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

3409 lines
99 KiB
C
Raw Normal View History

/*
* Copyright © 2012-2014 Intel Corporation
*
* 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, sublicense,
* 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 NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS 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.
*
* Authors:
* Eugeni Dodonov <eugeni.dodonov@intel.com>
* Daniel Vetter <daniel.vetter@ffwll.ch>
*
*/
#include <linux/pm_runtime.h>
#include <linux/vgaarb.h>
#include "i915_drv.h"
#include "intel_drv.h"
/**
* DOC: runtime pm
*
* The i915 driver supports dynamic enabling and disabling of entire hardware
* blocks at runtime. This is especially important on the display side where
* software is supposed to control many power gates manually on recent hardware,
* since on the GT side a lot of the power management is done by the hardware.
* But even there some manual control at the device level is required.
*
* Since i915 supports a diverse set of platforms with a unified codebase and
* hardware engineers just love to shuffle functionality around between power
* domains there's a sizeable amount of indirection required. This file provides
* generic functions to the driver for grabbing and releasing references for
* abstract power domains. It then maps those to the actual power wells
* present for a given platform.
*/
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv,
enum i915_power_well_id power_well_id);
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
static struct i915_power_well *
lookup_power_well(struct drm_i915_private *dev_priv,
enum i915_power_well_id power_well_id);
const char *
intel_display_power_domain_str(enum intel_display_power_domain domain)
{
switch (domain) {
case POWER_DOMAIN_PIPE_A:
return "PIPE_A";
case POWER_DOMAIN_PIPE_B:
return "PIPE_B";
case POWER_DOMAIN_PIPE_C:
return "PIPE_C";
case POWER_DOMAIN_PIPE_A_PANEL_FITTER:
return "PIPE_A_PANEL_FITTER";
case POWER_DOMAIN_PIPE_B_PANEL_FITTER:
return "PIPE_B_PANEL_FITTER";
case POWER_DOMAIN_PIPE_C_PANEL_FITTER:
return "PIPE_C_PANEL_FITTER";
case POWER_DOMAIN_TRANSCODER_A:
return "TRANSCODER_A";
case POWER_DOMAIN_TRANSCODER_B:
return "TRANSCODER_B";
case POWER_DOMAIN_TRANSCODER_C:
return "TRANSCODER_C";
case POWER_DOMAIN_TRANSCODER_EDP:
return "TRANSCODER_EDP";
case POWER_DOMAIN_TRANSCODER_DSI_A:
return "TRANSCODER_DSI_A";
case POWER_DOMAIN_TRANSCODER_DSI_C:
return "TRANSCODER_DSI_C";
case POWER_DOMAIN_PORT_DDI_A_LANES:
return "PORT_DDI_A_LANES";
case POWER_DOMAIN_PORT_DDI_B_LANES:
return "PORT_DDI_B_LANES";
case POWER_DOMAIN_PORT_DDI_C_LANES:
return "PORT_DDI_C_LANES";
case POWER_DOMAIN_PORT_DDI_D_LANES:
return "PORT_DDI_D_LANES";
case POWER_DOMAIN_PORT_DDI_E_LANES:
return "PORT_DDI_E_LANES";
case POWER_DOMAIN_PORT_DDI_F_LANES:
return "PORT_DDI_F_LANES";
case POWER_DOMAIN_PORT_DDI_A_IO:
return "PORT_DDI_A_IO";
case POWER_DOMAIN_PORT_DDI_B_IO:
return "PORT_DDI_B_IO";
case POWER_DOMAIN_PORT_DDI_C_IO:
return "PORT_DDI_C_IO";
case POWER_DOMAIN_PORT_DDI_D_IO:
return "PORT_DDI_D_IO";
case POWER_DOMAIN_PORT_DDI_E_IO:
return "PORT_DDI_E_IO";
case POWER_DOMAIN_PORT_DDI_F_IO:
return "PORT_DDI_F_IO";
case POWER_DOMAIN_PORT_DSI:
return "PORT_DSI";
case POWER_DOMAIN_PORT_CRT:
return "PORT_CRT";
case POWER_DOMAIN_PORT_OTHER:
return "PORT_OTHER";
case POWER_DOMAIN_VGA:
return "VGA";
case POWER_DOMAIN_AUDIO:
return "AUDIO";
case POWER_DOMAIN_PLLS:
return "PLLS";
case POWER_DOMAIN_AUX_A:
return "AUX_A";
case POWER_DOMAIN_AUX_B:
return "AUX_B";
case POWER_DOMAIN_AUX_C:
return "AUX_C";
case POWER_DOMAIN_AUX_D:
return "AUX_D";
case POWER_DOMAIN_AUX_F:
return "AUX_F";
case POWER_DOMAIN_GMBUS:
return "GMBUS";
case POWER_DOMAIN_INIT:
return "INIT";
case POWER_DOMAIN_MODESET:
return "MODESET";
drm/i915: Restore GT performance in headless mode with DMC loaded It seems that the DMC likes to transition between the DC states a lot when there are no connected displays (no active power domains) during command submission. This activity on DC states has a negative impact on the performance of the chip with huge latencies observed in the interrupt handlers and elsewhere. Simple tests like igt/gem_latency -n 0 are slowed down by a factor of eight. Work around it by introducing a new power domain named, POWER_DOMAIN_GT_IRQ, associtated with the "DC off" power well, which is held for the duration of command submission activity. CNL has the same problem which will be addressed as a follow-up. Doing that requires a fix for a DC6 context corruption problem in the CNL DMC firmware which is yet to be released. v2: * Add commit text as comment in i915_gem_mark_busy. (Chris Wilson) * Protect macro body with braces. (Jani Nikula) v3: * Add dedicated power domain for clarity. (Chris, Imre) * Commit message and comment text updates. * Apply to all big-core GEN9 parts apart for Skylake which is pending DMC firmware release. v4: * Power domain should be inner to device runtime pm. (Chris) * Simplify NEEDS_CSR_GT_PERF_WA macro. (Chris) * Handle async DMC loading by moving the GT_IRQ power domain logic into intel_runtime_pm. (Daniel, Chris) * Include small core GEN9 as well. (Imre) v5 * Special handling for async DMC load is not needed since on failure the power domain reference is kept permanently taken. (Imre) v6: * Drop the NEEDS_CSR_GT_PERF_WA macro since all firmwares have now been deployed. (Imre, Chris) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100572 Testcase: igt/gem_exec_nop/headless Cc: Imre Deak <imre.deak@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> (v2) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v5) Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> [Imre: Add note about applying the WA on CNL as a follow-up] Signed-off-by: Imre Deak <imre.deak@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171205132854.26380-1-tvrtko.ursulin@linux.intel.com
2017-12-05 21:28:54 +08:00
case POWER_DOMAIN_GT_IRQ:
return "GT_IRQ";
default:
MISSING_CASE(domain);
return "?";
}
}
static void intel_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
DRM_DEBUG_KMS("enabling %s\n", power_well->name);
power_well->ops->enable(dev_priv, power_well);
power_well->hw_enabled = true;
}
static void intel_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
DRM_DEBUG_KMS("disabling %s\n", power_well->name);
power_well->hw_enabled = false;
power_well->ops->disable(dev_priv, power_well);
}
static void intel_power_well_get(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
if (!power_well->count++)
intel_power_well_enable(dev_priv, power_well);
}
static void intel_power_well_put(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN(!power_well->count, "Use count on power well %s is already zero",
power_well->name);
if (!--power_well->count)
intel_power_well_disable(dev_priv, power_well);
}
/**
* __intel_display_power_is_enabled - unlocked check for a power domain
* @dev_priv: i915 device instance
* @domain: power domain to check
*
* This is the unlocked version of intel_display_power_is_enabled() and should
* only be used from error capture and recovery code where deadlocks are
* possible.
*
* Returns:
* True when the power domain is enabled, false otherwise.
*/
bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_well *power_well;
bool is_enabled;
if (dev_priv->runtime_pm.suspended)
return false;
is_enabled = true;
for_each_power_domain_well_rev(dev_priv, power_well, BIT_ULL(domain)) {
if (power_well->always_on)
continue;
if (!power_well->hw_enabled) {
is_enabled = false;
break;
}
}
return is_enabled;
}
/**
* intel_display_power_is_enabled - check for a power domain
* @dev_priv: i915 device instance
* @domain: power domain to check
*
* This function can be used to check the hw power domain state. It is mostly
* used in hardware state readout functions. Everywhere else code should rely
* upon explicit power domain reference counting to ensure that the hardware
* block is powered up before accessing it.
*
* Callers must hold the relevant modesetting locks to ensure that concurrent
* threads can't disable the power well while the caller tries to read a few
* registers.
*
* Returns:
* True when the power domain is enabled, false otherwise.
*/
bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_domains *power_domains;
bool ret;
power_domains = &dev_priv->power_domains;
mutex_lock(&power_domains->lock);
ret = __intel_display_power_is_enabled(dev_priv, domain);
mutex_unlock(&power_domains->lock);
return ret;
}
/**
* intel_display_set_init_power - set the initial power domain state
* @dev_priv: i915 device instance
* @enable: whether to enable or disable the initial power domain state
*
* For simplicity our driver load/unload and system suspend/resume code assumes
* that all power domains are always enabled. This functions controls the state
* of this little hack. While the initial power domain state is enabled runtime
* pm is effectively disabled.
*/
void intel_display_set_init_power(struct drm_i915_private *dev_priv,
bool enable)
{
if (dev_priv->power_domains.init_power_on == enable)
return;
if (enable)
intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
else
intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
dev_priv->power_domains.init_power_on = enable;
}
/*
* Starting with Haswell, we have a "Power Down Well" that can be turned off
* when not needed anymore. We have 4 registers that can request the power well
* to be enabled, and it will only be disabled if none of the registers is
* requesting it to be enabled.
*/
static void hsw_power_well_post_enable(struct drm_i915_private *dev_priv,
u8 irq_pipe_mask, bool has_vga)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
/*
* After we re-enable the power well, if we touch VGA register 0x3d5
* we'll get unclaimed register interrupts. This stops after we write
* anything to the VGA MSR register. The vgacon module uses this
* register all the time, so if we unbind our driver and, as a
* consequence, bind vgacon, we'll get stuck in an infinite loop at
* console_unlock(). So make here we touch the VGA MSR register, making
* sure vgacon can keep working normally without triggering interrupts
* and error messages.
*/
if (has_vga) {
vga_get_uninterruptible(pdev, VGA_RSRC_LEGACY_IO);
outb(inb(VGA_MSR_READ), VGA_MSR_WRITE);
vga_put(pdev, VGA_RSRC_LEGACY_IO);
}
if (irq_pipe_mask)
gen8_irq_power_well_post_enable(dev_priv, irq_pipe_mask);
}
static void hsw_power_well_pre_disable(struct drm_i915_private *dev_priv,
u8 irq_pipe_mask)
{
if (irq_pipe_mask)
gen8_irq_power_well_pre_disable(dev_priv, irq_pipe_mask);
}
static void hsw_wait_for_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
/* Timeout for PW1:10 us, AUX:not specified, other PWs:20 us. */
WARN_ON(intel_wait_for_register(dev_priv,
HSW_PWR_WELL_CTL_DRIVER(id),
HSW_PWR_WELL_CTL_STATE(id),
HSW_PWR_WELL_CTL_STATE(id),
1));
}
static u32 hsw_power_well_requesters(struct drm_i915_private *dev_priv,
enum i915_power_well_id id)
{
u32 req_mask = HSW_PWR_WELL_CTL_REQ(id);
u32 ret;
ret = I915_READ(HSW_PWR_WELL_CTL_BIOS(id)) & req_mask ? 1 : 0;
ret |= I915_READ(HSW_PWR_WELL_CTL_DRIVER(id)) & req_mask ? 2 : 0;
ret |= I915_READ(HSW_PWR_WELL_CTL_KVMR) & req_mask ? 4 : 0;
ret |= I915_READ(HSW_PWR_WELL_CTL_DEBUG(id)) & req_mask ? 8 : 0;
return ret;
}
static void hsw_wait_for_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
bool disabled;
u32 reqs;
/*
* Bspec doesn't require waiting for PWs to get disabled, but still do
* this for paranoia. The known cases where a PW will be forced on:
* - a KVMR request on any power well via the KVMR request register
* - a DMC request on PW1 and MISC_IO power wells via the BIOS and
* DEBUG request registers
* Skip the wait in case any of the request bits are set and print a
* diagnostic message.
*/
wait_for((disabled = !(I915_READ(HSW_PWR_WELL_CTL_DRIVER(id)) &
HSW_PWR_WELL_CTL_STATE(id))) ||
(reqs = hsw_power_well_requesters(dev_priv, id)), 1);
if (disabled)
return;
DRM_DEBUG_KMS("%s forced on (bios:%d driver:%d kvmr:%d debug:%d)\n",
power_well->name,
!!(reqs & 1), !!(reqs & 2), !!(reqs & 4), !!(reqs & 8));
}
static void gen9_wait_for_power_well_fuses(struct drm_i915_private *dev_priv,
enum skl_power_gate pg)
{
/* Timeout 5us for PG#0, for other PGs 1us */
WARN_ON(intel_wait_for_register(dev_priv, SKL_FUSE_STATUS,
SKL_FUSE_PG_DIST_STATUS(pg),
SKL_FUSE_PG_DIST_STATUS(pg), 1));
}
static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
bool wait_fuses = power_well->hsw.has_fuses;
enum skl_power_gate uninitialized_var(pg);
u32 val;
if (wait_fuses) {
pg = SKL_PW_TO_PG(id);
/*
* For PW1 we have to wait both for the PW0/PG0 fuse state
* before enabling the power well and PW1/PG1's own fuse
* state after the enabling. For all other power wells with
* fuses we only have to wait for that PW/PG's fuse state
* after the enabling.
*/
if (pg == SKL_PG1)
gen9_wait_for_power_well_fuses(dev_priv, SKL_PG0);
}
val = I915_READ(HSW_PWR_WELL_CTL_DRIVER(id));
I915_WRITE(HSW_PWR_WELL_CTL_DRIVER(id), val | HSW_PWR_WELL_CTL_REQ(id));
hsw_wait_for_power_well_enable(dev_priv, power_well);
/* Display WA #1178: cnl */
if (IS_CANNONLAKE(dev_priv) &&
(id == CNL_DISP_PW_AUX_B || id == CNL_DISP_PW_AUX_C ||
id == CNL_DISP_PW_AUX_D || id == CNL_DISP_PW_AUX_F)) {
val = I915_READ(CNL_AUX_ANAOVRD1(id));
val |= CNL_AUX_ANAOVRD1_ENABLE | CNL_AUX_ANAOVRD1_LDO_BYPASS;
I915_WRITE(CNL_AUX_ANAOVRD1(id), val);
}
if (wait_fuses)
gen9_wait_for_power_well_fuses(dev_priv, pg);
hsw_power_well_post_enable(dev_priv, power_well->hsw.irq_pipe_mask,
power_well->hsw.has_vga);
}
static void hsw_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
u32 val;
hsw_power_well_pre_disable(dev_priv, power_well->hsw.irq_pipe_mask);
val = I915_READ(HSW_PWR_WELL_CTL_DRIVER(id));
I915_WRITE(HSW_PWR_WELL_CTL_DRIVER(id),
val & ~HSW_PWR_WELL_CTL_REQ(id));
hsw_wait_for_power_well_disable(dev_priv, power_well);
}
/*
* We should only use the power well if we explicitly asked the hardware to
* enable it, so check if it's enabled and also check if we've requested it to
* be enabled.
