLUG_MASK;
else if (IS_BROADWELL(dev_priv))
de_port_enables |= GEN8_PORT_DP_A_HOTPLUG;
dev_priv->de_irq_mask[PIPE_A] = ~de_pipe_masked;
dev_priv->de_irq_mask[PIPE_B] = ~de_pipe_masked;
dev_priv->de_irq_mask[PIPE_C] = ~de_pipe_masked;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
for_each_pipe(dev_priv, pipe)
if (intel_display_power_is_enabled(dev_priv,
POWER_DOMAIN_PIPE(pipe)))
GEN8_IRQ_INIT_NDX(DE_PIPE, pipe,
dev_priv->de_irq_mask[pipe],
de_pipe_enables);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
GEN5_IRQ_INIT(GEN8_DE_PORT_, ~de_port_masked, de_port_enables);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
static int gen8_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (HAS_PCH_SPLIT(dev))
ibx_irq_pre_postinstall(dev);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
gen8_gt_irq_postinstall(dev_priv);
gen8_de_irq_postinstall(dev_priv);
if (HAS_PCH_SPLIT(dev))
ibx_irq_postinstall(dev);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
I915_WRITE(GEN8_MASTER_IRQ, DE_MASTER_IRQ_CONTROL);
POSTING_READ(GEN8_MASTER_IRQ);
return 0;
}
static int cherryview_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
gen8_gt_irq_postinstall(dev_priv);
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_postinstall(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
I915_WRITE(GEN8_MASTER_IRQ, MASTER_INTERRUPT_ENABLE);
POSTING_READ(GEN8_MASTER_IRQ);
return 0;
}
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
static void gen8_irq_uninstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv)
return;
gen8_irq_reset(dev);
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
}
static void valleyview_irq_uninstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv)
return;
I915_WRITE(VLV_MASTER_IER, 0);
gen5_gt_irq_reset(dev);
I915_WRITE(HWSTAM, 0xffffffff);
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_reset(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
}
static void cherryview_irq_uninstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv)
return;
I915_WRITE(GEN8_MASTER_IRQ, 0);
POSTING_READ(GEN8_MASTER_IRQ);
gen8_gt_irq_reset(dev_priv);
GEN5_IRQ_RESET(GEN8_PCU_);
spin_lock_irq(&dev_priv->irq_lock);
if (dev_priv->display_irqs_enabled)
vlv_display_irq_reset(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
}
static void ironlake_irq_uninstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (!dev_priv)
return;
ironlake_irq_reset(dev);
}
static void i8xx_irq_preinstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE16(IMR, 0xffff);
I915_WRITE16(IER, 0x0);
POSTING_READ16(IER);
}
static int i8xx_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
I915_WRITE16(EMR,
~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
/* Unmask the interrupts that we always want on. */
dev_priv->irq_mask =
~(I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT);
I915_WRITE16(IMR, dev_priv->irq_mask);
I915_WRITE16(IER,
I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
I915_USER_INTERRUPT);
POSTING_READ16(IER);
/* Interrupt setup is already guaranteed to be single-threaded, this is
* just to make the assert_spin_locked check happy. */
spin_lock_irq(&dev_priv->irq_lock);
i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
spin_unlock_irq(&dev_priv->irq_lock);
return 0;
}
/*
* Returns true when a page flip has completed.
*/
static bool i8xx_handle_vblank(struct drm_device *dev,
int plane, int pipe, u32 iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u16 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
drm/i915: Make sprite updates atomic Add a mechanism by which we can evade the leading edge of vblank. This guarantees that no two sprite register writes will straddle on either side of the vblank start, and that means all the writes will be latched together in one atomic operation. We do the vblank evade by checking the scanline counter, and if it's too close to the start of vblank (too close has been hardcoded to 100usec for now), we will wait for the vblank start to pass. In order to eliminate random delayes from the rest of the system, we operate with interrupts disabled, except when waiting for the vblank obviously. Note that we now go digging through pipe_to_crtc_mapping[] in the vblank interrupt handler, which is a bit dangerous since we set up interrupts before the crtcs. However in this case since it's the vblank interrupt, we don't actually unmask it until some piece of code requests it. v2: preempt_check_resched() calls after local_irq_enable() (Jesse) Hook up the vblank irq stuff on BDW as well v3: Pass intel_crtc instead of drm_crtc (Daniel) Warn if crtc.mutex isn't locked (Daniel) Add an explicit compiler barrier and document the barriers (Daniel) Note the irq vs. modeset setup madness in the commit message (Daniel) v4: Use prepare_to_wait() & co. directly and eliminate vbl_received v5: Refactor intel_pipe_handle_vblank() vs. drm_handle_vblank() (Chris) Check for min/max scanline <= 0 (Chris) Don't call intel_pipe_update_end() if start failed totally (Chris) Check that the vblank counters match on both sides of the critical section (Chris) v6: Fix atomic update for interlaced modes v7: Reorder code for better readability (Chris) v8: Drop preempt_check_resched(). It's not available to modules anymore and isn't even needed unless we ourselves cause a wakeup needing reschedule while interrupts are off Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:46 +08:00
if (!intel_pipe_handle_vblank(dev, pipe))
return false;
if ((iir & flip_pending) == 0)
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
goto check_page_flip;
/* We detect FlipDone by looking for the change in PendingFlip from '1'
* to '0' on the following vblank, i.e. IIR has the Pendingflip
* asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
* the flip is completed (no longer pending). Since this doesn't raise
* an interrupt per se, we watch for the change at vblank.