*/
static bool hsw_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
u32 mask = HSW_PWR_WELL_CTL_REQ(id) | HSW_PWR_WELL_CTL_STATE(id);
return (I915_READ(HSW_PWR_WELL_CTL_DRIVER(id)) & mask) == mask;
}
static void assert_can_enable_dc9(struct drm_i915_private *dev_priv)
{
enum i915_power_well_id id = SKL_DISP_PW_2;
WARN_ONCE((I915_READ(DC_STATE_EN) & DC_STATE_EN_DC9),
"DC9 already programmed to be enabled.\n");
WARN_ONCE(I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5,
"DC5 still not disabled to enable DC9.\n");
WARN_ONCE(I915_READ(HSW_PWR_WELL_CTL_DRIVER(id)) &
HSW_PWR_WELL_CTL_REQ(id),
"Power well 2 on.\n");
WARN_ONCE(intel_irqs_enabled(dev_priv),
"Interrupts not disabled yet.\n");
/*
* TODO: check for the following to verify the conditions to enter DC9
* state are satisfied:
* 1] Check relevant display engine registers to verify if mode set
* disable sequence was followed.
* 2] Check if display uninitialize sequence is initialized.
*/
}
static void assert_can_disable_dc9(struct drm_i915_private *dev_priv)
{
WARN_ONCE(intel_irqs_enabled(dev_priv),
"Interrupts not disabled yet.\n");
WARN_ONCE(I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5,
"DC5 still not disabled.\n");
/*
* TODO: check for the following to verify DC9 state was indeed
* entered before programming to disable it:
* 1] Check relevant display engine registers to verify if mode
* set disable sequence was followed.
* 2] Check if display uninitialize sequence is initialized.
*/
}
static void gen9_write_dc_state(struct drm_i915_private *dev_priv,
u32 state)
{
int rewrites = 0;
int rereads = 0;
u32 v;
I915_WRITE(DC_STATE_EN, state);
/* It has been observed that disabling the dc6 state sometimes
* doesn't stick and dmc keeps returning old value. Make sure
* the write really sticks enough times and also force rewrite until
* we are confident that state is exactly what we want.
*/
do {
v = I915_READ(DC_STATE_EN);
if (v != state) {
I915_WRITE(DC_STATE_EN, state);
rewrites++;
rereads = 0;
} else if (rereads++ > 5) {
break;
}
} while (rewrites < 100);
if (v != state)
DRM_ERROR("Writing dc state to 0x%x failed, now 0x%x\n",
state, v);
/* Most of the times we need one retry, avoid spam */
if (rewrites > 1)
DRM_DEBUG_KMS("Rewrote dc state to 0x%x %d times\n",
state, rewrites);
}
static u32 gen9_dc_mask(struct drm_i915_private *dev_priv)
{
u32 mask;
mask = DC_STATE_EN_UPTO_DC5;
if (IS_GEN9_LP(dev_priv))
mask |= DC_STATE_EN_DC9;
else
mask |= DC_STATE_EN_UPTO_DC6;
return mask;
}
void gen9_sanitize_dc_state(struct drm_i915_private *dev_priv)
{
u32 val;
val = I915_READ(DC_STATE_EN) & gen9_dc_mask(dev_priv);
DRM_DEBUG_KMS("Resetting DC state tracking from %02x to %02x\n",
dev_priv->csr.dc_state, val);
dev_priv->csr.dc_state = val;
}
static void gen9_set_dc_state(struct drm_i915_private *dev_priv, uint32_t state)
{
uint32_t val;
uint32_t mask;
if (WARN_ON_ONCE(state & ~dev_priv->csr.allowed_dc_mask))
state &= dev_priv->csr.allowed_dc_mask;
val = I915_READ(DC_STATE_EN);
mask = gen9_dc_mask(dev_priv);
DRM_DEBUG_KMS("Setting DC state from %02x to %02x\n",
val & mask, state);
/* Check if DMC is ignoring our DC state requests */
if ((val & mask) != dev_priv->csr.dc_state)
DRM_ERROR("DC state mismatch (0x%x -> 0x%x)\n",
dev_priv->csr.dc_state, val & mask);
val &= ~mask;
val |= state;
gen9_write_dc_state(dev_priv, val);
dev_priv->csr.dc_state = val & mask;
}
void bxt_enable_dc9(struct drm_i915_private *dev_priv)
{
assert_can_enable_dc9(dev_priv);
DRM_DEBUG_KMS("Enabling DC9\n");
intel_power_sequencer_reset(dev_priv);
gen9_set_dc_state(dev_priv, DC_STATE_EN_DC9);
}
void bxt_disable_dc9(struct drm_i915_private *dev_priv)
{
assert_can_disable_dc9(dev_priv);
DRM_DEBUG_KMS("Disabling DC9\n");
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
intel_pps_unlock_regs_wa(dev_priv);
}
static void assert_csr_loaded(struct drm_i915_private *dev_priv)
{
WARN_ONCE(!I915_READ(CSR_PROGRAM(0)),
"CSR program storage start is NULL\n");
WARN_ONCE(!I915_READ(CSR_SSP_BASE), "CSR SSP Base Not fine\n");
WARN_ONCE(!I915_READ(CSR_HTP_SKL), "CSR HTP Not fine\n");
}
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
static void assert_can_enable_dc5(struct drm_i915_private *dev_priv)
drm/i915/skl: Add DC5 Trigger Sequence Add triggers as per expectations mentioned in gen9_enable_dc5 and gen9_disable_dc5 patch. Also call POSTING_READ for every write to a register to ensure that its written immediately. v1: Remove POSTING_READ calls as they've already been added in previous patches. v2: Rebase to move all runtime pm specific changes to intel_runtime_pm.c file. Modified as per review comments from Imre: 1] Change variable name 'dc5_allowed' to 'dc5_enabled' to correspond to relevant functions. 2] Move the check dc5_enabled in skl_set_power_well() to disable DC5 into gen9_disable_DC5 which is a more appropriate place. 3] Convert checks for 'pm.dc5_enabled' and 'pm.suspended' in skl_set_power_well() to warnings. However, removing them for now as they'll be included in a future patch asserting DC-state entry/exit criteria. 4] Enable DC5, only when CSR firmware is verified to be loaded. Create new structure to track 'enabled' and 'deferred' status of DC5. 5] Ensure runtime PM reference is obtained, if CSR is not loaded, to avoid entering runtime-suspend and release it when it's loaded. 6] Protect necessary CSR-related code with locks. 7] Move CSR-loading call to runtime PM initialization, as power domains needed to be accessed during deferred DC5-enabling, are not initialized earlier. v3: Rebase to latest. Modified as per review comments from Imre: 1] Use blocking wait for CSR-loading to finish to enable DC5 for simplicity, instead of deferring enabling DC5 until CSR is loaded. 2] Obtain runtime PM reference during CSR-loading initialization itself as deferred DC5- enabling is removed and release it at the end of CSR-loading functionality. 3] Revert calling CSR-loading functionality to the beginning of i915 driver-load functionality to avoid any delay in loading. 4] Define another variable to track whether CSR-loading failed and use it to avoid enabling DC5 if it's true. 5] Define CSR-load-status accessor functions for use later. v4: 1] Disable DC5 before enabling PG2 instead of after it. 2] DC5 was being mistaken enabled even when CSR-loading timed-out. Fix that. 3] Enable DC5-related functionality using a macro. 4] Remove dc5_enabled tracking variable and its use as it's not needed now. v5: 1] Mark CSR failed to load where necessary in finish_csr_load function. 2] Use mutex-protected accessor function to check if CSR loaded instead of directly accessing the variable. 3] Prefix csr_load_status_get/set function names with intel_. v6: rebase to latest. v7: Rebase on top of nightly (Damien) v8: Squashed the patch from Imre - added csr helper pointers to simplify the code. (Imre) v9: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v10: Added a enum for different csr states, suggested by Imre. (Animesh) v11: Based on review comments from Imre, Damien and Daniel following changes done - enum name chnaged to csr_state (singular form). - FW_UNINITIALIZED used as zeroth element in enum csr_state. - Prototype changed for helper function(set/get csr status), using enum csr_state instead of bool. v12: Based on review comment from Imre, introduced bool fw_loaded local to finish_csr_load() which helps calling once to set the csr status. The same flag used to fail RPM if find any issue during firmware loading. Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-17 22:16:16 +08:00
{
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
bool pg2_enabled = intel_display_power_well_is_enabled(dev_priv,
SKL_DISP_PW_2);
WARN_ONCE(pg2_enabled, "PG2 not disabled to enable DC5.\n");
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
WARN_ONCE((I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5),
"DC5 already programmed to be enabled.\n");
assert_rpm_wakelock_held(dev_priv);
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
assert_csr_loaded(dev_priv);
}
void gen9_enable_dc5(struct drm_i915_private *dev_priv)
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
{
assert_can_enable_dc5(dev_priv);
drm/i915/skl: Implement enable/disable for Display C5 state. This patch just implements the basic enable and disable functions of DC5 state which is needed for both SKL and BXT. Its important to load respective CSR program before calling enable, which anyways will happen as CSR program is executed during boot. DC5 is a power saving state where hardware dynamically disables power well 1 and the CDCLK PLL and saves the associated registers. DC5 can be entered when software allows it, power well 2 is disabled, and hardware detects that all pipes are disabled or pipe A is enabled with PSR active. Its better to configure display engine to have power well 2 disabled before getting into DC5 enable function. Hence rpm framework will have to ensure to check status of power well 2 before calling gen9_enable_dc5. Rather dc5 entry criteria should be decided based on power well 2 status. If disabled, then call gen9_enable_dc5. v2: Replace HAS_ with IS_ check as per Daniel's review comments v3: Cleared the bits dc5/dc6 enable of DC_STATE_EN register before setting them as per Satheesh's review comments. v4: call POSTING_READ for every write to a register to ensure that its written immediately. v5: Modified as per review comments from Imre. - Squashed register definitions into this patch. - Finetuned comments and functions. v6: Avoid redundant writes in gen9_set_dc_state_debugmask_memory_up function. v7: - Rebase to latest. - Move all runtime PM functions defined in intel_display.c to intel_runtime_pm.c. v8: Rebased to drm-intel-nightly. (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:09 +08:00
DRM_DEBUG_KMS("Enabling DC5\n");
/* Wa Display #1183: skl,kbl,cfl */
if (IS_GEN9_BC(dev_priv))
I915_WRITE(GEN8_CHICKEN_DCPR_1, I915_READ(GEN8_CHICKEN_DCPR_1) |
SKL_SELECT_ALTERNATE_DC_EXIT);
gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC5);
drm/i915/skl: Add DC5 Trigger Sequence Add triggers as per expectations mentioned in gen9_enable_dc5 and gen9_disable_dc5 patch. Also call POSTING_READ for every write to a register to ensure that its written immediately. v1: Remove POSTING_READ calls as they've already been added in previous patches. v2: Rebase to move all runtime pm specific changes to intel_runtime_pm.c file. Modified as per review comments from Imre: 1] Change variable name 'dc5_allowed' to 'dc5_enabled' to correspond to relevant functions. 2] Move the check dc5_enabled in skl_set_power_well() to disable DC5 into gen9_disable_DC5 which is a more appropriate place. 3] Convert checks for 'pm.dc5_enabled' and 'pm.suspended' in skl_set_power_well() to warnings. However, removing them for now as they'll be included in a future patch asserting DC-state entry/exit criteria. 4] Enable DC5, only when CSR firmware is verified to be loaded. Create new structure to track 'enabled' and 'deferred' status of DC5. 5] Ensure runtime PM reference is obtained, if CSR is not loaded, to avoid entering runtime-suspend and release it when it's loaded. 6] Protect necessary CSR-related code with locks. 7] Move CSR-loading call to runtime PM initialization, as power domains needed to be accessed during deferred DC5-enabling, are not initialized earlier. v3: Rebase to latest. Modified as per review comments from Imre: 1] Use blocking wait for CSR-loading to finish to enable DC5 for simplicity, instead of deferring enabling DC5 until CSR is loaded. 2] Obtain runtime PM reference during CSR-loading initialization itself as deferred DC5- enabling is removed and release it at the end of CSR-loading functionality. 3] Revert calling CSR-loading functionality to the beginning of i915 driver-load functionality to avoid any delay in loading. 4] Define another variable to track whether CSR-loading failed and use it to avoid enabling DC5 if it's true. 5] Define CSR-load-status accessor functions for use later. v4: 1] Disable DC5 before enabling PG2 instead of after it. 2] DC5 was being mistaken enabled even when CSR-loading timed-out. Fix that. 3] Enable DC5-related functionality using a macro. 4] Remove dc5_enabled tracking variable and its use as it's not needed now. v5: 1] Mark CSR failed to load where necessary in finish_csr_load function. 2] Use mutex-protected accessor function to check if CSR loaded instead of directly accessing the variable. 3] Prefix csr_load_status_get/set function names with intel_. v6: rebase to latest. v7: Rebase on top of nightly (Damien) v8: Squashed the patch from Imre - added csr helper pointers to simplify the code. (Imre) v9: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v10: Added a enum for different csr states, suggested by Imre. (Animesh) v11: Based on review comments from Imre, Damien and Daniel following changes done - enum name chnaged to csr_state (singular form). - FW_UNINITIALIZED used as zeroth element in enum csr_state. - Prototype changed for helper function(set/get csr status), using enum csr_state instead of bool. v12: Based on review comment from Imre, introduced bool fw_loaded local to finish_csr_load() which helps calling once to set the csr status. The same flag used to fail RPM if find any issue during firmware loading. Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-17 22:16:16 +08:00
}
2015-04-16 16:52:13 +08:00
static void assert_can_enable_dc6(struct drm_i915_private *dev_priv)
drm/i915/skl: Add DC6 Trigger sequence. Add triggers for DC6 as per details provided in skl_enable_dc6 and skl_disable_dc6 implementations. Also Call POSTING_READ for every write to a register to ensure it is written to immediately v1: Remove POSTING_READ and intel_prepare_ddi calls as they've been added in previous patches. v2: 1] Remove check for backlight disabled as it should be the case by that time. 2] Mark DC5 as disabled when enabling DC6. 3] Return from DC5-disabling function early if DC5 is already be disabled which can happen due to DC6-enabling earlier. 3] Ensure CSR firmware is loaded after resume from DC6 as corresponding memory contents won't be retained after runtime-suspend. 4] Ensure that CSR isn't identified as loaded before CSR-loading program is called during runtime-resume. v3: Rebase to latest Modified as per review comments from Imre and after discussion with Art: 1] DC6 should be preferably enabled when PG2 is disabled by SW as the check for PG1 being disabled is taken of by HW to enter DC6, and disabled when PG2 is enabled respectively. This helps save more power, especially in the case when display is disabled but GT is enabled. Accordingly, replacing DC5 trigger sequence with DC6 for SKL. 2] DC6 could be enabled from intel_runtime_suspend() function, if DC5 is already enabled. 3] Move CSR-load-status setting code from intel_runtime_suspend function to a new function. v4: 1] Enable/disable DC6 only when toggling the power-well using a newly defined macro ENABLE_DC6. v5: 1] Load CSR on system resume too as firmware may be lost on system suspend preventing enabling DC5, DC6. 2] DDI buffers shouldn't be programmed during driver-load/resume as it's already done during modeset initialization then and also that the encoder list is still uninitialized by then. Therefore, call intel_prepare_ddi function right after disabling DC6 but outside skl_disable_dc6 function and not during driver-load/resume. v6: 1] Rebase to latest. 2] Move SKL_ENABLE_DC6 macro definition from intel_display.c to intel_runtime_pm.c. v7: 1) Refactored the code for removing the warning got from checkpatch. 2) After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v8: - Reverted the changes done in v7. - Removed the condition check in skl_prepare_resune(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:11 +08:00
{
WARN_ONCE(I915_READ(UTIL_PIN_CTL) & UTIL_PIN_ENABLE,
"Backlight is not disabled.\n");
WARN_ONCE((I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC6),
"DC6 already programmed to be enabled.\n");
2015-04-16 16:52:13 +08:00
assert_csr_loaded(dev_priv);
}
2015-09-29 13:31:59 +08:00
void skl_enable_dc6(struct drm_i915_private *dev_priv)
2015-04-16 16:52:13 +08:00
{
assert_can_enable_dc6(dev_priv);
Implement enable/disable for Display C6 state This patch just implements the basic enable and disable functions of DC6 state which is needed for SKL platform. Its important to load SKL CSR program before calling enable. DC6 is a deeper power saving state where hardware dynamically disables power well 0 and saves the associated registers. DC6 can be entered when software allows it, the conditions for DC5 are met, and the PCU allows DC6. DC6 cannot be used if the backlight is being driven from the display utility pin. Its better to configure display engine to have power well 2 disabled before getting into DC6 enable function. Hence rpm framework will ensure to check status of power well 2 and DC5 before calling skl_enable_dc6. v2: Replace HAS_ with IS_ check as per Daniel's review comments v3: Cleared the bits dc5/dc6 enable of DC_STATE_EN register before setting them as per Satheesh's review comments. v4: No need to call gen9_disable_dc5 inside enable sequence of DC6, as its already take care above. v5: call POSTING_READ for every write to a register to ensure that its written immediately. Call intel_prepare_ddi during DC6 exit as it's required on low-power exit. v6: Protect DC6-enabling-disabling functionality with locks to synchronize with CSR-loading code. v7: Remove grabbing CSR-related mutex in skl_enable/disable_dc6 functions as deferred DC5-enabling functionality is now removed. v8: Remove 'Disabling DC5' from the debug comment during DC6 enabling as when DC6 is allowed, DC5 is not programmed at all. v9: - Rebase to latest. - Move all DC6-related functions from intel_display.c to intel_runtime_pm.c. v10: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:12 +08:00
DRM_DEBUG_KMS("Enabling DC6\n");
gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6);
drm/i915/skl: Add DC6 Trigger sequence. Add triggers for DC6 as per details provided in skl_enable_dc6 and skl_disable_dc6 implementations. Also Call POSTING_READ for every write to a register to ensure it is written to immediately v1: Remove POSTING_READ and intel_prepare_ddi calls as they've been added in previous patches. v2: 1] Remove check for backlight disabled as it should be the case by that time. 2] Mark DC5 as disabled when enabling DC6. 3] Return from DC5-disabling function early if DC5 is already be disabled which can happen due to DC6-enabling earlier. 3] Ensure CSR firmware is loaded after resume from DC6 as corresponding memory contents won't be retained after runtime-suspend. 4] Ensure that CSR isn't identified as loaded before CSR-loading program is called during runtime-resume. v3: Rebase to latest Modified as per review comments from Imre and after discussion with Art: 1] DC6 should be preferably enabled when PG2 is disabled by SW as the check for PG1 being disabled is taken of by HW to enter DC6, and disabled when PG2 is enabled respectively. This helps save more power, especially in the case when display is disabled but GT is enabled. Accordingly, replacing DC5 trigger sequence with DC6 for SKL. 2] DC6 could be enabled from intel_runtime_suspend() function, if DC5 is already enabled. 3] Move CSR-load-status setting code from intel_runtime_suspend function to a new function. v4: 1] Enable/disable DC6 only when toggling the power-well using a newly defined macro ENABLE_DC6. v5: 1] Load CSR on system resume too as firmware may be lost on system suspend preventing enabling DC5, DC6. 2] DDI buffers shouldn't be programmed during driver-load/resume as it's already done during modeset initialization then and also that the encoder list is still uninitialized by then. Therefore, call intel_prepare_ddi function right after disabling DC6 but outside skl_disable_dc6 function and not during driver-load/resume. v6: 1] Rebase to latest. 2] Move SKL_ENABLE_DC6 macro definition from intel_display.c to intel_runtime_pm.c. v7: 1) Refactored the code for removing the warning got from checkpatch. 2) After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v8: - Reverted the changes done in v7. - Removed the condition check in skl_prepare_resune(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:11 +08:00
}
2015-09-29 13:31:59 +08:00
void skl_disable_dc6(struct drm_i915_private *dev_priv)
drm/i915/skl: Add DC6 Trigger sequence. Add triggers for DC6 as per details provided in skl_enable_dc6 and skl_disable_dc6 implementations. Also Call POSTING_READ for every write to a register to ensure it is written to immediately v1: Remove POSTING_READ and intel_prepare_ddi calls as they've been added in previous patches. v2: 1] Remove check for backlight disabled as it should be the case by that time. 2] Mark DC5 as disabled when enabling DC6. 3] Return from DC5-disabling function early if DC5 is already be disabled which can happen due to DC6-enabling earlier. 3] Ensure CSR firmware is loaded after resume from DC6 as corresponding memory contents won't be retained after runtime-suspend. 4] Ensure that CSR isn't identified as loaded before CSR-loading program is called during runtime-resume. v3: Rebase to latest Modified as per review comments from Imre and after discussion with Art: 1] DC6 should be preferably enabled when PG2 is disabled by SW as the check for PG1 being disabled is taken of by HW to enter DC6, and disabled when PG2 is enabled respectively. This helps save more power, especially in the case when display is disabled but GT is enabled. Accordingly, replacing DC5 trigger sequence with DC6 for SKL. 2] DC6 could be enabled from intel_runtime_suspend() function, if DC5 is already enabled. 3] Move CSR-load-status setting code from intel_runtime_suspend function to a new function. v4: 1] Enable/disable DC6 only when toggling the power-well using a newly defined macro ENABLE_DC6. v5: 1] Load CSR on system resume too as firmware may be lost on system suspend preventing enabling DC5, DC6. 2] DDI buffers shouldn't be programmed during driver-load/resume as it's already done during modeset initialization then and also that the encoder list is still uninitialized by then. Therefore, call intel_prepare_ddi function right after disabling DC6 but outside skl_disable_dc6 function and not during driver-load/resume. v6: 1] Rebase to latest. 2] Move SKL_ENABLE_DC6 macro definition from intel_display.c to intel_runtime_pm.c. v7: 1) Refactored the code for removing the warning got from checkpatch. 2) After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v8: - Reverted the changes done in v7. - Removed the condition check in skl_prepare_resune(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:11 +08:00
{
Implement enable/disable for Display C6 state This patch just implements the basic enable and disable functions of DC6 state which is needed for SKL platform. Its important to load SKL CSR program before calling enable. DC6 is a deeper power saving state where hardware dynamically disables power well 0 and saves the associated registers. DC6 can be entered when software allows it, the conditions for DC5 are met, and the PCU allows DC6. DC6 cannot be used if the backlight is being driven from the display utility pin. Its better to configure display engine to have power well 2 disabled before getting into DC6 enable function. Hence rpm framework will ensure to check status of power well 2 and DC5 before calling skl_enable_dc6. v2: Replace HAS_ with IS_ check as per Daniel's review comments v3: Cleared the bits dc5/dc6 enable of DC_STATE_EN register before setting them as per Satheesh's review comments. v4: No need to call gen9_disable_dc5 inside enable sequence of DC6, as its already take care above. v5: call POSTING_READ for every write to a register to ensure that its written immediately. Call intel_prepare_ddi during DC6 exit as it's required on low-power exit. v6: Protect DC6-enabling-disabling functionality with locks to synchronize with CSR-loading code. v7: Remove grabbing CSR-related mutex in skl_enable/disable_dc6 functions as deferred DC5-enabling functionality is now removed. v8: Remove 'Disabling DC5' from the debug comment during DC6 enabling as when DC6 is allowed, DC5 is not programmed at all. v9: - Rebase to latest. - Move all DC6-related functions from intel_display.c to intel_runtime_pm.c. v10: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:12 +08:00
DRM_DEBUG_KMS("Disabling DC6\n");
/* Wa Display #1183: skl,kbl,cfl */
if (IS_GEN9_BC(dev_priv))
I915_WRITE(GEN8_CHICKEN_DCPR_1, I915_READ(GEN8_CHICKEN_DCPR_1) |
SKL_SELECT_ALTERNATE_DC_EXIT);
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
drm/i915/skl: Add DC6 Trigger sequence. Add triggers for DC6 as per details provided in skl_enable_dc6 and skl_disable_dc6 implementations. Also Call POSTING_READ for every write to a register to ensure it is written to immediately v1: Remove POSTING_READ and intel_prepare_ddi calls as they've been added in previous patches. v2: 1] Remove check for backlight disabled as it should be the case by that time. 2] Mark DC5 as disabled when enabling DC6. 3] Return from DC5-disabling function early if DC5 is already be disabled which can happen due to DC6-enabling earlier. 3] Ensure CSR firmware is loaded after resume from DC6 as corresponding memory contents won't be retained after runtime-suspend. 4] Ensure that CSR isn't identified as loaded before CSR-loading program is called during runtime-resume. v3: Rebase to latest Modified as per review comments from Imre and after discussion with Art: 1] DC6 should be preferably enabled when PG2 is disabled by SW as the check for PG1 being disabled is taken of by HW to enter DC6, and disabled when PG2 is enabled respectively. This helps save more power, especially in the case when display is disabled but GT is enabled. Accordingly, replacing DC5 trigger sequence with DC6 for SKL. 2] DC6 could be enabled from intel_runtime_suspend() function, if DC5 is already enabled. 3] Move CSR-load-status setting code from intel_runtime_suspend function to a new function. v4: 1] Enable/disable DC6 only when toggling the power-well using a newly defined macro ENABLE_DC6. v5: 1] Load CSR on system resume too as firmware may be lost on system suspend preventing enabling DC5, DC6. 2] DDI buffers shouldn't be programmed during driver-load/resume as it's already done during modeset initialization then and also that the encoder list is still uninitialized by then. Therefore, call intel_prepare_ddi function right after disabling DC6 but outside skl_disable_dc6 function and not during driver-load/resume. v6: 1] Rebase to latest. 2] Move SKL_ENABLE_DC6 macro definition from intel_display.c to intel_runtime_pm.c. v7: 1) Refactored the code for removing the warning got from checkpatch. 2) After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v8: - Reverted the changes done in v7. - Removed the condition check in skl_prepare_resune(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:11 +08:00
}
static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id id = power_well->id;
u32 mask = HSW_PWR_WELL_CTL_REQ(id);
u32 bios_req = I915_READ(HSW_PWR_WELL_CTL_BIOS(id));
/* Take over the request bit if set by BIOS. */
if (bios_req & mask) {
u32 drv_req = I915_READ(HSW_PWR_WELL_CTL_DRIVER(id));
if (!(drv_req & mask))
I915_WRITE(HSW_PWR_WELL_CTL_DRIVER(id), drv_req | mask);
I915_WRITE(HSW_PWR_WELL_CTL_BIOS(id), bios_req & ~mask);
}
}
static void bxt_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
bxt_ddi_phy_init(dev_priv, power_well->bxt.phy);
}
static void bxt_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
bxt_ddi_phy_uninit(dev_priv, power_well->bxt.phy);
}
static bool bxt_dpio_cmn_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
return bxt_ddi_phy_is_enabled(dev_priv, power_well->bxt.phy);
}
static void bxt_verify_ddi_phy_power_wells(struct drm_i915_private *dev_priv)
{
struct i915_power_well *power_well;
power_well = lookup_power_well(dev_priv, BXT_DPIO_CMN_A);
if (power_well->count > 0)
bxt_ddi_phy_verify_state(dev_priv, power_well->bxt.phy);
power_well = lookup_power_well(dev_priv, BXT_DPIO_CMN_BC);
if (power_well->count > 0)
bxt_ddi_phy_verify_state(dev_priv, power_well->bxt.phy);
if (IS_GEMINILAKE(dev_priv)) {
power_well = lookup_power_well(dev_priv, GLK_DPIO_CMN_C);
if (power_well->count > 0)
bxt_ddi_phy_verify_state(dev_priv, power_well->bxt.phy);
}
}
static bool gen9_dc_off_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
return (I915_READ(DC_STATE_EN) & DC_STATE_EN_UPTO_DC5_DC6_MASK) == 0;
}
static void gen9_assert_dbuf_enabled(struct drm_i915_private *dev_priv)
{
u32 tmp = I915_READ(DBUF_CTL);
WARN((tmp & (DBUF_POWER_STATE | DBUF_POWER_REQUEST)) !=
(DBUF_POWER_STATE | DBUF_POWER_REQUEST),
"Unexpected DBuf power power state (0x%08x)\n", tmp);
}
static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
struct intel_cdclk_state cdclk_state = {};
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
dev_priv->display.get_cdclk(dev_priv, &cdclk_state);
/* Can't read out voltage_level so can't use intel_cdclk_changed() */
WARN_ON(intel_cdclk_needs_modeset(&dev_priv->cdclk.hw, &cdclk_state));
gen9_assert_dbuf_enabled(dev_priv);
if (IS_GEN9_LP(dev_priv))
bxt_verify_ddi_phy_power_wells(dev_priv);
}
static void gen9_dc_off_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
if (!dev_priv->csr.dmc_payload)
return;
if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC6)
skl_enable_dc6(dev_priv);
else if (dev_priv->csr.allowed_dc_mask & DC_STATE_EN_UPTO_DC5)
gen9_enable_dc5(dev_priv);
}
static void i9xx_power_well_sync_hw_noop(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
}
static void i9xx_always_on_power_well_noop(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
}
static bool i9xx_always_on_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
return true;
}
static void i830_pipes_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
if ((I915_READ(PIPECONF(PIPE_A)) & PIPECONF_ENABLE) == 0)
i830_enable_pipe(dev_priv, PIPE_A);
if ((I915_READ(PIPECONF(PIPE_B)) & PIPECONF_ENABLE) == 0)
i830_enable_pipe(dev_priv, PIPE_B);
}
static void i830_pipes_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
i830_disable_pipe(dev_priv, PIPE_B);
i830_disable_pipe(dev_priv, PIPE_A);
}
static bool i830_pipes_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
return I915_READ(PIPECONF(PIPE_A)) & PIPECONF_ENABLE &&
I915_READ(PIPECONF(PIPE_B)) & PIPECONF_ENABLE;
}
static void i830_pipes_power_well_sync_hw(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
if (power_well->count > 0)
i830_pipes_power_well_enable(dev_priv, power_well);
else
i830_pipes_power_well_disable(dev_priv, power_well);
}
static void vlv_set_power_well(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well, bool enable)
{
enum i915_power_well_id power_well_id = power_well->id;
u32 mask;
u32 state;
u32 ctrl;
mask = PUNIT_PWRGT_MASK(power_well_id);
state = enable ? PUNIT_PWRGT_PWR_ON(power_well_id) :
PUNIT_PWRGT_PWR_GATE(power_well_id);
mutex_lock(&dev_priv->pcu_lock);
#define COND \
((vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask) == state)
if (COND)
goto out;
ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL);
ctrl &= ~mask;
ctrl |= state;
vlv_punit_write(dev_priv, PUNIT_REG_PWRGT_CTRL, ctrl);
if (wait_for(COND, 100))
DRM_ERROR("timeout setting power well state %08x (%08x)\n",
state,
vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL));
#undef COND
out:
mutex_unlock(&dev_priv->pcu_lock);
}
static void vlv_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
vlv_set_power_well(dev_priv, power_well, true);
}
static void vlv_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
vlv_set_power_well(dev_priv, power_well, false);
}
static bool vlv_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum i915_power_well_id power_well_id = power_well->id;
bool enabled = false;
u32 mask;
u32 state;
u32 ctrl;
mask = PUNIT_PWRGT_MASK(power_well_id);
ctrl = PUNIT_PWRGT_PWR_ON(power_well_id);
mutex_lock(&dev_priv->pcu_lock);
state = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_STATUS) & mask;
/*
* We only ever set the power-on and power-gate states, anything
* else is unexpected.
*/
WARN_ON(state != PUNIT_PWRGT_PWR_ON(power_well_id) &&
state != PUNIT_PWRGT_PWR_GATE(power_well_id));
if (state == ctrl)
enabled = true;
/*
* A transient state at this point would mean some unexpected party
* is poking at the power controls too.
*/
ctrl = vlv_punit_read(dev_priv, PUNIT_REG_PWRGT_CTRL) & mask;
WARN_ON(ctrl != state);
mutex_unlock(&dev_priv->pcu_lock);
return enabled;
}
static void vlv_init_display_clock_gating(struct drm_i915_private *dev_priv)
{
u32 val;
/*
* On driver load, a pipe may be active and driving a DSI display.