*/
if (I915_READ16(ISR) & flip_pending)
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
goto check_page_flip;
drm/i915: Don't call intel_prepare_page_flip() multiple times on gen2-4 The flip stall detector kicks in when pending>=INTEL_FLIP_COMPLETE. That means if we first call intel_prepare_page_flip() but don't call intel_finish_page_flip(), the next stall check will erroneosly think the page flip was somehow stuck. With enough debug spew emitted from the interrupt handler my 830 hangs when this happens. My theory is that the previous vblank interrupt gets sufficiently delayed that the handler will see the pending bit set in IIR, but ISR still has the bit set as well (ie. the flip was processed by CS but didn't complete yet). In this case the handler will proceed to call intel_check_page_flip() immediately after intel_prepare_page_flip(). It then tries to print a backtrace for the stuck flip WARN, which apparetly results in way too much debug spew delaying interrupt processing further. That then seems to cause an endless loop in the interrupt handler, and the machine is dead until the watchdog kicks in and reboots. At least limiting the number of iterations of the loop in the interrupt handler also prevented the hang. So it seems better to not call intel_prepare_page_flip() without immediately calling intel_finish_page_flip(). The IIR/ISR trickery avoids races here so this is a perfectly safe thing to do. v2: Fix typo in commit message (checkpatch) Cc: stable@vger.kernel.org Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88381 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85888 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2014-12-18 05:08:03 +08:00
intel_prepare_page_flip(dev, plane);
intel_finish_page_flip(dev, pipe);
return true;
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
check_page_flip:
intel_check_page_flip(dev, pipe);
return false;
}
static irqreturn_t i8xx_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u16 iir, new_iir;
u32 pipe_stats[2];
int pipe;
u16 flip_mask =
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
irqreturn_t ret;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
ret = IRQ_NONE;
iir = I915_READ16(IIR);
if (iir == 0)
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
goto out;
while (iir & ~flip_mask) {
/* Can't rely on pipestat interrupt bit in iir as it might
* have been cleared after the pipestat interrupt was received.
* It doesn't set the bit in iir again, but it still produces
* interrupts (for non-MSI).
*/
spin_lock(&dev_priv->irq_lock);
if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
for_each_pipe(dev_priv, pipe) {
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = PIPESTAT(pipe);
pipe_stats[pipe] = I915_READ(reg);
/*
* Clear the PIPE*STAT regs before the IIR
*/
if (pipe_stats[pipe] & 0x8000ffff)
I915_WRITE(reg, pipe_stats[pipe]);
}
spin_unlock(&dev_priv->irq_lock);
I915_WRITE16(IIR, iir & ~flip_mask);
new_iir = I915_READ16(IIR); /* Flush posted writes */
if (iir & I915_USER_INTERRUPT)
notify_ring(&dev_priv->engine[RCS]);
for_each_pipe(dev_priv, pipe) {
int plane = pipe;
if (HAS_FBC(dev))
plane = !plane;
if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
i8xx_handle_vblank(dev, plane, pipe, iir))
flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
i9xx_pipe_crc_irq_handler(dev, pipe);
if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
intel_cpu_fifo_underrun_irq_handler(dev_priv,
pipe);
}
iir = new_iir;
}
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
ret = IRQ_HANDLED;
out:
enable_rpm_wakeref_asserts(dev_priv);
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
return ret;
}
static void i8xx_irq_uninstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
for_each_pipe(dev_priv, pipe) {
/* Clear enable bits; then clear status bits */
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
}
I915_WRITE16(IMR, 0xffff);
I915_WRITE16(IER, 0x0);
I915_WRITE16(IIR, I915_READ16(IIR));
}
static void i915_irq_preinstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
if (I915_HAS_HOTPLUG(dev)) {
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
}
I915_WRITE16(HWSTAM, 0xeffe);
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE(IMR, 0xffffffff);
I915_WRITE(IER, 0x0);
POSTING_READ(IER);
}
static int i915_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 enable_mask;
I915_WRITE(EMR, ~(I915_ERROR_PAGE_TABLE | I915_ERROR_MEMORY_REFRESH));
/* Unmask the interrupts that we always want on. */
dev_priv->irq_mask =
~(I915_ASLE_INTERRUPT |
I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT);
enable_mask =
I915_ASLE_INTERRUPT |
I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
I915_USER_INTERRUPT;
if (I915_HAS_HOTPLUG(dev)) {
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
POSTING_READ(PORT_HOTPLUG_EN);
/* Enable in IER... */
enable_mask |= I915_DISPLAY_PORT_INTERRUPT;
/* and unmask in IMR */
dev_priv->irq_mask &= ~I915_DISPLAY_PORT_INTERRUPT;
}
I915_WRITE(IMR, dev_priv->irq_mask);
I915_WRITE(IER, enable_mask);
POSTING_READ(IER);
i915_enable_asle_pipestat(dev);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
/* Interrupt setup is already guaranteed to be single-threaded, this is
* just to make the assert_spin_locked check happy. */
spin_lock_irq(&dev_priv->irq_lock);
i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
spin_unlock_irq(&dev_priv->irq_lock);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
return 0;
}
/*
* Returns true when a page flip has completed.