* Preserve DPOUNIT_CLOCK_GATE_DISABLE to avoid the pipe getting stuck
* (and never recovering) in this case. intel_dsi_post_disable() will
* clear it when we turn off the display.
*/
val = I915_READ(DSPCLK_GATE_D);
val &= DPOUNIT_CLOCK_GATE_DISABLE;
val |= VRHUNIT_CLOCK_GATE_DISABLE;
I915_WRITE(DSPCLK_GATE_D, val);
/*
* Disable trickle feed and enable pnd deadline calculation
*/
I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
I915_WRITE(CBR1_VLV, 0);
WARN_ON(dev_priv->rawclk_freq == 0);
I915_WRITE(RAWCLK_FREQ_VLV,
DIV_ROUND_CLOSEST(dev_priv->rawclk_freq, 1000));
}
static void vlv_display_power_well_init(struct drm_i915_private *dev_priv)
{
struct intel_encoder *encoder;
enum pipe pipe;
/*
* Enable the CRI clock source so we can get at the
* display and the reference clock for VGA
* hotplug / manual detection. Supposedly DSI also
* needs the ref clock up and running.
*
* CHV DPLL B/C have some issues if VGA mode is enabled.
*/
for_each_pipe(dev_priv, pipe) {
u32 val = I915_READ(DPLL(pipe));
val |= DPLL_REF_CLK_ENABLE_VLV | DPLL_VGA_MODE_DIS;
if (pipe != PIPE_A)
val |= DPLL_INTEGRATED_CRI_CLK_VLV;
I915_WRITE(DPLL(pipe), val);
}
vlv_init_display_clock_gating(dev_priv);
spin_lock_irq(&dev_priv->irq_lock);
valleyview_enable_display_irqs(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
/*
* During driver initialization/resume we can avoid restoring the
* part of the HW/SW state that will be inited anyway explicitly.
*/
if (dev_priv->power_domains.initializing)
return;
intel_hpd_init(dev_priv);
/* Re-enable the ADPA, if we have one */
for_each_intel_encoder(&dev_priv->drm, encoder) {
if (encoder->type == INTEL_OUTPUT_ANALOG)
intel_crt_reset(&encoder->base);
}
i915_redisable_vga_power_on(dev_priv);
intel_pps_unlock_regs_wa(dev_priv);
}
static void vlv_display_power_well_deinit(struct drm_i915_private *dev_priv)
{
spin_lock_irq(&dev_priv->irq_lock);
valleyview_disable_display_irqs(dev_priv);
spin_unlock_irq(&dev_priv->irq_lock);
/* make sure we're done processing display irqs */
synchronize_irq(dev_priv->drm.irq);
intel_power_sequencer_reset(dev_priv);
drm/i915: Enable polling when we don't have hpd Unfortunately, there's two situations where we lose hpd right now: - Runtime suspend - When we've shut off all of the power wells on Valleyview/Cherryview While it would be nice if this didn't cause issues, this has the ability to get us in some awkward states where a user won't be able to get their display to turn on. For instance; if we boot a Valleyview system without any monitors connected, it won't need any of it's power wells and thus shut them off. Since this causes us to lose HPD, this means that unless the user knows how to ssh into their machine and do a manual reprobe for monitors, none of the monitors they connect after booting will actually work. Eventually we should come up with a better fix then having to enable polling for this, since this makes rpm a lot less useful, but for now the infrastructure in i915 just isn't there yet to get hpd in these situations. Changes since v1: - Add comment explaining the addition of the if (!mode_config->poll_running) in intel_hpd_init() - Remove unneeded if (!dev->mode_config.poll_enabled) in i915_hpd_poll_init_work() - Call to drm_helper_hpd_irq_event() after we disable polling - Add cancel_work_sync() call to intel_hpd_cancel_work() Changes since v2: - Apparently dev->mode_config.poll_running doesn't actually reflect whether or not a poll is currently in progress, and is actually used for dynamic module paramter enabling/disabling. So now we instead keep track of our own poll_running variable in dev_priv->hotplug - Clean i915_hpd_poll_init_work() a little bit Changes since v3: - Remove the now-redundant connector loop in intel_hpd_init(), just rely on intel_hpd_poll_enable() for setting connector->polled correctly on each connector - Get rid of poll_running - Don't assign enabled in i915_hpd_poll_init_work before we actually lock dev->mode_config.mutex - Wrap enabled assignment in i915_hpd_poll_init_work() in READ_ONCE() for doc purposes - Do the same for dev_priv->hotplug.poll_enabled with WRITE_ONCE in intel_hpd_poll_enable() - Add some comments about racing not mattering in intel_hpd_poll_enable Changes since v4: - Rename intel_hpd_poll_enable() to intel_hpd_poll_init() - Drop the bool argument from intel_hpd_poll_init() - Remove redundant calls to intel_hpd_poll_init() - Rename poll_enable_work to poll_init_work - Add some kerneldoc for intel_hpd_poll_init() - Cross-reference intel_hpd_poll_init() in intel_hpd_init() - Just copy the loop from intel_hpd_init() in intel_hpd_poll_init() Changes since v5: - Minor kerneldoc nitpicks Cc: stable@vger.kernel.org Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Lyude <cpaul@redhat.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2016-06-22 05:03:44 +08:00
/* Prevent us from re-enabling polling on accident in late suspend */
if (!dev_priv->drm.dev->power.is_suspended)
intel_hpd_poll_init(dev_priv);
}
static void vlv_display_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DISP2D);
vlv_set_power_well(dev_priv, power_well, true);
vlv_display_power_well_init(dev_priv);
}
static void vlv_display_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DISP2D);
vlv_display_power_well_deinit(dev_priv);
vlv_set_power_well(dev_priv, power_well, false);
}
static void vlv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DPIO_CMN_BC);
/* since ref/cri clock was enabled */
udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
vlv_set_power_well(dev_priv, power_well, true);
/*
* From VLV2A0_DP_eDP_DPIO_driver_vbios_notes_10.docx -
* 6. De-assert cmn_reset/side_reset. Same as VLV X0.
* a. GUnit 0x2110 bit[0] set to 1 (def 0)
* b. The other bits such as sfr settings / modesel may all
* be set to 0.
*
* This should only be done on init and resume from S3 with
* both PLLs disabled, or we risk losing DPIO and PLL
* synchronization.
*/
I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) | DPIO_CMNRST);
}
static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum pipe pipe;
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DPIO_CMN_BC);
for_each_pipe(dev_priv, pipe)
assert_pll_disabled(dev_priv, pipe);
/* Assert common reset */
I915_WRITE(DPIO_CTL, I915_READ(DPIO_CTL) & ~DPIO_CMNRST);
vlv_set_power_well(dev_priv, power_well, false);
}
#define POWER_DOMAIN_MASK (GENMASK_ULL(POWER_DOMAIN_NUM - 1, 0))
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
static struct i915_power_well *
lookup_power_well(struct drm_i915_private *dev_priv,
enum i915_power_well_id power_well_id)
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
int i;
for (i = 0; i < power_domains->power_well_count; i++) {
struct i915_power_well *power_well;
power_well = &power_domains->power_wells[i];
if (power_well->id == power_well_id)
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
return power_well;
}
return NULL;
}
#define BITS_SET(val, bits) (((val) & (bits)) == (bits))
static void assert_chv_phy_status(struct drm_i915_private *dev_priv)
{
struct i915_power_well *cmn_bc =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
struct i915_power_well *cmn_d =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_D);
u32 phy_control = dev_priv->chv_phy_control;
u32 phy_status = 0;
u32 phy_status_mask = 0xffffffff;
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
/*
* The BIOS can leave the PHY is some weird state
* where it doesn't fully power down some parts.
* Disable the asserts until the PHY has been fully
* reset (ie. the power well has been disabled at
* least once).
*/
if (!dev_priv->chv_phy_assert[DPIO_PHY0])
phy_status_mask &= ~(PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH0) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 0) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 1) |
PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH1) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 0) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 1));
if (!dev_priv->chv_phy_assert[DPIO_PHY1])
phy_status_mask &= ~(PHY_STATUS_CMN_LDO(DPIO_PHY1, DPIO_CH0) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 0) |
PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 1));
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
if (cmn_bc->ops->is_enabled(dev_priv, cmn_bc)) {
phy_status |= PHY_POWERGOOD(DPIO_PHY0);
/* this assumes override is only used to enable lanes */
if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH0)) == 0)
phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH0);
if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH1)) == 0)
phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1);
/* CL1 is on whenever anything is on in either channel */
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH0) |
PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1)))
phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH0);
/*
* The DPLLB check accounts for the pipe B + port A usage
* with CL2 powered up but all the lanes in the second channel
* powered down.
*/
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY0, DPIO_CH1)) &&
(I915_READ(DPLL(PIPE_B)) & DPLL_VCO_ENABLE) == 0)
phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY0, DPIO_CH1);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY0, DPIO_CH0)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 0);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY0, DPIO_CH0)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH0, 1);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY0, DPIO_CH1)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 0);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY0, DPIO_CH1)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY0, DPIO_CH1, 1);
}
if (cmn_d->ops->is_enabled(dev_priv, cmn_d)) {
phy_status |= PHY_POWERGOOD(DPIO_PHY1);
/* this assumes override is only used to enable lanes */
if ((phy_control & PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY1, DPIO_CH0)) == 0)
phy_control |= PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY1, DPIO_CH0);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xf, DPIO_PHY1, DPIO_CH0)))
phy_status |= PHY_STATUS_CMN_LDO(DPIO_PHY1, DPIO_CH0);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0x3, DPIO_PHY1, DPIO_CH0)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 0);
if (BITS_SET(phy_control,
PHY_CH_POWER_DOWN_OVRD(0xc, DPIO_PHY1, DPIO_CH0)))
phy_status |= PHY_STATUS_SPLINE_LDO(DPIO_PHY1, DPIO_CH0, 1);
}
phy_status &= phy_status_mask;
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
/*
* The PHY may be busy with some initial calibration and whatnot,
* so the power state can take a while to actually change.
*/
if (intel_wait_for_register(dev_priv,
DISPLAY_PHY_STATUS,
phy_status_mask,
phy_status,
10))
DRM_ERROR("Unexpected PHY_STATUS 0x%08x, expected 0x%08x (PHY_CONTROL=0x%08x)\n",
I915_READ(DISPLAY_PHY_STATUS) & phy_status_mask,
phy_status, dev_priv->chv_phy_control);
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
}
#undef BITS_SET
static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum dpio_phy phy;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
enum pipe pipe;
uint32_t tmp;
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DPIO_CMN_BC &&
power_well->id != PUNIT_POWER_WELL_DPIO_CMN_D);
if (power_well->id == PUNIT_POWER_WELL_DPIO_CMN_BC) {
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
pipe = PIPE_A;
phy = DPIO_PHY0;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
} else {
pipe = PIPE_C;
phy = DPIO_PHY1;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
}
/* since ref/cri clock was enabled */
udelay(1); /* >10ns for cmnreset, >0ns for sidereset */
vlv_set_power_well(dev_priv, power_well, true);
/* Poll for phypwrgood signal */
if (intel_wait_for_register(dev_priv,
DISPLAY_PHY_STATUS,
PHY_POWERGOOD(phy),
PHY_POWERGOOD(phy),
1))
DRM_ERROR("Display PHY %d is not power up\n", phy);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
mutex_lock(&dev_priv->sb_lock);
/* Enable dynamic power down */
tmp = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW28);
tmp |= DPIO_DYNPWRDOWNEN_CH0 | DPIO_CL1POWERDOWNEN |
DPIO_SUS_CLK_CONFIG_GATE_CLKREQ;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW28, tmp);
if (power_well->id == PUNIT_POWER_WELL_DPIO_CMN_BC) {
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
tmp = vlv_dpio_read(dev_priv, pipe, _CHV_CMN_DW6_CH1);
tmp |= DPIO_DYNPWRDOWNEN_CH1;
vlv_dpio_write(dev_priv, pipe, _CHV_CMN_DW6_CH1, tmp);
} else {
/*
* Force the non-existing CL2 off. BXT does this
* too, so maybe it saves some power even though
* CL2 doesn't exist?
*/
tmp = vlv_dpio_read(dev_priv, pipe, CHV_CMN_DW30);
tmp |= DPIO_CL2_LDOFUSE_PWRENB;
vlv_dpio_write(dev_priv, pipe, CHV_CMN_DW30, tmp);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
}
mutex_unlock(&dev_priv->sb_lock);
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
dev_priv->chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(phy);
I915_WRITE(DISPLAY_PHY_CONTROL, dev_priv->chv_phy_control);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
DRM_DEBUG_KMS("Enabled DPIO PHY%d (PHY_CONTROL=0x%08x)\n",
phy, dev_priv->chv_phy_control);
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
assert_chv_phy_status(dev_priv);
}
static void chv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum dpio_phy phy;
WARN_ON_ONCE(power_well->id != PUNIT_POWER_WELL_DPIO_CMN_BC &&
power_well->id != PUNIT_POWER_WELL_DPIO_CMN_D);
if (power_well->id == PUNIT_POWER_WELL_DPIO_CMN_BC) {
phy = DPIO_PHY0;
assert_pll_disabled(dev_priv, PIPE_A);
assert_pll_disabled(dev_priv, PIPE_B);
} else {
phy = DPIO_PHY1;
assert_pll_disabled(dev_priv, PIPE_C);
}
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
dev_priv->chv_phy_control &= ~PHY_COM_LANE_RESET_DEASSERT(phy);
I915_WRITE(DISPLAY_PHY_CONTROL, dev_priv->chv_phy_control);
vlv_set_power_well(dev_priv, power_well, false);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
DRM_DEBUG_KMS("Disabled DPIO PHY%d (PHY_CONTROL=0x%08x)\n",
phy, dev_priv->chv_phy_control);
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
/* PHY is fully reset now, so we can enable the PHY state asserts */
dev_priv->chv_phy_assert[phy] = true;
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
assert_chv_phy_status(dev_priv);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
}
static void assert_chv_phy_powergate(struct drm_i915_private *dev_priv, enum dpio_phy phy,
enum dpio_channel ch, bool override, unsigned int mask)
{
enum pipe pipe = phy == DPIO_PHY0 ? PIPE_A : PIPE_C;
u32 reg, val, expected, actual;
/*
* The BIOS can leave the PHY is some weird state
* where it doesn't fully power down some parts.
* Disable the asserts until the PHY has been fully
* reset (ie. the power well has been disabled at
* least once).
*/
if (!dev_priv->chv_phy_assert[phy])
return;
if (ch == DPIO_CH0)
reg = _CHV_CMN_DW0_CH0;
else
reg = _CHV_CMN_DW6_CH1;
mutex_lock(&dev_priv->sb_lock);
val = vlv_dpio_read(dev_priv, pipe, reg);
mutex_unlock(&dev_priv->sb_lock);
/*
* This assumes !override is only used when the port is disabled.
* All lanes should power down even without the override when
* the port is disabled.
*/
if (!override || mask == 0xf) {
expected = DPIO_ALLDL_POWERDOWN | DPIO_ANYDL_POWERDOWN;
/*
* If CH1 common lane is not active anymore
* (eg. for pipe B DPLL) the entire channel will
* shut down, which causes the common lane registers
* to read as 0. That means we can't actually check
* the lane power down status bits, but as the entire
* register reads as 0 it's a good indication that the
* channel is indeed entirely powered down.
*/
if (ch == DPIO_CH1 && val == 0)
expected = 0;
} else if (mask != 0x0) {
expected = DPIO_ANYDL_POWERDOWN;
} else {
expected = 0;
}
if (ch == DPIO_CH0)
actual = val >> DPIO_ANYDL_POWERDOWN_SHIFT_CH0;
else
actual = val >> DPIO_ANYDL_POWERDOWN_SHIFT_CH1;
actual &= DPIO_ALLDL_POWERDOWN | DPIO_ANYDL_POWERDOWN;
WARN(actual != expected,
"Unexpected DPIO lane power down: all %d, any %d. Expected: all %d, any %d. (0x%x = 0x%08x)\n",
!!(actual & DPIO_ALLDL_POWERDOWN), !!(actual & DPIO_ANYDL_POWERDOWN),
!!(expected & DPIO_ALLDL_POWERDOWN), !!(expected & DPIO_ANYDL_POWERDOWN),
reg, val);
}
drm/i915: Trick CL2 into life on CHV when using pipe B with port B Normmally the common lane in a PHY channel gets powered up when some of the data lanes get powered up. But when we're driving port B with pipe B we don't want to enabled any of the data lanes, and just want the DPLL in the common lane to be active. To make that happens we have to temporarily enable some data lanes after which we can access the DPLL registers in the common lane. Once the pipe is up and running we can drop the power override on the data lanes allowing them to shut down. From this point forward the common lane will in fact stay powered on until the data lanes in the other channel get powered down. Ville's extended explanation from the review thread: On Wed, Aug 19, 2015 at 07:47:41AM +0530, Deepak wrote: > One Q, why only for port B? Port C is also in same common lane right? Port B is in the first PHY channel which also houses CL1. CL1 always powers up whenever any lanes in either PHY channel are powered up. CL2 only powers up if lanes in the second channel (ie. the one with port C) powers up. So in this scenario (pipe B->port B) we want the DPLL from CL2, but ideally we only want to power up the lanes for port B. Powering up port B lanes will only power up CL1, but as we need CL2 instead we need to, temporarily, power up some lanes in port C as well. Crossing the streams the other way (pipe A->port C) is not a problem since CL1 powers up whenever anything else powers up. So powering up some port C lanes is enough on its own to make the CL1 DPLL operational, even though CL1 and the lanes live in separate channels. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> [danvet: Amend commit message with extended explanation.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:55 +08:00
bool chv_phy_powergate_ch(struct drm_i915_private *dev_priv, enum dpio_phy phy,
enum dpio_channel ch, bool override)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
bool was_override;
mutex_lock(&power_domains->lock);
was_override = dev_priv->chv_phy_control & PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
if (override == was_override)
goto out;
if (override)
dev_priv->chv_phy_control |= PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
else
dev_priv->chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
I915_WRITE(DISPLAY_PHY_CONTROL, dev_priv->chv_phy_control);
DRM_DEBUG_KMS("Power gating DPIO PHY%d CH%d (DPIO_PHY_CONTROL=0x%08x)\n",
phy, ch, dev_priv->chv_phy_control);
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
assert_chv_phy_status(dev_priv);
drm/i915: Trick CL2 into life on CHV when using pipe B with port B Normmally the common lane in a PHY channel gets powered up when some of the data lanes get powered up. But when we're driving port B with pipe B we don't want to enabled any of the data lanes, and just want the DPLL in the common lane to be active. To make that happens we have to temporarily enable some data lanes after which we can access the DPLL registers in the common lane. Once the pipe is up and running we can drop the power override on the data lanes allowing them to shut down. From this point forward the common lane will in fact stay powered on until the data lanes in the other channel get powered down. Ville's extended explanation from the review thread: On Wed, Aug 19, 2015 at 07:47:41AM +0530, Deepak wrote: > One Q, why only for port B? Port C is also in same common lane right? Port B is in the first PHY channel which also houses CL1. CL1 always powers up whenever any lanes in either PHY channel are powered up. CL2 only powers up if lanes in the second channel (ie. the one with port C) powers up. So in this scenario (pipe B->port B) we want the DPLL from CL2, but ideally we only want to power up the lanes for port B. Powering up port B lanes will only power up CL1, but as we need CL2 instead we need to, temporarily, power up some lanes in port C as well. Crossing the streams the other way (pipe A->port C) is not a problem since CL1 powers up whenever anything else powers up. So powering up some port C lanes is enough on its own to make the CL1 DPLL operational, even though CL1 and the lanes live in separate channels. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> [danvet: Amend commit message with extended explanation.] Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:55 +08:00
out:
mutex_unlock(&power_domains->lock);
return was_override;
}
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
void chv_phy_powergate_lanes(struct intel_encoder *encoder,
bool override, unsigned int mask)
{
struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
struct i915_power_domains *power_domains = &dev_priv->power_domains;
enum dpio_phy phy = vlv_dport_to_phy(enc_to_dig_port(&encoder->base));
enum dpio_channel ch = vlv_dport_to_channel(enc_to_dig_port(&encoder->base));
mutex_lock(&power_domains->lock);
dev_priv->chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD(0xf, phy, ch);
dev_priv->chv_phy_control |= PHY_CH_POWER_DOWN_OVRD(mask, phy, ch);
if (override)
dev_priv->chv_phy_control |= PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
else
dev_priv->chv_phy_control &= ~PHY_CH_POWER_DOWN_OVRD_EN(phy, ch);
I915_WRITE(DISPLAY_PHY_CONTROL, dev_priv->chv_phy_control);
DRM_DEBUG_KMS("Power gating DPIO PHY%d CH%d lanes 0x%x (PHY_CONTROL=0x%08x)\n",
phy, ch, mask, dev_priv->chv_phy_control);
drm/i915: Add CHV PHY LDO power sanity checks At various points when changing the DPIO lane/phy power states, construct an expected value of the DISPLAY_PHY_STATUS register and compare it with the real thing. To construct the expected value we look at our shadow PHY_CONTROL register value (which should match what we've just written to the hardware), and we also need to look at the actual state of the cmn power wells as a disabled power well causes the relevant LDO status to be reported as 'on' in DISPLAY_PHY_STATUS. When initially powering up the PHY it performs various internal calibrations for which it fully powers up. That means that if we check for the expetected power state immediately upon releasing cmnreset we would get the occasional false positive. But we can of course poll until the expected value appears. It shouldn't be too long so this shouldn't make modesets substantially longer. One extra complication is introduced when we cross the streams, ie. drive port B with pipe B. In this case we trick CL2 (where the DPLL lives) into life by temporaily powering up the lanes in the second channel, and once the pipe is up and runnign we release the lane power override. At that point the power state of CL2 has somehow gotten entangled with the power state of the first channel. That means that constructing the expected DISPLAY_PHY_STATUS value is a bit tricky since based on the lane power states in the second channel, CL2 should also be powered down. But we can use the DPLL enable bit to determine when CL2 should be alive even if the lanes are powered down. However the power state of CL2 isn't actually tied in with the DPLL state, but to the state of the lanes in first channel, so we have to avoid checking the expected state between shutting down the DPLL and powering down the lanes in the first channel. So no calling assert_chv_phy_status() before the DISPLAY_PHY_CONTROL write in chv_phy_powergate_lanes(), but after the write is a safe time to check. Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:46:01 +08:00
assert_chv_phy_status(dev_priv);
assert_chv_phy_powergate(dev_priv, phy, ch, override, mask);
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
mutex_unlock(&power_domains->lock);
}
static bool chv_pipe_power_well_enabled(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
enum pipe pipe = PIPE_A;
bool enabled;
u32 state, ctrl;
mutex_lock(&dev_priv->pcu_lock);
state = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe);
/*
* We only ever set the power-on and power-gate states, anything
* else is unexpected.
*/
WARN_ON(state != DP_SSS_PWR_ON(pipe) && state != DP_SSS_PWR_GATE(pipe));
enabled = state == DP_SSS_PWR_ON(pipe);
/*
* A transient state at this point would mean some unexpected party
* is poking at the power controls too.
*/
ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSC_MASK(pipe);
WARN_ON(ctrl << 16 != state);
mutex_unlock(&dev_priv->pcu_lock);
return enabled;
}
static void chv_set_pipe_power_well(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well,
bool enable)
{
enum pipe pipe = PIPE_A;
u32 state;
u32 ctrl;
state = enable ? DP_SSS_PWR_ON(pipe) : DP_SSS_PWR_GATE(pipe);
mutex_lock(&dev_priv->pcu_lock);
#define COND \
((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) & DP_SSS_MASK(pipe)) == state)
if (COND)
goto out;
ctrl = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
ctrl &= ~DP_SSC_MASK(pipe);
ctrl |= enable ? DP_SSC_PWR_ON(pipe) : DP_SSC_PWR_GATE(pipe);
vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, ctrl);
if (wait_for(COND, 100))
DRM_ERROR("timeout setting power well state %08x (%08x)\n",
state,
vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ));
#undef COND
out:
mutex_unlock(&dev_priv->pcu_lock);
}
static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN_ON_ONCE(power_well->id != CHV_DISP_PW_PIPE_A);
chv_set_pipe_power_well(dev_priv, power_well, true);
vlv_display_power_well_init(dev_priv);
}
static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
struct i915_power_well *power_well)
{
WARN_ON_ONCE(power_well->id != CHV_DISP_PW_PIPE_A);
vlv_display_power_well_deinit(dev_priv);
chv_set_pipe_power_well(dev_priv, power_well, false);
}
static void
__intel_display_power_get_domain(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *power_well;
for_each_power_domain_well(dev_priv, power_well, BIT_ULL(domain))
intel_power_well_get(dev_priv, power_well);
power_domains->domain_use_count[domain]++;
}
/**
* intel_display_power_get - grab a power domain reference
* @dev_priv: i915 device instance
* @domain: power domain to reference
*
* This function grabs a power domain reference for @domain and ensures that the
* power domain and all its parents are powered up. Therefore users should only
* grab a reference to the innermost power domain they need.
*
* Any power domain reference obtained by this function must have a symmetric
* call to intel_display_power_put() to release the reference again.
*/
void intel_display_power_get(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
intel_runtime_pm_get(dev_priv);
mutex_lock(&power_domains->lock);
__intel_display_power_get_domain(dev_priv, domain);
mutex_unlock(&power_domains->lock);
}
/**
* intel_display_power_get_if_enabled - grab a reference for an enabled display power domain
* @dev_priv: i915 device instance
* @domain: power domain to reference
*
* This function grabs a power domain reference for @domain and ensures that the
* power domain and all its parents are powered up. Therefore users should only
* grab a reference to the innermost power domain they need.
*
* Any power domain reference obtained by this function must have a symmetric
* call to intel_display_power_put() to release the reference again.
*/
bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
bool is_enabled;
if (!intel_runtime_pm_get_if_in_use(dev_priv))
return false;
mutex_lock(&power_domains->lock);
if (__intel_display_power_is_enabled(dev_priv, domain)) {
__intel_display_power_get_domain(dev_priv, domain);
is_enabled = true;
} else {
is_enabled = false;
}
mutex_unlock(&power_domains->lock);
if (!is_enabled)
intel_runtime_pm_put(dev_priv);
return is_enabled;
}
/**
* intel_display_power_put - release a power domain reference
* @dev_priv: i915 device instance
* @domain: power domain to reference
*
* This function drops the power domain reference obtained by
* intel_display_power_get() and might power down the corresponding hardware
* block right away if this is the last reference.
*/
void intel_display_power_put(struct drm_i915_private *dev_priv,
enum intel_display_power_domain domain)
{
struct i915_power_domains *power_domains;
struct i915_power_well *power_well;
power_domains = &dev_priv->power_domains;
mutex_lock(&power_domains->lock);
WARN(!power_domains->domain_use_count[domain],
"Use count on domain %s is already zero\n",
intel_display_power_domain_str(domain));
power_domains->domain_use_count[domain]--;
for_each_power_domain_well_rev(dev_priv, power_well, BIT_ULL(domain))
intel_power_well_put(dev_priv, power_well);
mutex_unlock(&power_domains->lock);
intel_runtime_pm_put(dev_priv);
}
#define I830_PIPES_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PIPE_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DISPLAY_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PIPE_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DSI) | \
BIT_ULL(POWER_DOMAIN_PORT_CRT) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_GMBUS) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DPIO_CMN_BC_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_CRT) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CHV_DISPLAY_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PIPE_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DSI) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_AUX_D) | \
BIT_ULL(POWER_DOMAIN_GMBUS) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CHV_DPIO_CMN_BC_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CHV_DPIO_CMN_D_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_D) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define HSW_DISPLAY_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_CRT) | /* DDI E */ \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define BDW_DISPLAY_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_CRT) | /* DDI E */ \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_POWERWELL_2_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_E_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_AUX_D) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_DDI_IO_A_E_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_A_IO) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_E_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_DDI_IO_B_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_DDI_IO_C_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_DDI_IO_D_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define SKL_DISPLAY_DC_OFF_POWER_DOMAINS ( \
SKL_DISPLAY_POWERWELL_2_POWER_DOMAINS | \
drm/i915: Restore GT performance in headless mode with DMC loaded It seems that the DMC likes to transition between the DC states a lot when there are no connected displays (no active power domains) during command submission. This activity on DC states has a negative impact on the performance of the chip with huge latencies observed in the interrupt handlers and elsewhere. Simple tests like igt/gem_latency -n 0 are slowed down by a factor of eight. Work around it by introducing a new power domain named, POWER_DOMAIN_GT_IRQ, associtated with the "DC off" power well, which is held for the duration of command submission activity. CNL has the same problem which will be addressed as a follow-up. Doing that requires a fix for a DC6 context corruption problem in the CNL DMC firmware which is yet to be released. v2: * Add commit text as comment in i915_gem_mark_busy. (Chris Wilson) * Protect macro body with braces. (Jani Nikula) v3: * Add dedicated power domain for clarity. (Chris, Imre) * Commit message and comment text updates. * Apply to all big-core GEN9 parts apart for Skylake which is pending DMC firmware release. v4: * Power domain should be inner to device runtime pm. (Chris) * Simplify NEEDS_CSR_GT_PERF_WA macro. (Chris) * Handle async DMC loading by moving the GT_IRQ power domain logic into intel_runtime_pm. (Daniel, Chris) * Include small core GEN9 as well. (Imre) v5 * Special handling for async DMC load is not needed since on failure the power domain reference is kept permanently taken. (Imre) v6: * Drop the NEEDS_CSR_GT_PERF_WA macro since all firmwares have now been deployed. (Imre, Chris) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100572 Testcase: igt/gem_exec_nop/headless Cc: Imre Deak <imre.deak@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> (v2) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v5) Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> [Imre: Add note about applying the WA on CNL as a follow-up] Signed-off-by: Imre Deak <imre.deak@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171205132854.26380-1-tvrtko.ursulin@linux.intel.com
2017-12-05 21:28:54 +08:00
BIT_ULL(POWER_DOMAIN_GT_IRQ) | \
BIT_ULL(POWER_DOMAIN_MODESET) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define BXT_DISPLAY_POWERWELL_2_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define BXT_DISPLAY_DC_OFF_POWER_DOMAINS ( \
BXT_DISPLAY_POWERWELL_2_POWER_DOMAINS | \
drm/i915: Restore GT performance in headless mode with DMC loaded It seems that the DMC likes to transition between the DC states a lot when there are no connected displays (no active power domains) during command submission. This activity on DC states has a negative impact on the performance of the chip with huge latencies observed in the interrupt handlers and elsewhere. Simple tests like igt/gem_latency -n 0 are slowed down by a factor of eight. Work around it by introducing a new power domain named, POWER_DOMAIN_GT_IRQ, associtated with the "DC off" power well, which is held for the duration of command submission activity. CNL has the same problem which will be addressed as a follow-up. Doing that requires a fix for a DC6 context corruption problem in the CNL DMC firmware which is yet to be released. v2: * Add commit text as comment in i915_gem_mark_busy. (Chris Wilson) * Protect macro body with braces. (Jani Nikula) v3: * Add dedicated power domain for clarity. (Chris, Imre) * Commit message and comment text updates. * Apply to all big-core GEN9 parts apart for Skylake which is pending DMC firmware release. v4: * Power domain should be inner to device runtime pm. (Chris) * Simplify NEEDS_CSR_GT_PERF_WA macro. (Chris) * Handle async DMC loading by moving the GT_IRQ power domain logic into intel_runtime_pm. (Daniel, Chris) * Include small core GEN9 as well. (Imre) v5 * Special handling for async DMC load is not needed since on failure the power domain reference is kept permanently taken. (Imre) v6: * Drop the NEEDS_CSR_GT_PERF_WA macro since all firmwares have now been deployed. (Imre, Chris) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100572 Testcase: igt/gem_exec_nop/headless Cc: Imre Deak <imre.deak@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> (v2) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v5) Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> [Imre: Add note about applying the WA on CNL as a follow-up] Signed-off-by: Imre Deak <imre.deak@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171205132854.26380-1-tvrtko.ursulin@linux.intel.com