*/
static bool i915_handle_vblank(struct drm_device *dev,
int plane, int pipe, u32 iir)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
drm/i915: Make sprite updates atomic Add a mechanism by which we can evade the leading edge of vblank. This guarantees that no two sprite register writes will straddle on either side of the vblank start, and that means all the writes will be latched together in one atomic operation. We do the vblank evade by checking the scanline counter, and if it's too close to the start of vblank (too close has been hardcoded to 100usec for now), we will wait for the vblank start to pass. In order to eliminate random delayes from the rest of the system, we operate with interrupts disabled, except when waiting for the vblank obviously. Note that we now go digging through pipe_to_crtc_mapping[] in the vblank interrupt handler, which is a bit dangerous since we set up interrupts before the crtcs. However in this case since it's the vblank interrupt, we don't actually unmask it until some piece of code requests it. v2: preempt_check_resched() calls after local_irq_enable() (Jesse) Hook up the vblank irq stuff on BDW as well v3: Pass intel_crtc instead of drm_crtc (Daniel) Warn if crtc.mutex isn't locked (Daniel) Add an explicit compiler barrier and document the barriers (Daniel) Note the irq vs. modeset setup madness in the commit message (Daniel) v4: Use prepare_to_wait() & co. directly and eliminate vbl_received v5: Refactor intel_pipe_handle_vblank() vs. drm_handle_vblank() (Chris) Check for min/max scanline <= 0 (Chris) Don't call intel_pipe_update_end() if start failed totally (Chris) Check that the vblank counters match on both sides of the critical section (Chris) v6: Fix atomic update for interlaced modes v7: Reorder code for better readability (Chris) v8: Drop preempt_check_resched(). It's not available to modules anymore and isn't even needed unless we ourselves cause a wakeup needing reschedule while interrupts are off Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> Reviewed-by: Sourab Gupta <sourabgupta@gmail.com> Reviewed-by: Akash Goel <akash.goels@gmail.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-04-29 18:35:46 +08:00
if (!intel_pipe_handle_vblank(dev, pipe))
return false;
if ((iir & flip_pending) == 0)
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
goto check_page_flip;
/* We detect FlipDone by looking for the change in PendingFlip from '1'
* to '0' on the following vblank, i.e. IIR has the Pendingflip
* asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
* the flip is completed (no longer pending). Since this doesn't raise
* an interrupt per se, we watch for the change at vblank.
*/
if (I915_READ(ISR) & flip_pending)
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
goto check_page_flip;
drm/i915: Don't call intel_prepare_page_flip() multiple times on gen2-4 The flip stall detector kicks in when pending>=INTEL_FLIP_COMPLETE. That means if we first call intel_prepare_page_flip() but don't call intel_finish_page_flip(), the next stall check will erroneosly think the page flip was somehow stuck. With enough debug spew emitted from the interrupt handler my 830 hangs when this happens. My theory is that the previous vblank interrupt gets sufficiently delayed that the handler will see the pending bit set in IIR, but ISR still has the bit set as well (ie. the flip was processed by CS but didn't complete yet). In this case the handler will proceed to call intel_check_page_flip() immediately after intel_prepare_page_flip(). It then tries to print a backtrace for the stuck flip WARN, which apparetly results in way too much debug spew delaying interrupt processing further. That then seems to cause an endless loop in the interrupt handler, and the machine is dead until the watchdog kicks in and reboots. At least limiting the number of iterations of the loop in the interrupt handler also prevented the hang. So it seems better to not call intel_prepare_page_flip() without immediately calling intel_finish_page_flip(). The IIR/ISR trickery avoids races here so this is a perfectly safe thing to do. v2: Fix typo in commit message (checkpatch) Cc: stable@vger.kernel.org Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88381 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85888 Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2014-12-18 05:08:03 +08:00
intel_prepare_page_flip(dev, plane);
intel_finish_page_flip(dev, pipe);
return true;
drm/i915: Check for a stalled page flip after each vblank Long ago, back in the racy haydays of 915gm interrupt handling, page flips would occasionally go astray and leave the hardware stuck, and the display not updating. This annoyed people who relied on their systems being able to display continuously updating information 24/7, and so some code to detect when the driver missed the page flip completion signal was added. Until recently, it was presumed that the interrupt handling was now flawless, but once again Simon Farnsworth has found a system whose display will stall. Reinstate the pageflip stall detection, which works by checking to see if the hardware has been updated to the new framebuffer address following each vblank. If the hardware is scanning out from the new framebuffer, but we still think the flip is pending, then we kick our driver into submision. This is a continuation of the effort started with commit 4e5359cd053bfb7d8dabe4a63624a5726848ffbc Author: Simon Farnsworth <simon.farnsworth@onelan.co.uk> Date: Wed Sep 1 17:47:52 2010 +0100 drm/i915: Avoid pageflipping freeze when we miss the flip prepare interrupt This now includes a belt-and-braces approach to make sure the driver (or the hardware) doesn't miss an interrupt and cause us to stop updating the display should the unthinkable happen and the pageflip fail - i.e. that the user is able to continue submitting flips. v2: Cleanup, refactor, and rename v3: Only start counting vblanks after the flip command has been seen by the hardware. v4: Record the seqno after we touch the ring, or else there may be no seqno allocated yet. v5: Rebase on mmio-flip. v6: Rebase, rebase. Reported-by: Simon Farnsworth <simon@farnz.org.uk> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75502 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> [v4] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2014-09-05 14:13:24 +08:00
check_page_flip:
intel_check_page_flip(dev, pipe);
return false;
}
static irqreturn_t i915_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 iir, new_iir, pipe_stats[I915_MAX_PIPES];
u32 flip_mask =
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
int pipe, ret = IRQ_NONE;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
iir = I915_READ(IIR);
do {
bool irq_received = (iir & ~flip_mask) != 0;
bool blc_event = false;
/* Can't rely on pipestat interrupt bit in iir as it might
* have been cleared after the pipestat interrupt was received.
* It doesn't set the bit in iir again, but it still produces
* interrupts (for non-MSI).