2017-12-05 21:28:54 +08:00
BIT_ULL(POWER_DOMAIN_GT_IRQ) | \
BIT_ULL(POWER_DOMAIN_MODESET) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_GMBUS) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define BXT_DPIO_CMN_A_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_A_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define BXT_DPIO_CMN_BC_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_POWERWELL_2_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_DDI_IO_A_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_A_IO))
#define GLK_DISPLAY_DDI_IO_B_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_IO))
#define GLK_DISPLAY_DDI_IO_C_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_IO))
#define GLK_DPIO_CMN_A_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_A_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DPIO_CMN_B_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DPIO_CMN_C_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_AUX_A_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_AUX_B_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_AUX_C_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define GLK_DISPLAY_DC_OFF_POWER_DOMAINS ( \
GLK_DISPLAY_POWERWELL_2_POWER_DOMAINS | \
drm/i915: Restore GT performance in headless mode with DMC loaded It seems that the DMC likes to transition between the DC states a lot when there are no connected displays (no active power domains) during command submission. This activity on DC states has a negative impact on the performance of the chip with huge latencies observed in the interrupt handlers and elsewhere. Simple tests like igt/gem_latency -n 0 are slowed down by a factor of eight. Work around it by introducing a new power domain named, POWER_DOMAIN_GT_IRQ, associtated with the "DC off" power well, which is held for the duration of command submission activity. CNL has the same problem which will be addressed as a follow-up. Doing that requires a fix for a DC6 context corruption problem in the CNL DMC firmware which is yet to be released. v2: * Add commit text as comment in i915_gem_mark_busy. (Chris Wilson) * Protect macro body with braces. (Jani Nikula) v3: * Add dedicated power domain for clarity. (Chris, Imre) * Commit message and comment text updates. * Apply to all big-core GEN9 parts apart for Skylake which is pending DMC firmware release. v4: * Power domain should be inner to device runtime pm. (Chris) * Simplify NEEDS_CSR_GT_PERF_WA macro. (Chris) * Handle async DMC loading by moving the GT_IRQ power domain logic into intel_runtime_pm. (Daniel, Chris) * Include small core GEN9 as well. (Imre) v5 * Special handling for async DMC load is not needed since on failure the power domain reference is kept permanently taken. (Imre) v6: * Drop the NEEDS_CSR_GT_PERF_WA macro since all firmwares have now been deployed. (Imre, Chris) Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100572 Testcase: igt/gem_exec_nop/headless Cc: Imre Deak <imre.deak@intel.com> Acked-by: Chris Wilson <chris@chris-wilson.co.uk> (v2) Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v5) Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> [Imre: Add note about applying the WA on CNL as a follow-up] Signed-off-by: Imre Deak <imre.deak@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20171205132854.26380-1-tvrtko.ursulin@linux.intel.com
2017-12-05 21:28:54 +08:00
BIT_ULL(POWER_DOMAIN_GT_IRQ) | \
BIT_ULL(POWER_DOMAIN_MODESET) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_GMBUS) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_POWERWELL_2_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_TRANSCODER_A) | \
BIT_ULL(POWER_DOMAIN_PIPE_B) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_B) | \
BIT_ULL(POWER_DOMAIN_PIPE_C) | \
BIT_ULL(POWER_DOMAIN_TRANSCODER_C) | \
BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) | \
BIT_ULL(POWER_DOMAIN_PORT_DDI_F_LANES) | \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_AUX_D) | \
BIT_ULL(POWER_DOMAIN_AUX_F) | \
BIT_ULL(POWER_DOMAIN_AUDIO) | \
BIT_ULL(POWER_DOMAIN_VGA) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DDI_A_IO_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_A_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DDI_B_IO_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_B_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DDI_C_IO_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_C_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DDI_D_IO_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_D_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_AUX_A_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_AUX_B_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_B) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_AUX_C_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_C) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_AUX_D_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_D) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_AUX_F_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_AUX_F) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DDI_F_IO_POWER_DOMAINS ( \
BIT_ULL(POWER_DOMAIN_PORT_DDI_F_IO) | \
BIT_ULL(POWER_DOMAIN_INIT))
#define CNL_DISPLAY_DC_OFF_POWER_DOMAINS ( \
CNL_DISPLAY_POWERWELL_2_POWER_DOMAINS | \
BIT_ULL(POWER_DOMAIN_GT_IRQ) | \
BIT_ULL(POWER_DOMAIN_MODESET) | \
BIT_ULL(POWER_DOMAIN_AUX_A) | \
BIT_ULL(POWER_DOMAIN_INIT))
static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = i9xx_always_on_power_well_noop,
.disable = i9xx_always_on_power_well_noop,
.is_enabled = i9xx_always_on_power_well_enabled,
};
static const struct i915_power_well_ops chv_pipe_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = chv_pipe_power_well_enable,
.disable = chv_pipe_power_well_disable,
.is_enabled = chv_pipe_power_well_enabled,
};
static const struct i915_power_well_ops chv_dpio_cmn_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = chv_dpio_cmn_power_well_enable,
.disable = chv_dpio_cmn_power_well_disable,
.is_enabled = vlv_power_well_enabled,
};
static struct i915_power_well i9xx_always_on_power_well[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
};
static const struct i915_power_well_ops i830_pipes_power_well_ops = {
.sync_hw = i830_pipes_power_well_sync_hw,
.enable = i830_pipes_power_well_enable,
.disable = i830_pipes_power_well_disable,
.is_enabled = i830_pipes_power_well_enabled,
};
static struct i915_power_well i830_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "pipes",
.domains = I830_PIPES_POWER_DOMAINS,
.ops = &i830_pipes_power_well_ops,
.id = I830_DISP_PW_PIPES,
},
};
static const struct i915_power_well_ops hsw_power_well_ops = {
.sync_hw = hsw_power_well_sync_hw,
.enable = hsw_power_well_enable,
.disable = hsw_power_well_disable,
.is_enabled = hsw_power_well_enabled,
};
static const struct i915_power_well_ops gen9_dc_off_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = gen9_dc_off_power_well_enable,
.disable = gen9_dc_off_power_well_disable,
.is_enabled = gen9_dc_off_power_well_enabled,
};
static const struct i915_power_well_ops bxt_dpio_cmn_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = bxt_dpio_cmn_power_well_enable,
.disable = bxt_dpio_cmn_power_well_disable,
.is_enabled = bxt_dpio_cmn_power_well_enabled,
};
static struct i915_power_well hsw_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "display",
.domains = HSW_DISPLAY_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = HSW_DISP_PW_GLOBAL,
{
.hsw.has_vga = true,
},
},
};
static struct i915_power_well bdw_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "display",
.domains = BDW_DISPLAY_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = HSW_DISP_PW_GLOBAL,
{
.hsw.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
.hsw.has_vga = true,
},
},
};
static const struct i915_power_well_ops vlv_display_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = vlv_display_power_well_enable,
.disable = vlv_display_power_well_disable,
.is_enabled = vlv_power_well_enabled,
};
static const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = vlv_dpio_cmn_power_well_enable,
.disable = vlv_dpio_cmn_power_well_disable,
.is_enabled = vlv_power_well_enabled,
};
static const struct i915_power_well_ops vlv_dpio_power_well_ops = {
.sync_hw = i9xx_power_well_sync_hw_noop,
.enable = vlv_power_well_enable,
.disable = vlv_power_well_disable,
.is_enabled = vlv_power_well_enabled,
};
static struct i915_power_well vlv_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "display",
.domains = VLV_DISPLAY_POWER_DOMAINS,
.id = PUNIT_POWER_WELL_DISP2D,
.ops = &vlv_display_power_well_ops,
},
{
.name = "dpio-tx-b-01",
.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
.ops = &vlv_dpio_power_well_ops,
.id = PUNIT_POWER_WELL_DPIO_TX_B_LANES_01,
},
{
.name = "dpio-tx-b-23",
.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
.ops = &vlv_dpio_power_well_ops,
.id = PUNIT_POWER_WELL_DPIO_TX_B_LANES_23,
},
{
.name = "dpio-tx-c-01",
.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
.ops = &vlv_dpio_power_well_ops,
.id = PUNIT_POWER_WELL_DPIO_TX_C_LANES_01,
},
{
.name = "dpio-tx-c-23",
.domains = VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS |
VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS,
.ops = &vlv_dpio_power_well_ops,
.id = PUNIT_POWER_WELL_DPIO_TX_C_LANES_23,
},
{
.name = "dpio-common",
.domains = VLV_DPIO_CMN_BC_POWER_DOMAINS,
.id = PUNIT_POWER_WELL_DPIO_CMN_BC,
.ops = &vlv_dpio_cmn_power_well_ops,
},
};
static struct i915_power_well chv_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "display",
/*
* Pipe A power well is the new disp2d well. Pipe B and C
* power wells don't actually exist. Pipe A power well is
* required for any pipe to work.
*/
.domains = CHV_DISPLAY_POWER_DOMAINS,
.id = CHV_DISP_PW_PIPE_A,
.ops = &chv_pipe_power_well_ops,
},
{
.name = "dpio-common-bc",
.domains = CHV_DPIO_CMN_BC_POWER_DOMAINS,
.id = PUNIT_POWER_WELL_DPIO_CMN_BC,
.ops = &chv_dpio_cmn_power_well_ops,
},
{
.name = "dpio-common-d",
.domains = CHV_DPIO_CMN_D_POWER_DOMAINS,
.id = PUNIT_POWER_WELL_DPIO_CMN_D,
.ops = &chv_dpio_cmn_power_well_ops,
},
};
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv,
enum i915_power_well_id power_well_id)
drm/i915/skl: Assert the requirements to enter or exit DC5. Warn if the conditions to enter or exit DC5 are not satisfied such as support for runtime PM, state of power well, CSR loading etc. v2: Removed camelcase in functions and variables. v3: Do some minimal check to assert if CSR program is not loaded. v4: 1] Used an appropriate function lookup_power_well() to identify power well, instead of using a magic number which can change in future. 2] Split the conditions further in assert_can_enable_DC5() and added more checks. 3] Removed all WARNs from assert_can_disable_DC5 as they were unnecessary and added two new ones. 4] Changed variable names as updated in earlier patches. v5: 1] Change lookup_power_well function to take an int power well id. 2] Define a new intel_display_power_well_is_enabled helper function to check whether a particular power well is enabled. 3] Use CSR-related mutex in assert_csr_loaded function. v6: Remove use of dc5_enabled variable as it's no longer needed. v7: 1] Rebase to latest. 2] Move all DC5-related functions from intel_display.c to intel_runtime_pm.c. v8: After adding dmc ver 1.0 support rebased on top of nightly. (Animesh) v9: Modified below changes based on review comments from Imre. - Moved intel_display_power_well_is_enabled() to intel_runtime_pm.c. - Removed mutex lock from assert_csr_loaded(). (Animesh) Issue: VIZ-2819 Signed-off-by: A.Sunil Kamath <sunil.kamath@intel.com> Signed-off-by: Suketu Shah <suketu.j.shah@intel.com> Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Animesh Manna <animesh.manna@intel.com> Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-16 16:52:10 +08:00
{
struct i915_power_well *power_well;
bool ret;
power_well = lookup_power_well(dev_priv, power_well_id);
ret = power_well->ops->is_enabled(dev_priv, power_well);
return ret;
}
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
static struct i915_power_well skl_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "power well 1",
/* Handled by the DMC firmware */
.domains = 0,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_1,
{
.hsw.has_fuses = true,
},
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "MISC IO power well",
/* Handled by the DMC firmware */
.domains = 0,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_MISC_IO,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "DC off",
.domains = SKL_DISPLAY_DC_OFF_POWER_DOMAINS,
.ops = &gen9_dc_off_power_well_ops,
.id = SKL_DISP_PW_DC_OFF,
},
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
{
.name = "power well 2",
.domains = SKL_DISPLAY_POWERWELL_2_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_2,
{
.hsw.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
.hsw.has_vga = true,
.hsw.has_fuses = true,
},
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "DDI A/E IO power well",
.domains = SKL_DISPLAY_DDI_IO_A_E_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_A_E,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "DDI B IO power well",
.domains = SKL_DISPLAY_DDI_IO_B_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_B,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "DDI C IO power well",
.domains = SKL_DISPLAY_DDI_IO_C_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_C,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
{
.name = "DDI D IO power well",
.domains = SKL_DISPLAY_DDI_IO_D_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_D,
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
},
};
static struct i915_power_well bxt_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "power well 1",
.domains = 0,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_1,
{
.hsw.has_fuses = true,
},
},
{
.name = "DC off",
.domains = BXT_DISPLAY_DC_OFF_POWER_DOMAINS,
.ops = &gen9_dc_off_power_well_ops,
.id = SKL_DISP_PW_DC_OFF,
},
{
.name = "power well 2",
.domains = BXT_DISPLAY_POWERWELL_2_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_2,
{
.hsw.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
.hsw.has_vga = true,
.hsw.has_fuses = true,
},
},
{
.name = "dpio-common-a",
.domains = BXT_DPIO_CMN_A_POWER_DOMAINS,
.ops = &bxt_dpio_cmn_power_well_ops,
.id = BXT_DPIO_CMN_A,
{
.bxt.phy = DPIO_PHY1,
},
},
{
.name = "dpio-common-bc",
.domains = BXT_DPIO_CMN_BC_POWER_DOMAINS,
.ops = &bxt_dpio_cmn_power_well_ops,
.id = BXT_DPIO_CMN_BC,
{
.bxt.phy = DPIO_PHY0,
},
},
};
static struct i915_power_well glk_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "power well 1",
/* Handled by the DMC firmware */
.domains = 0,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_1,
{
.hsw.has_fuses = true,
},
},
{
.name = "DC off",
.domains = GLK_DISPLAY_DC_OFF_POWER_DOMAINS,
.ops = &gen9_dc_off_power_well_ops,
.id = SKL_DISP_PW_DC_OFF,
},
{
.name = "power well 2",
.domains = GLK_DISPLAY_POWERWELL_2_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_2,
{
.hsw.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
.hsw.has_vga = true,
.hsw.has_fuses = true,
},
},
{
.name = "dpio-common-a",
.domains = GLK_DPIO_CMN_A_POWER_DOMAINS,
.ops = &bxt_dpio_cmn_power_well_ops,
.id = BXT_DPIO_CMN_A,
{
.bxt.phy = DPIO_PHY1,
},
},
{
.name = "dpio-common-b",
.domains = GLK_DPIO_CMN_B_POWER_DOMAINS,
.ops = &bxt_dpio_cmn_power_well_ops,
.id = BXT_DPIO_CMN_BC,
{
.bxt.phy = DPIO_PHY0,
},
},
{
.name = "dpio-common-c",
.domains = GLK_DPIO_CMN_C_POWER_DOMAINS,
.ops = &bxt_dpio_cmn_power_well_ops,
.id = GLK_DPIO_CMN_C,
{
.bxt.phy = DPIO_PHY2,
},
},
{
.name = "AUX A",
.domains = GLK_DISPLAY_AUX_A_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = GLK_DISP_PW_AUX_A,
},
{
.name = "AUX B",
.domains = GLK_DISPLAY_AUX_B_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = GLK_DISP_PW_AUX_B,
},
{
.name = "AUX C",
.domains = GLK_DISPLAY_AUX_C_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = GLK_DISP_PW_AUX_C,
},
{
.name = "DDI A IO power well",
.domains = GLK_DISPLAY_DDI_IO_A_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = GLK_DISP_PW_DDI_A,
},
{
.name = "DDI B IO power well",
.domains = GLK_DISPLAY_DDI_IO_B_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_B,
},
{
.name = "DDI C IO power well",
.domains = GLK_DISPLAY_DDI_IO_C_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_C,
},
};
static struct i915_power_well cnl_power_wells[] = {
{
.name = "always-on",
.always_on = 1,
.domains = POWER_DOMAIN_MASK,
.ops = &i9xx_always_on_power_well_ops,
.id = I915_DISP_PW_ALWAYS_ON,
},
{
.name = "power well 1",
/* Handled by the DMC firmware */
.domains = 0,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_1,
{
.hsw.has_fuses = true,
},
},
{
.name = "AUX A",
.domains = CNL_DISPLAY_AUX_A_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_AUX_A,
},
{
.name = "AUX B",
.domains = CNL_DISPLAY_AUX_B_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_AUX_B,
},
{
.name = "AUX C",
.domains = CNL_DISPLAY_AUX_C_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_AUX_C,
},
{
.name = "AUX D",
.domains = CNL_DISPLAY_AUX_D_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_AUX_D,
},
{
.name = "DC off",
.domains = CNL_DISPLAY_DC_OFF_POWER_DOMAINS,
.ops = &gen9_dc_off_power_well_ops,
.id = SKL_DISP_PW_DC_OFF,
},
{
.name = "power well 2",
.domains = CNL_DISPLAY_POWERWELL_2_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_2,
{
.hsw.irq_pipe_mask = BIT(PIPE_B) | BIT(PIPE_C),
.hsw.has_vga = true,
.hsw.has_fuses = true,
},
},
{
.name = "DDI A IO power well",
.domains = CNL_DISPLAY_DDI_A_IO_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_DDI_A,
},
{
.name = "DDI B IO power well",
.domains = CNL_DISPLAY_DDI_B_IO_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_B,
},
{
.name = "DDI C IO power well",
.domains = CNL_DISPLAY_DDI_C_IO_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_C,
},
{
.name = "DDI D IO power well",
.domains = CNL_DISPLAY_DDI_D_IO_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = SKL_DISP_PW_DDI_D,
},
{
.name = "DDI F IO power well",
.domains = CNL_DISPLAY_DDI_F_IO_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_DDI_F,
},
{
.name = "AUX F",
.domains = CNL_DISPLAY_AUX_F_POWER_DOMAINS,
.ops = &hsw_power_well_ops,
.id = CNL_DISP_PW_AUX_F,
},
};
static int
sanitize_disable_power_well_option(const struct drm_i915_private *dev_priv,
int disable_power_well)
{
if (disable_power_well >= 0)
return !!disable_power_well;
return 1;
}
static uint32_t get_allowed_dc_mask(const struct drm_i915_private *dev_priv,
int enable_dc)
{
uint32_t mask;
int requested_dc;
int max_dc;
if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) {
max_dc = 2;
mask = 0;
} else if (IS_GEN9_LP(dev_priv)) {
max_dc = 1;
/*
* DC9 has a separate HW flow from the rest of the DC states,
* not depending on the DMC firmware. It's needed by system
* suspend/resume, so allow it unconditionally.