*/
spin_lock(&dev_priv->irq_lock);
if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
for_each_pipe(dev_priv, pipe) {
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = PIPESTAT(pipe);
pipe_stats[pipe] = I915_READ(reg);
/* Clear the PIPE*STAT regs before the IIR */
if (pipe_stats[pipe] & 0x8000ffff) {
I915_WRITE(reg, pipe_stats[pipe]);
irq_received = true;
}
}
spin_unlock(&dev_priv->irq_lock);
if (!irq_received)
break;
/* Consume port. Then clear IIR or we'll miss events */
if (I915_HAS_HOTPLUG(dev) &&
iir & I915_DISPLAY_PORT_INTERRUPT)
i9xx_hpd_irq_handler(dev);
I915_WRITE(IIR, iir & ~flip_mask);
new_iir = I915_READ(IIR); /* Flush posted writes */
if (iir & I915_USER_INTERRUPT)
notify_ring(&dev_priv->engine[RCS]);
for_each_pipe(dev_priv, pipe) {
int plane = pipe;
if (HAS_FBC(dev))
plane = !plane;
if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
i915_handle_vblank(dev, plane, pipe, iir))
flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
blc_event = true;
if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
i9xx_pipe_crc_irq_handler(dev, pipe);
if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
intel_cpu_fifo_underrun_irq_handler(dev_priv,
pipe);
}
if (blc_event || (iir & I915_ASLE_INTERRUPT))
intel_opregion_asle_intr(dev);
/* With MSI, interrupts are only generated when iir
* transitions from zero to nonzero. If another bit got
* set while we were handling the existing iir bits, then
* we would never get another interrupt.
*
* This is fine on non-MSI as well, as if we hit this path
* we avoid exiting the interrupt handler only to generate
* another one.
*
* Note that for MSI this could cause a stray interrupt report
* if an interrupt landed in the time between writing IIR and
* the posting read. This should be rare enough to never
* trigger the 99% of 100,000 interrupts test for disabling
* stray interrupts.
*/
ret = IRQ_HANDLED;
iir = new_iir;
} while (iir & ~flip_mask);
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
enable_rpm_wakeref_asserts(dev_priv);
return ret;
}
static void i915_irq_uninstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
if (I915_HAS_HOTPLUG(dev)) {
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
}
I915_WRITE16(HWSTAM, 0xffff);
for_each_pipe(dev_priv, pipe) {
/* Clear enable bits; then clear status bits */
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE(PIPESTAT(pipe), I915_READ(PIPESTAT(pipe)));
}
I915_WRITE(IMR, 0xffffffff);
I915_WRITE(IER, 0x0);
I915_WRITE(IIR, I915_READ(IIR));
}
static void i965_irq_preinstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
I915_WRITE(HWSTAM, 0xeffe);
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE(IMR, 0xffffffff);
I915_WRITE(IER, 0x0);
POSTING_READ(IER);
}
static int i965_irq_postinstall(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
u32 enable_mask;
u32 error_mask;
/* Unmask the interrupts that we always want on. */
dev_priv->irq_mask = ~(I915_ASLE_INTERRUPT |
I915_DISPLAY_PORT_INTERRUPT |
I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT |
I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT);
enable_mask = ~dev_priv->irq_mask;
enable_mask &= ~(I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT);
enable_mask |= I915_USER_INTERRUPT;
if (IS_G4X(dev))
enable_mask |= I915_BSD_USER_INTERRUPT;
/* Interrupt setup is already guaranteed to be single-threaded, this is
* just to make the assert_spin_locked check happy. */
spin_lock_irq(&dev_priv->irq_lock);
i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
i915_enable_pipestat(dev_priv, PIPE_A, PIPE_CRC_DONE_INTERRUPT_STATUS);
i915_enable_pipestat(dev_priv, PIPE_B, PIPE_CRC_DONE_INTERRUPT_STATUS);
spin_unlock_irq(&dev_priv->irq_lock);
/*
* Enable some error detection, note the instruction error mask
* bit is reserved, so we leave it masked.
*/
if (IS_G4X(dev)) {
error_mask = ~(GM45_ERROR_PAGE_TABLE |
GM45_ERROR_MEM_PRIV |
GM45_ERROR_CP_PRIV |
I915_ERROR_MEMORY_REFRESH);
} else {
error_mask = ~(I915_ERROR_PAGE_TABLE |
I915_ERROR_MEMORY_REFRESH);
}
I915_WRITE(EMR, error_mask);
I915_WRITE(IMR, dev_priv->irq_mask);
I915_WRITE(IER, enable_mask);
POSTING_READ(IER);
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
POSTING_READ(PORT_HOTPLUG_EN);
i915_enable_asle_pipestat(dev);
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
return 0;
}
static void i915_hpd_irq_setup(struct drm_device *dev)
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
{
struct drm_i915_private *dev_priv = dev->dev_private;
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
u32 hotplug_en;
assert_spin_locked(&dev_priv->irq_lock);
/* Note HDMI and DP share hotplug bits */
/* enable bits are the same for all generations */
hotplug_en = intel_hpd_enabled_irqs(dev, hpd_mask_i915);
/* Programming the CRT detection parameters tends
to generate a spurious hotplug event about three
seconds later. So just do it once.
*/
if (IS_G4X(dev))
hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64;
hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50;
/* Ignore TV since it's buggy */
i915_hotplug_interrupt_update_locked(dev_priv,
HOTPLUG_INT_EN_MASK |
CRT_HOTPLUG_VOLTAGE_COMPARE_MASK |
CRT_HOTPLUG_ACTIVATION_PERIOD_64,
hotplug_en);
}
static irqreturn_t i965_irq_handler(int irq, void *arg)
{
struct drm_device *dev = arg;
struct drm_i915_private *dev_priv = dev->dev_private;
u32 iir, new_iir;
u32 pipe_stats[I915_MAX_PIPES];
int ret = IRQ_NONE, pipe;
u32 flip_mask =
I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT |
I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
if (!intel_irqs_enabled(dev_priv))
return IRQ_NONE;
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
/* IRQs are synced during runtime_suspend, we don't require a wakeref */
disable_rpm_wakeref_asserts(dev_priv);
iir = I915_READ(IIR);
for (;;) {
bool irq_received = (iir & ~flip_mask) != 0;
bool blc_event = false;
/* Can't rely on pipestat interrupt bit in iir as it might
* have been cleared after the pipestat interrupt was received.
* It doesn't set the bit in iir again, but it still produces
* interrupts (for non-MSI).
*/
spin_lock(&dev_priv->irq_lock);
if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
DRM_DEBUG("Command parser error, iir 0x%08x\n", iir);
for_each_pipe(dev_priv, pipe) {
drm/i915: Type safe register read/write Make I915_READ and I915_WRITE more type safe by wrapping the register offset in a struct. This should eliminate most of the fumbles we've had with misplaced parens. This only takes care of normal mmio registers. We could extend the idea to other register types and define each with its own struct. That way you wouldn't be able to accidentally pass the wrong thing to a specific register access function. The gpio_reg setup is probably the ugliest thing left. But I figure I'd just leave it for now, and wait for some divine inspiration to strike before making it nice. As for the generated code, it's actually a bit better sometimes. Eg. looking at i915_irq_handler(), we can see the following change: lea 0x70024(%rdx,%rax,1),%r9d mov $0x1,%edx - movslq %r9d,%r9 - mov %r9,%rsi - mov %r9,-0x58(%rbp) - callq *0xd8(%rbx) + mov %r9d,%esi + mov %r9d,-0x48(%rbp) callq *0xd8(%rbx) So previously gcc thought the register offset might be signed and decided to sign extend it, just in case. The rest appears to be mostly just minor shuffling of instructions. v2: i915_mmio_reg_{offset,equal,valid}() helpers added s/_REG/_MMIO/ in the register defines mo more switch statements left to worry about ring_emit stuff got sorted in a prep patch cmd parser, lrc context and w/a batch buildup also in prep patch vgpu stuff cleaned up and moved to a prep patch all other unrelated changes split out v3: Rebased due to BXT DSI/BLC, MOCS, etc. v4: Rebased due to churn, s/i915_mmio_reg_t/i915_reg_t/ Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1447853606-2751-1-git-send-email-ville.syrjala@linux.intel.com
2015-11-18 21:33:26 +08:00
i915_reg_t reg = PIPESTAT(pipe);
pipe_stats[pipe] = I915_READ(reg);
/*
* Clear the PIPE*STAT regs before the IIR
*/
if (pipe_stats[pipe] & 0x8000ffff) {
I915_WRITE(reg, pipe_stats[pipe]);
irq_received = true;
}
}
spin_unlock(&dev_priv->irq_lock);
if (!irq_received)
break;
ret = IRQ_HANDLED;
/* Consume port. Then clear IIR or we'll miss events */
if (iir & I915_DISPLAY_PORT_INTERRUPT)
i9xx_hpd_irq_handler(dev);
I915_WRITE(IIR, iir & ~flip_mask);
new_iir = I915_READ(IIR); /* Flush posted writes */
if (iir & I915_USER_INTERRUPT)
notify_ring(&dev_priv->engine[RCS]);
if (iir & I915_BSD_USER_INTERRUPT)
notify_ring(&dev_priv->engine[VCS]);
for_each_pipe(dev_priv, pipe) {
if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
i915_handle_vblank(dev, pipe, pipe, iir))
flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
blc_event = true;
if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
i9xx_pipe_crc_irq_handler(dev, pipe);
if (pipe_stats[pipe] & PIPE_FIFO_UNDERRUN_STATUS)
intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
}
if (blc_event || (iir & I915_ASLE_INTERRUPT))
intel_opregion_asle_intr(dev);
if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
gmbus_irq_handler(dev);
/* With MSI, interrupts are only generated when iir
* transitions from zero to nonzero. If another bit got
* set while we were handling the existing iir bits, then
* we would never get another interrupt.
*
* This is fine on non-MSI as well, as if we hit this path
* we avoid exiting the interrupt handler only to generate
* another one.
*
* Note that for MSI this could cause a stray interrupt report
* if an interrupt landed in the time between writing IIR and
* the posting read. This should be rare enough to never
* trigger the 99% of 100,000 interrupts test for disabling
* stray interrupts.