*/
mask = DC_STATE_EN_DC9;
} else {
max_dc = 0;
mask = 0;
}
if (!i915_modparams.disable_power_well)
max_dc = 0;
if (enable_dc >= 0 && enable_dc <= max_dc) {
requested_dc = enable_dc;
} else if (enable_dc == -1) {
requested_dc = max_dc;
} else if (enable_dc > max_dc && enable_dc <= 2) {
DRM_DEBUG_KMS("Adjusting requested max DC state (%d->%d)\n",
enable_dc, max_dc);
requested_dc = max_dc;
} else {
DRM_ERROR("Unexpected value for enable_dc (%d)\n", enable_dc);
requested_dc = max_dc;
}
if (requested_dc > 1)
mask |= DC_STATE_EN_UPTO_DC6;
if (requested_dc > 0)
mask |= DC_STATE_EN_UPTO_DC5;
DRM_DEBUG_KMS("Allowed DC state mask %02x\n", mask);
return mask;
}
static void assert_power_well_ids_unique(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
u64 power_well_ids;
int i;
power_well_ids = 0;
for (i = 0; i < power_domains->power_well_count; i++) {
enum i915_power_well_id id = power_domains->power_wells[i].id;
WARN_ON(id >= sizeof(power_well_ids) * 8);
WARN_ON(power_well_ids & BIT_ULL(id));
power_well_ids |= BIT_ULL(id);
}
}
#define set_power_wells(power_domains, __power_wells) ({ \
(power_domains)->power_wells = (__power_wells); \
(power_domains)->power_well_count = ARRAY_SIZE(__power_wells); \
})
/**
* intel_power_domains_init - initializes the power domain structures
* @dev_priv: i915 device instance
*
* Initializes the power domain structures for @dev_priv depending upon the
* supported platform.
*/
int intel_power_domains_init(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
i915_modparams.disable_power_well =
sanitize_disable_power_well_option(dev_priv,
i915_modparams.disable_power_well);
dev_priv->csr.allowed_dc_mask =
get_allowed_dc_mask(dev_priv, i915_modparams.enable_dc);
BUILD_BUG_ON(POWER_DOMAIN_NUM > 64);
mutex_init(&power_domains->lock);
/*
* The enabling order will be from lower to higher indexed wells,
* the disabling order is reversed.
*/
if (IS_HASWELL(dev_priv)) {
set_power_wells(power_domains, hsw_power_wells);
} else if (IS_BROADWELL(dev_priv)) {
set_power_wells(power_domains, bdw_power_wells);
} else if (IS_GEN9_BC(dev_priv)) {
drm/i915/skl: Implementation of SKL display power well support This patch implements core logic of SKL display power well. v2: Addressed Imre's comments - Added respective DDIs under power well #1 and #2 - Simplified repetitive code in power well programming v3: Implemented Imre's comments - Further simplified power well programming - Made sure that PW 1 is enabled prior to PW 2 v4: Fix minor conflict with the the cherryview support (Damien) v5: Add the PLL power domain to the always on power well (Damien) v6: Disable BIOS power well (Imre) Use power well data for comparison (Imre) Put the PLL power domain into PW1 as its needed for CDCLK (Satheesh, Damien) v7: Addressed Imre's comments - Lowered the time out to 1ms - Added parantheses in macro - Moved debug message and fixed wait_for interval v8: - Add a WARN() when swiching on an unknown power well (Imre, done by Damien) - Whitespace fixes (spaces instead of tabs) (Damien) v9: (Imre, done by Damien) - Merge the register definitions with this patch - Merge the MISC IO power well in this patch v10: (Imre, done by Damien) - Define the Misc I/O power domains to be the power well 1 ones as Misc I/O needs to be enabled with PW1 - Added Transcoder A and VGA domains to PW 2 - Remove the MISC_IO power domains as well in the the always on domains definition - Move Misc I/O power well at the top of the power well list so it's turned on right after PW1. Reviewed-by: Imre Deak <imre.deak@intel.com> Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v3,v6,v7) Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-02-04 21:57:44 +08:00
set_power_wells(power_domains, skl_power_wells);
} else if (IS_CANNONLAKE(dev_priv)) {
set_power_wells(power_domains, cnl_power_wells);
/*
* DDI and Aux IO are getting enabled for all ports
* regardless the presence or use. So, in order to avoid
* timeouts, lets remove them from the list
* for the SKUs without port F.
*/
if (!IS_CNL_WITH_PORT_F(dev_priv))
power_domains->power_well_count -= 2;
} else if (IS_BROXTON(dev_priv)) {
set_power_wells(power_domains, bxt_power_wells);
} else if (IS_GEMINILAKE(dev_priv)) {
set_power_wells(power_domains, glk_power_wells);
} else if (IS_CHERRYVIEW(dev_priv)) {
set_power_wells(power_domains, chv_power_wells);
} else if (IS_VALLEYVIEW(dev_priv)) {
set_power_wells(power_domains, vlv_power_wells);
} else if (IS_I830(dev_priv)) {
set_power_wells(power_domains, i830_power_wells);
} else {
set_power_wells(power_domains, i9xx_always_on_power_well);
}
assert_power_well_ids_unique(dev_priv);
return 0;
}
/**
* intel_power_domains_fini - finalizes the power domain structures
* @dev_priv: i915 device instance
*
* Finalizes the power domain structures for @dev_priv depending upon the
* supported platform. This function also disables runtime pm and ensures that
* the device stays powered up so that the driver can be reloaded.
*/
void intel_power_domains_fini(struct drm_i915_private *dev_priv)
{
struct device *kdev = &dev_priv->drm.pdev->dev;
/*
* The i915.ko module is still not prepared to be loaded when
* the power well is not enabled, so just enable it in case
* we're going to unload/reload.
* The following also reacquires the RPM reference the core passed
* to the driver during loading, which is dropped in
* intel_runtime_pm_enable(). We have to hand back the control of the
* device to the core with this reference held.
*/
intel_display_set_init_power(dev_priv, true);
/* Remove the refcount we took to keep power well support disabled. */
if (!i915_modparams.disable_power_well)
intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
/*
* Remove the refcount we took in intel_runtime_pm_enable() in case
* the platform doesn't support runtime PM.
*/
if (!HAS_RUNTIME_PM(dev_priv))
pm_runtime_put(kdev);
}
static void intel_power_domains_sync_hw(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *power_well;
mutex_lock(&power_domains->lock);
for_each_power_well(dev_priv, power_well) {
power_well->ops->sync_hw(dev_priv, power_well);
power_well->hw_enabled = power_well->ops->is_enabled(dev_priv,
power_well);
}
mutex_unlock(&power_domains->lock);
}
static void gen9_dbuf_enable(struct drm_i915_private *dev_priv)
{
I915_WRITE(DBUF_CTL, I915_READ(DBUF_CTL) | DBUF_POWER_REQUEST);
POSTING_READ(DBUF_CTL);
udelay(10);
if (!(I915_READ(DBUF_CTL) & DBUF_POWER_STATE))
DRM_ERROR("DBuf power enable timeout\n");
}
static void gen9_dbuf_disable(struct drm_i915_private *dev_priv)
{
I915_WRITE(DBUF_CTL, I915_READ(DBUF_CTL) & ~DBUF_POWER_REQUEST);
POSTING_READ(DBUF_CTL);
udelay(10);
if (I915_READ(DBUF_CTL) & DBUF_POWER_STATE)
DRM_ERROR("DBuf power disable timeout!\n");
}
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
static void skl_display_core_init(struct drm_i915_private *dev_priv,
bool resume)
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
uint32_t val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
/* enable PCH reset handshake */
val = I915_READ(HSW_NDE_RSTWRN_OPT);
I915_WRITE(HSW_NDE_RSTWRN_OPT, val | RESET_PCH_HANDSHAKE_ENABLE);
/* enable PG1 and Misc I/O */
mutex_lock(&power_domains->lock);
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_enable(dev_priv, well);
well = lookup_power_well(dev_priv, SKL_DISP_PW_MISC_IO);
intel_power_well_enable(dev_priv, well);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
mutex_unlock(&power_domains->lock);
skl_init_cdclk(dev_priv);
gen9_dbuf_enable(dev_priv);
if (resume && dev_priv->csr.dmc_payload)
intel_csr_load_program(dev_priv);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
}
static void skl_display_core_uninit(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
gen9_dbuf_disable(dev_priv);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
skl_uninit_cdclk(dev_priv);
/* The spec doesn't call for removing the reset handshake flag */
/* disable PG1 and Misc I/O */
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
mutex_lock(&power_domains->lock);
/*
* BSpec says to keep the MISC IO power well enabled here, only
* remove our request for power well 1.
* Note that even though the driver's request is removed power well 1
* may stay enabled after this due to DMC's own request on it.
*/
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_disable(dev_priv, well);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
mutex_unlock(&power_domains->lock);
usleep_range(10, 30); /* 10 us delay per Bspec */
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
}
void bxt_display_core_init(struct drm_i915_private *dev_priv,
bool resume)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
uint32_t val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
/*
* NDE_RSTWRN_OPT RST PCH Handshake En must always be 0b on BXT
* or else the reset will hang because there is no PCH to respond.
* Move the handshake programming to initialization sequence.
* Previously was left up to BIOS.
*/
val = I915_READ(HSW_NDE_RSTWRN_OPT);
val &= ~RESET_PCH_HANDSHAKE_ENABLE;
I915_WRITE(HSW_NDE_RSTWRN_OPT, val);
/* Enable PG1 */
mutex_lock(&power_domains->lock);
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_enable(dev_priv, well);
mutex_unlock(&power_domains->lock);
bxt_init_cdclk(dev_priv);
gen9_dbuf_enable(dev_priv);
if (resume && dev_priv->csr.dmc_payload)
intel_csr_load_program(dev_priv);
}
void bxt_display_core_uninit(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
gen9_dbuf_disable(dev_priv);
bxt_uninit_cdclk(dev_priv);
/* The spec doesn't call for removing the reset handshake flag */
/*
* Disable PW1 (PG1).
* Note that even though the driver's request is removed power well 1
* may stay enabled after this due to DMC's own request on it.
*/
mutex_lock(&power_domains->lock);
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_disable(dev_priv, well);
mutex_unlock(&power_domains->lock);
usleep_range(10, 30); /* 10 us delay per Bspec */
}
enum {
PROCMON_0_85V_DOT_0,
PROCMON_0_95V_DOT_0,
PROCMON_0_95V_DOT_1,
PROCMON_1_05V_DOT_0,
PROCMON_1_05V_DOT_1,
};
static const struct cnl_procmon {
u32 dw1, dw9, dw10;
} cnl_procmon_values[] = {
[PROCMON_0_85V_DOT_0] =
{ .dw1 = 0x00000000, .dw9 = 0x62AB67BB, .dw10 = 0x51914F96, },
[PROCMON_0_95V_DOT_0] =
{ .dw1 = 0x00000000, .dw9 = 0x86E172C7, .dw10 = 0x77CA5EAB, },
[PROCMON_0_95V_DOT_1] =
{ .dw1 = 0x00000000, .dw9 = 0x93F87FE1, .dw10 = 0x8AE871C5, },
[PROCMON_1_05V_DOT_0] =
{ .dw1 = 0x00000000, .dw9 = 0x98FA82DD, .dw10 = 0x89E46DC1, },
[PROCMON_1_05V_DOT_1] =
{ .dw1 = 0x00440000, .dw9 = 0x9A00AB25, .dw10 = 0x8AE38FF1, },
};
/*
* CNL has just one set of registers, while ICL has two sets: one for port A and
* the other for port B. The CNL registers are equivalent to the ICL port A
* registers, that's why we call the ICL macros even though the function has CNL
* on its name.
*/
static void cnl_set_procmon_ref_values(struct drm_i915_private *dev_priv,
enum port port)
{
const struct cnl_procmon *procmon;
u32 val;
val = I915_READ(ICL_PORT_COMP_DW3(port));
switch (val & (PROCESS_INFO_MASK | VOLTAGE_INFO_MASK)) {
default:
MISSING_CASE(val);
case VOLTAGE_INFO_0_85V | PROCESS_INFO_DOT_0:
procmon = &cnl_procmon_values[PROCMON_0_85V_DOT_0];
break;
case VOLTAGE_INFO_0_95V | PROCESS_INFO_DOT_0:
procmon = &cnl_procmon_values[PROCMON_0_95V_DOT_0];
break;
case VOLTAGE_INFO_0_95V | PROCESS_INFO_DOT_1:
procmon = &cnl_procmon_values[PROCMON_0_95V_DOT_1];
break;
case VOLTAGE_INFO_1_05V | PROCESS_INFO_DOT_0:
procmon = &cnl_procmon_values[PROCMON_1_05V_DOT_0];
break;
case VOLTAGE_INFO_1_05V | PROCESS_INFO_DOT_1:
procmon = &cnl_procmon_values[PROCMON_1_05V_DOT_1];
break;
}
val = I915_READ(ICL_PORT_COMP_DW1(port));
val &= ~((0xff << 16) | 0xff);
val |= procmon->dw1;
I915_WRITE(ICL_PORT_COMP_DW1(port), val);
I915_WRITE(ICL_PORT_COMP_DW9(port), procmon->dw9);
I915_WRITE(ICL_PORT_COMP_DW10(port), procmon->dw10);
}
static void cnl_display_core_init(struct drm_i915_private *dev_priv, bool resume)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
u32 val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
/* 1. Enable PCH Reset Handshake */
val = I915_READ(HSW_NDE_RSTWRN_OPT);
val |= RESET_PCH_HANDSHAKE_ENABLE;
I915_WRITE(HSW_NDE_RSTWRN_OPT, val);
/* 2. Enable Comp */
val = I915_READ(CHICKEN_MISC_2);
val &= ~CNL_COMP_PWR_DOWN;
I915_WRITE(CHICKEN_MISC_2, val);
/* Dummy PORT_A to get the correct CNL register from the ICL macro */
cnl_set_procmon_ref_values(dev_priv, PORT_A);
val = I915_READ(CNL_PORT_COMP_DW0);
val |= COMP_INIT;
I915_WRITE(CNL_PORT_COMP_DW0, val);
/* 3. */
val = I915_READ(CNL_PORT_CL1CM_DW5);
val |= CL_POWER_DOWN_ENABLE;
I915_WRITE(CNL_PORT_CL1CM_DW5, val);
/*
* 4. Enable Power Well 1 (PG1).