*/
iir = new_iir;
}
drm/i915: add support for checking if we hold an RPM reference Atm, we assert that the device is not suspended until the point when the device is truly put to a suspended state. This is fine, but we can catch more problems if we check that RPM refcount is non-zero. After that one drops to zero we shouldn't access the device any more, even if the actual device suspend may be delayed. Change assert_rpm_wakelock_held() accordingly to check for a non-zero RPM refcount in addition to the current device-not-suspended check. For the new asserts to work we need to annotate every place explicitly in the code where we expect that the device is powered. The places where we only assume this, but may not hold an RPM reference: - driver load We assume the device to be powered until we enable RPM. Make this explicit by taking an RPM reference around the load function. - system and runtime sudpend/resume handlers These handlers are called when the RPM reference becomes 0 and know the exact point after which the device can get powered off. Disable the RPM-reference-held check for their duration. - the IRQ, hangcheck and RPS work handlers These handlers are flushed in the system/runtime suspend handler before the device is powered off, so it's guaranteed that they won't run while the device is powered off even though they don't hold any RPM reference. Disable the RPM-reference-held check for their duration. In all these cases we still check that the device is not suspended. These explicit annotations also have the positive side effect of documenting our assumptions better. This caught additional WARNs from the atomic modeset path, those should be fixed separately. v2: - remove the redundant HAS_RUNTIME_PM check (moved to patch 1) (Ville) v3: - use a new dedicated RPM wakelock refcount to also catch cases where our own RPM get/put functions were not called (Chris) - assert also that the new RPM wakelock refcount is 0 in the RPM suspend handler (Chris) - change the assert error message to be more meaningful (Chris) - prevent false assert errors and check that the RPM wakelock is 0 in the RPM resume handler too - prevent false assert errors in the hangcheck work too - add a device not suspended assert check to the hangcheck work v4: - rename disable/enable_rpm_asserts to disable/enable_rpm_wakeref_asserts and wakelock_count to wakeref_count - disable the wakeref asserts in the IRQ handlers and RPS work too - update/clarify commit message v5: - mark places we plan to change to use proper RPM refcounting with separate DISABLE/ENABLE_RPM_WAKEREF_ASSERTS aliases (Chris) Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Link: http://patchwork.freedesktop.org/patch/msgid/1450227139-13471-1-git-send-email-imre.deak@intel.com
2015-12-16 08:52:19 +08:00
enable_rpm_wakeref_asserts(dev_priv);
return ret;
}
static void i965_irq_uninstall(struct drm_device * dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
int pipe;
if (!dev_priv)
return;
i915_hotplug_interrupt_update(dev_priv, 0xffffffff, 0);
I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
I915_WRITE(HWSTAM, 0xffffffff);
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe), 0);
I915_WRITE(IMR, 0xffffffff);
I915_WRITE(IER, 0x0);
for_each_pipe(dev_priv, pipe)
I915_WRITE(PIPESTAT(pipe),
I915_READ(PIPESTAT(pipe)) & 0x8000ffff);
I915_WRITE(IIR, I915_READ(IIR));
}
/**
* intel_irq_init - initializes irq support
* @dev_priv: i915 device instance
*
* This function initializes all the irq support including work items, timers
* and all the vtables. It does not setup the interrupt itself though.
*/
void intel_irq_init(struct drm_i915_private *dev_priv)
{
struct drm_device *dev = dev_priv->dev;
intel_hpd_init_work(dev_priv);
INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work);
INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
/* Let's track the enabled rps events */
if (IS_VALLEYVIEW(dev_priv))
/* WaGsvRC0ResidencyMethod:vlv */
dev_priv->pm_rps_events = GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED;
else
dev_priv->pm_rps_events = GEN6_PM_RPS_EVENTS;
INIT_DELAYED_WORK(&dev_priv->gpu_error.hangcheck_work,
i915_hangcheck_elapsed);
if (IS_GEN2(dev_priv)) {
dev->max_vblank_count = 0;
dev->driver->get_vblank_counter = i8xx_get_vblank_counter;
} else if (IS_G4X(dev_priv) || INTEL_INFO(dev_priv)->gen >= 5) {
dev->max_vblank_count = 0xffffffff; /* full 32 bit counter */
dev->driver->get_vblank_counter = g4x_get_vblank_counter;
} else {
dev->driver->get_vblank_counter = i915_get_vblank_counter;
dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
}
/*
* Opt out of the vblank disable timer on everything except gen2.
* Gen2 doesn't have a hardware frame counter and so depends on
* vblank interrupts to produce sane vblank seuquence numbers.
*/
if (!IS_GEN2(dev_priv))
dev->vblank_disable_immediate = true;
dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
if (IS_CHERRYVIEW(dev_priv)) {
dev->driver->irq_handler = cherryview_irq_handler;
dev->driver->irq_preinstall = cherryview_irq_preinstall;
dev->driver->irq_postinstall = cherryview_irq_postinstall;
dev->driver->irq_uninstall = cherryview_irq_uninstall;
dev->driver->enable_vblank = valleyview_enable_vblank;
dev->driver->disable_vblank = valleyview_disable_vblank;
dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
} else if (IS_VALLEYVIEW(dev_priv)) {
dev->driver->irq_handler = valleyview_irq_handler;
dev->driver->irq_preinstall = valleyview_irq_preinstall;
dev->driver->irq_postinstall = valleyview_irq_postinstall;
dev->driver->irq_uninstall = valleyview_irq_uninstall;
dev->driver->enable_vblank = valleyview_enable_vblank;
dev->driver->disable_vblank = valleyview_disable_vblank;
dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
} else if (INTEL_INFO(dev_priv)->gen >= 8) {
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
dev->driver->irq_handler = gen8_irq_handler;
dev->driver->irq_preinstall = gen8_irq_reset;
drm/i915/bdw: Implement interrupt changes The interrupt handling implementation remains the same as previous generations with the 4 types of registers, status, identity, mask, and enable. However the layout of where the bits go have changed entirely. To address these changes, all of the interrupt vfuncs needed special gen8 code. The way it works is there is a top level status register now which informs the interrupt service routine which unit caused the interrupt, and therefore which interrupt registers to read to process the interrupt. For display the division is quite logical, a set of interrupt registers for each pipe, and in addition to those, a set each for "misc" and port. For GT the things get a bit hairy, as seen by the code. Each of the GT units has it's own bits defined. They all look *very similar* and resides in 16 bits of a GT register. As an example, RCS and BCS share register 0. To compact the code a bit, at a slight expense to complexity, this is exactly how the code works as well. 2 structures are added to the ring buffer so that our ring buffer interrupt handling code knows which ring shares the interrupt registers, and a shift value (ie. the top or bottom 16 bits of the register). The above allows us to kept the interrupt register caching scheme, the per interrupt enables, and the code to mask and unmask interrupts relatively clean (again at the cost of some more complexity). Most of the GT units mentioned above are command streamers, and so the symmetry should work quite well for even the yet to be implemented rings which Broadwell adds. v2: Fixes up a couple of bugs, and is more verbose about errors in the Broadwell interrupt handler. v3: fix DE_MISC IER offset v4: Simplify interrupts: I totally misread the docs the first time I implemented interrupts, and so this should greatly simplify the mess. Unlike GEN6, we never touch the regular mask registers in irq_get/put. v5: Rebased on to of recent pch hotplug setup changes. v6: Fixup on top of moving num_pipes to intel_info. v7: Rebased on top of Egbert Eich's hpd irq handling rework. Also wired up ibx_hpd_irq_setup for gen8. v8: Rebase on top of Jani's asle handling rework. v9: Rebase on top of Ben's VECS enabling for Haswell, where he unfortunately went OCD on the gt irq #defines. Not that they're still not yet fully consistent: - Used the GT_RENDER_ #defines + bdw shifts. - Dropped the shift from the L3_PARITY stuff, seemed clearer. - s/irq_refcount/irq_refcount.gt/ v10: Squash in VECS enabling patches and the gen8_gt_irq_handler refactoring from Zhao Yakui <yakui.zhao@intel.com> v11: Rebase on top of the interrupt cleanups in upstream. v12: Rebase on top of Ben's DPF changes in upstream. v13: Drop bdw from the HAS_L3_DPF feature flag for now, it's unclear what exactly needs to be done. Requested by Ben. v14: Fix the patch. - Drop the mask of reserved bits and assorted logic, it doesn't match the spec. - Do the posting read inconditionally instead of commenting it out. - Add a GEN8_MASTER_IRQ_CONTROL definition and use it. - Fix up the GEN8_PIPE interrupt defines and give the GEN8_ prefixes - we actually will need to use them. - Enclose macros in do {} while (0) (checkpatch). - Clear DE_MISC interrupt bits only after having processed them. - Fix whitespace fail (checkpatch). - Fix overtly long lines where appropriate (checkpatch). - Don't use typedef'ed private_t (maintainer-scripts). - Align the function parameter list correctly. Signed-off-by: Ben Widawsky <ben@bwidawsk.net> (v4) Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> bikeshed
2013-11-03 12:07:09 +08:00
dev->driver->irq_postinstall = gen8_irq_postinstall;
dev->driver->irq_uninstall = gen8_irq_uninstall;
dev->driver->enable_vblank = gen8_enable_vblank;
dev->driver->disable_vblank = gen8_disable_vblank;
if (IS_BROXTON(dev))
dev_priv->display.hpd_irq_setup = bxt_hpd_irq_setup;
else if (HAS_PCH_SPT(dev))
dev_priv->display.hpd_irq_setup = spt_hpd_irq_setup;
else
dev_priv->display.hpd_irq_setup = ilk_hpd_irq_setup;
} else if (HAS_PCH_SPLIT(dev)) {
dev->driver->irq_handler = ironlake_irq_handler;
dev->driver->irq_preinstall = ironlake_irq_reset;
dev->driver->irq_postinstall = ironlake_irq_postinstall;
dev->driver->irq_uninstall = ironlake_irq_uninstall;
dev->driver->enable_vblank = ironlake_enable_vblank;
dev->driver->disable_vblank = ironlake_disable_vblank;
dev_priv->display.hpd_irq_setup = ilk_hpd_irq_setup;
} else {
if (INTEL_INFO(dev_priv)->gen == 2) {
dev->driver->irq_preinstall = i8xx_irq_preinstall;
dev->driver->irq_postinstall = i8xx_irq_postinstall;
dev->driver->irq_handler = i8xx_irq_handler;
dev->driver->irq_uninstall = i8xx_irq_uninstall;
} else if (INTEL_INFO(dev_priv)->gen == 3) {
dev->driver->irq_preinstall = i915_irq_preinstall;
dev->driver->irq_postinstall = i915_irq_postinstall;
dev->driver->irq_uninstall = i915_irq_uninstall;
dev->driver->irq_handler = i915_irq_handler;
} else {
dev->driver->irq_preinstall = i965_irq_preinstall;
dev->driver->irq_postinstall = i965_irq_postinstall;
dev->driver->irq_uninstall = i965_irq_uninstall;
dev->driver->irq_handler = i965_irq_handler;
}
if (I915_HAS_HOTPLUG(dev_priv))
dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
dev->driver->enable_vblank = i915_enable_vblank;
dev->driver->disable_vblank = i915_disable_vblank;
}
}
drm/i915: Fixup hpd irq register setup ordering For GMCH platforms we set up the hpd irq registers in the irq postinstall hook. But since we only enable the irq sources we actually need in PORT_HOTPLUG_EN/STATUS, taking dev_priv->hotplug_supported_mask into account, no hpd interrupt sources is enabled since commit 52d7ecedac3f96fb562cb482c139015372728638 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Sat Dec 1 21:03:22 2012 +0100 drm/i915: reorder setup sequence to have irqs for output setup Wrongly set-up interrupts also lead to broken hw-based load-detection on at least GM45, resulting in ghost VGA/TV-out outputs. To fix this, delay the hotplug register setup until after all outputs are set up, by moving it into a new dev_priv->display.hpd_irq_callback. We might also move the PCH_SPLIT platforms to such a setup eventually. Another funny part is that we need to delay the fbdev initial config probing until after the hpd regs are setup, for otherwise it'll detect ghost outputs. But we can only enable the hpd interrupt handling itself (and the output polling) _after_ that initial scan, due to massive locking brain-damage in the fbdev setup code. Add a big comment to explain this cute little dragon lair. v2: Encapsulate all the fbdev handling by wrapping the move call into intel_fbdev_initial_config in intel_fb.c. Requested by Chris Wilson. v3: Applied bikeshed from Jesse Barnes. v4: Imre Deak noticed that we also need to call intel_hpd_init after the drm_irqinstall calls in the gpu reset and resume paths - otherwise hotplug will be broken. Also improve the comment a bit about why hpd_init needs to be called before we set up the initial fbdev config. Bugzilla: Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54943 Reported-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v3) Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2012-12-11 21:05:07 +08:00
/**
* intel_irq_install - enables the hardware interrupt
* @dev_priv: i915 device instance
*
* This function enables the hardware interrupt handling, but leaves the hotplug
* handling still disabled. It is called after intel_irq_init().
*
* In the driver load and resume code we need working interrupts in a few places
* but don't want to deal with the hassle of concurrent probe and hotplug
* workers. Hence the split into this two-stage approach.
*/
int intel_irq_install(struct drm_i915_private *dev_priv)
{
/*
* We enable some interrupt sources in our postinstall hooks, so mark
* interrupts as enabled _before_ actually enabling them to avoid
* special cases in our ordering checks.
*/
dev_priv->pm.irqs_enabled = true;
return drm_irq_install(dev_priv->dev, dev_priv->dev->pdev->irq);
}
/**
* intel_irq_uninstall - finilizes all irq handling
* @dev_priv: i915 device instance
*
* This stops interrupt and hotplug handling and unregisters and frees all
* resources acquired in the init functions.
*/
void intel_irq_uninstall(struct drm_i915_private *dev_priv)
{
drm_irq_uninstall(dev_priv->dev);
intel_hpd_cancel_work(dev_priv);
dev_priv->pm.irqs_enabled = false;
}
/**
* intel_runtime_pm_disable_interrupts - runtime interrupt disabling
* @dev_priv: i915 device instance
*
* This function is used to disable interrupts at runtime, both in the runtime
* pm and the system suspend/resume code.
*/
void intel_runtime_pm_disable_interrupts(struct drm_i915_private *dev_priv)
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
{
dev_priv->dev->driver->irq_uninstall(dev_priv->dev);
dev_priv->pm.irqs_enabled = false;
drm/i915: avoid processing spurious/shared interrupts in low-power states Atm, it's possible that the interrupt handler is called when the device is in D3 or some other low-power state. It can be due to another device that is still in D0 state and shares the interrupt line with i915, or on some platforms there could be spurious interrupts even without sharing the interrupt line. The latter case was reported by Klaus Ethgen using a Lenovo x61p machine (gen 4). He noticed this issue via a system suspend/resume hang and bisected it to the following commit: commit e11aa362308f5de467ce355a2a2471321b15a35c Author: Jesse Barnes <jbarnes@virtuousgeek.org> Date: Wed Jun 18 09:52:55 2014 -0700 drm/i915: use runtime irq suspend/resume in freeze/thaw This is a problem, since in low-power states IIR will always read 0xffffffff resulting in an endless IRQ servicing loop. Fix this by handling interrupts only when the driver explicitly enables them and so it's guaranteed that the interrupt registers return a valid value. Note that this issue existed even before the above commit, since during runtime suspend/resume we never unregistered the handler. v2: - clarify the purpose of smp_mb() vs. synchronize_irq() in the code comment (Chris) v3: - no need for an explicit smp_mb(), we can assume that synchronize_irq() and the mmio read/writes in the install hooks provide for this (Daniel) - remove code comment as the remaining synchronize_irq() is self explanatory (Daniel) v4: - drm_irq_uninstall() implies synchronize_irq(), so no need to call it explicitly (Daniel) Reference: https://lkml.org/lkml/2015/2/11/205 Reported-and-bisected-by: Klaus Ethgen <Klaus@Ethgen.ch> Cc: stable@vger.kernel.org Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-02-24 17:14:30 +08:00
synchronize_irq(dev_priv->dev->irq);
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
}
/**
* intel_runtime_pm_enable_interrupts - runtime interrupt enabling
* @dev_priv: i915 device instance
*
* This function is used to enable interrupts at runtime, both in the runtime
* pm and the system suspend/resume code.
*/
void intel_runtime_pm_enable_interrupts(struct drm_i915_private *dev_priv)
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
{
dev_priv->pm.irqs_enabled = true;
dev_priv->dev->driver->irq_preinstall(dev_priv->dev);
dev_priv->dev->driver->irq_postinstall(dev_priv->dev);
drm/i915: allow package C8+ states on Haswell (disabled) This patch allows PC8+ states on Haswell. These states can only be reached when all the display outputs are disabled, and they allow some more power savings. The fact that the graphics device is allowing PC8+ doesn't mean that the machine will actually enter PC8+: all the other devices also need to allow PC8+. For now this option is disabled by default. You need i915.allow_pc8=1 if you want it. This patch adds a big comment inside i915_drv.h explaining how it works and how it tracks things. Read it. v2: (this is not really v2, many previous versions were already sent, but they had different names) - Use the new functions to enable/disable GTIMR and GEN6_PMIMR - Rename almost all variables and functions to names suggested by Chris - More WARNs on the IRQ handling code - Also disable PC8 when there's GPU work to do (thanks to Ben for the help on this), so apps can run caster - Enable PC8 on a delayed work function that is delayed for 5 seconds. This makes sure we only enable PC8+ if we're really idle - Make sure we're not in PC8+ when suspending v3: - WARN if IRQs are disabled on __wait_seqno - Replace some DRM_ERRORs with WARNs - Fix calls to restore GT and PM interrupts - Use intel_mark_busy instead of intel_ring_advance to disable PC8 v4: - Use the force_wake, Luke! v5: - Remove the "IIR is not zero" WARNs - Move the force_wake chunk to its own patch - Only restore what's missing from RC6, not everything Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2013-08-20 00:18:09 +08:00
}