* The AUX IO power wells will be enabled on demand.
*/
mutex_lock(&power_domains->lock);
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_enable(dev_priv, well);
mutex_unlock(&power_domains->lock);
/* 5. Enable CD clock */
cnl_init_cdclk(dev_priv);
/* 6. Enable DBUF */
gen9_dbuf_enable(dev_priv);
if (resume && dev_priv->csr.dmc_payload)
intel_csr_load_program(dev_priv);
}
static void cnl_display_core_uninit(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *well;
u32 val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
/* 1. Disable all display engine functions -> aready done */
/* 2. Disable DBUF */
gen9_dbuf_disable(dev_priv);
/* 3. Disable CD clock */
cnl_uninit_cdclk(dev_priv);
/*
* 4. Disable Power Well 1 (PG1).
* The AUX IO power wells are toggled on demand, so they are already
* disabled at this point.
*/
mutex_lock(&power_domains->lock);
well = lookup_power_well(dev_priv, SKL_DISP_PW_1);
intel_power_well_disable(dev_priv, well);
mutex_unlock(&power_domains->lock);
usleep_range(10, 30); /* 10 us delay per Bspec */
/* 5. Disable Comp */
val = I915_READ(CHICKEN_MISC_2);
val |= CNL_COMP_PWR_DOWN;
I915_WRITE(CHICKEN_MISC_2, val);
}
static void icl_display_core_init(struct drm_i915_private *dev_priv,
bool resume)
{
enum port port;
u32 val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
/* 1. Enable PCH reset handshake. */
val = I915_READ(HSW_NDE_RSTWRN_OPT);
val |= RESET_PCH_HANDSHAKE_ENABLE;
I915_WRITE(HSW_NDE_RSTWRN_OPT, val);
for (port = PORT_A; port <= PORT_B; port++) {
/* 2. Enable DDI combo PHY comp. */
val = I915_READ(ICL_PHY_MISC(port));
val &= ~ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN;
I915_WRITE(ICL_PHY_MISC(port), val);
cnl_set_procmon_ref_values(dev_priv, port);
val = I915_READ(ICL_PORT_COMP_DW0(port));
val |= COMP_INIT;
I915_WRITE(ICL_PORT_COMP_DW0(port), val);
/* 3. Set power down enable. */
val = I915_READ(ICL_PORT_CL_DW5(port));
val |= CL_POWER_DOWN_ENABLE;
I915_WRITE(ICL_PORT_CL_DW5(port), val);
}
/* 4. Enable power well 1 (PG1) and aux IO power. */
/* FIXME: ICL power wells code not here yet. */
/* 5. Enable CDCLK. */
icl_init_cdclk(dev_priv);
/* 6. Enable DBUF. */
gen9_dbuf_enable(dev_priv);
/* 7. Setup MBUS. */
/* FIXME: MBUS code not here yet. */
/* 8. CHICKEN_DCPR_1 */
I915_WRITE(GEN8_CHICKEN_DCPR_1, I915_READ(GEN8_CHICKEN_DCPR_1) |
CNL_DDI_CLOCK_REG_ACCESS_ON);
}
static void icl_display_core_uninit(struct drm_i915_private *dev_priv)
{
enum port port;
u32 val;
gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
/* 1. Disable all display engine functions -> aready done */
/* 2. Disable DBUF */
gen9_dbuf_disable(dev_priv);
/* 3. Disable CD clock */
icl_uninit_cdclk(dev_priv);
/* 4. Disable Power Well 1 (PG1) and Aux IO Power */
/* FIXME: ICL power wells code not here yet. */
/* 5. Disable Comp */
for (port = PORT_A; port <= PORT_B; port++) {
val = I915_READ(ICL_PHY_MISC(port));
val |= ICL_PHY_MISC_DE_IO_COMP_PWR_DOWN;
I915_WRITE(ICL_PHY_MISC(port), val);
}
}
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
static void chv_phy_control_init(struct drm_i915_private *dev_priv)
{
struct i915_power_well *cmn_bc =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
struct i915_power_well *cmn_d =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_D);
/*
* DISPLAY_PHY_CONTROL can get corrupted if read. As a
* workaround never ever read DISPLAY_PHY_CONTROL, and
* instead maintain a shadow copy ourselves. Use the actual
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
* power well state and lane status to reconstruct the
* expected initial value.
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
*/
dev_priv->chv_phy_control =
PHY_LDO_SEQ_DELAY(PHY_LDO_DELAY_600NS, DPIO_PHY0) |
PHY_LDO_SEQ_DELAY(PHY_LDO_DELAY_600NS, DPIO_PHY1) |
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY0, DPIO_CH0) |
PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY0, DPIO_CH1) |
PHY_CH_POWER_MODE(PHY_CH_DEEP_PSR, DPIO_PHY1, DPIO_CH0);
/*
* If all lanes are disabled we leave the override disabled
* with all power down bits cleared to match the state we
* would use after disabling the port. Otherwise enable the
* override and set the lane powerdown bits accding to the
* current lane status.
*/
if (cmn_bc->ops->is_enabled(dev_priv, cmn_bc)) {
uint32_t status = I915_READ(DPLL(PIPE_A));
unsigned int mask;
mask = status & DPLL_PORTB_READY_MASK;
if (mask == 0xf)
mask = 0x0;
else
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH0);
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY0, DPIO_CH0);
mask = (status & DPLL_PORTC_READY_MASK) >> 4;
if (mask == 0xf)
mask = 0x0;
else
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY0, DPIO_CH1);
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY0, DPIO_CH1);
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
dev_priv->chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(DPIO_PHY0);
dev_priv->chv_phy_assert[DPIO_PHY0] = false;
} else {
dev_priv->chv_phy_assert[DPIO_PHY0] = true;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
}
if (cmn_d->ops->is_enabled(dev_priv, cmn_d)) {
uint32_t status = I915_READ(DPIO_PHY_STATUS);
unsigned int mask;
mask = status & DPLL_PORTD_READY_MASK;
if (mask == 0xf)
mask = 0x0;
else
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD_EN(DPIO_PHY1, DPIO_CH0);
dev_priv->chv_phy_control |=
PHY_CH_POWER_DOWN_OVRD(mask, DPIO_PHY1, DPIO_CH0);
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
dev_priv->chv_phy_control |= PHY_COM_LANE_RESET_DEASSERT(DPIO_PHY1);
dev_priv->chv_phy_assert[DPIO_PHY1] = false;
} else {
dev_priv->chv_phy_assert[DPIO_PHY1] = true;
drm/i915: Implement PHY lane power gating for CHV Powergate the PHY lanes when they're not needed. For HDMI all four lanes are needed always, but for DP we can enable only the needed lanes. To power down the unused lanes we use some power down override bits in the DISPLAY_PHY_CONTROL register. Without the overrides it appears that the hardware always powers on all the lanes. When the port is disabled the power down override is not needed and the lanes will shut off on their own. That also means the override is critical to actually be able to access the DPIO registers before the port is actually enabled. Additionally the common lanes will power down when not needed. CL1 remains on as long as anything else is on, CL2 will shut down when all the lanes in the same channel will shut down. There is one exception for CL2 that will be dealt in a separate patch for clarity. With potentially some lanes powered down, the DP code now has to check the number of active lanes before accessing PCS/TX registers. All registers in powered down blocks will reads as 0xffffffff, and soe we would drown in warnings from vlv_dpio_read() if we allowed the code to access all those registers. Another important detail in the DP code is the "TX latency optimal" setting. Normally the second TX lane acts as some kind of reset master, with the other lanes as slaves. But when only a single lane is enabled, that single lane obviously has to be the master. A bit of extra care is needed to reconstruct the initial state of the DISPLAY_PHY_CONTROL register since it can't be read safely. So instead read the actual lane status from the DPLL/PHY_STATUS registers and use that to determine which lanes ought to be powergated initially. We also need to switch the PHY power modes to "deep PSR" to avoid a hard system hang when powering down the single channel PHY. Also sprinkle a few debug prints around so that we can monitor the DISPLAY_PHY_STATUS changes without having to read it and risk corrupting it. v2: Add locking to chv_powergate_phy_lanes() v3: Actually enable dynamic powerdown in the PHY and deal with the fallout Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-07-09 04:45:54 +08:00
}
I915_WRITE(DISPLAY_PHY_CONTROL, dev_priv->chv_phy_control);
DRM_DEBUG_KMS("Initial PHY_CONTROL=0x%08x\n",
dev_priv->chv_phy_control);
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
}
static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
{
struct i915_power_well *cmn =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DPIO_CMN_BC);
struct i915_power_well *disp2d =
lookup_power_well(dev_priv, PUNIT_POWER_WELL_DISP2D);
/* If the display might be already active skip this */
if (cmn->ops->is_enabled(dev_priv, cmn) &&
disp2d->ops->is_enabled(dev_priv, disp2d) &&
I915_READ(DPIO_CTL) & DPIO_CMNRST)
return;
DRM_DEBUG_KMS("toggling display PHY side reset\n");
/* cmnlane needs DPLL registers */
disp2d->ops->enable(dev_priv, disp2d);
/*
* From VLV2A0_DP_eDP_HDMI_DPIO_driver_vbios_notes_11.docx:
* Need to assert and de-assert PHY SB reset by gating the
* common lane power, then un-gating it.
* Simply ungating isn't enough to reset the PHY enough to get
* ports and lanes running.
*/
cmn->ops->disable(dev_priv, cmn);
}
/**
* intel_power_domains_init_hw - initialize hardware power domain state
* @dev_priv: i915 device instance
* @resume: Called from resume code paths or not
*
* This function initializes the hardware power domain state and enables all
* power wells belonging to the INIT power domain. Power wells in other
* domains (and not in the INIT domain) are referenced or disabled during the
* modeset state HW readout. After that the reference count of each power well
* must match its HW enabled state, see intel_power_domains_verify_state().
*/
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
power_domains->initializing = true;
if (IS_ICELAKE(dev_priv)) {
icl_display_core_init(dev_priv, resume);
} else if (IS_CANNONLAKE(dev_priv)) {
cnl_display_core_init(dev_priv, resume);
} else if (IS_GEN9_BC(dev_priv)) {
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
skl_display_core_init(dev_priv, resume);
} else if (IS_GEN9_LP(dev_priv)) {
bxt_display_core_init(dev_priv, resume);
} else if (IS_CHERRYVIEW(dev_priv)) {
mutex_lock(&power_domains->lock);
drm/i915: Work around DISPLAY_PHY_CONTROL register corruption on CHV Sometimes (exactly when is a bit unclear) DISPLAY_PHY_CONTROL appears to get corrupted. The values I've managed to read from it seem to have some pattern but vary quite a lot. The corruption doesn't seem to just happen when the register is accessed, but can also happen spontaneosly during modeset. When this happens during a modeset things go south and the display doesn't light up. I've managed to hit the problemn when toggling HDMI on port D on and off. When things get corrupted the display doesn't light up, but as soon as I manually write the correct value to the register the display comes up. First I was suspicious that we ourselves accidentally overwrite it with garbage, but didn't catch anything with the reg_rw tracepoint. Also I sprinkled check all over the modeset path to see exactly when the corruption happens, and eg. the read back value was fine just before intel_dp_set_m(), and corrupted immediately after it. I also made my check function repair the register value whenever it was wrong, and with this approach the corruption repeated several times during the modeset operation, always seeming to trigger in the same exact calls to the check function, while other calls to the function never caught anything. So far I've not seen this problem occurring when carefully avoiding all read accesses to DISPLAY_PHY_CONTROL. Not sure if that's just pure luck or an actual workaround, but we can hope it works. So let's avoid reading the register and instead track the desired value of the register in dev_priv. v2: Read out the power well state to determine initial register value v3: Use DPIO_CHx names instead of raw numbers Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Reviewed-by: Deepak S <deepak.s@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-04-10 23:21:28 +08:00
chv_phy_control_init(dev_priv);
mutex_unlock(&power_domains->lock);
} else if (IS_VALLEYVIEW(dev_priv)) {
mutex_lock(&power_domains->lock);
vlv_cmnlane_wa(dev_priv);
mutex_unlock(&power_domains->lock);
}
/* For now, we need the power well to be always enabled. */
intel_display_set_init_power(dev_priv, true);
/* Disable power support if the user asked so. */
if (!i915_modparams.disable_power_well)
intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
intel_power_domains_sync_hw(dev_priv);
power_domains->initializing = false;
}
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
/**
* intel_power_domains_suspend - suspend power domain state
* @dev_priv: i915 device instance
*
* This function prepares the hardware power domain state before entering
* system suspend. It must be paired with intel_power_domains_init_hw().
*/
void intel_power_domains_suspend(struct drm_i915_private *dev_priv)
{
/*
* Even if power well support was disabled we still want to disable
* power wells while we are system suspended.
*/
if (!i915_modparams.disable_power_well)
intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
if (IS_ICELAKE(dev_priv))
icl_display_core_uninit(dev_priv);
else if (IS_CANNONLAKE(dev_priv))
cnl_display_core_uninit(dev_priv);
else if (IS_GEN9_BC(dev_priv))
skl_display_core_uninit(dev_priv);
else if (IS_GEN9_LP(dev_priv))
bxt_display_core_uninit(dev_priv);
drm/i915/skl: init/uninit display core as part of the HW power domain state We need to initialize the display core part early, before initializing the rest of the display power state. This is also described in the bspec termed "Display initialization sequence". Atm we run this sequence during driver loading after power domain HW state initialization which is too late and during runtime suspend/resume which is unneeded and can interere with DMC functionality which handles HW resources toggled by this init/uninit sequence automatically. The init sequence must be run as the first step of HW power state initialization and during system resume. The uninit sequence must be run during system suspend. To address the above move the init sequence to the initial HW power state setup and the uninit sequence to a new power domains suspend function called during system suspend. As part of the init sequence we also have to reprogram the DMC firmware as it's lost across a system suspend/resume cycle. After this change CD clock initialization during driver loading will happen only later after other dependent HW/SW parts are initialized, while during system resume it will get initialized as the last step of the init sequence. This distinction can be removed by some refactoring of platform independent parts. I left this refactoring out from this series since I didn't want to change non-SKL parts. This is a TODO for later. v2: - fix error path in i915_drm_suspend_late() - don't try to re-program the DMC firmware if it failed to load Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1447774433-20834-1-git-send-email-imre.deak@intel.com
2015-11-17 23:33:53 +08:00
}
static void intel_power_domains_dump_info(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *power_well;
for_each_power_well(dev_priv, power_well) {
enum intel_display_power_domain domain;
DRM_DEBUG_DRIVER("%-25s %d\n",
power_well->name, power_well->count);
for_each_power_domain(domain, power_well->domains)
DRM_DEBUG_DRIVER(" %-23s %d\n",
intel_display_power_domain_str(domain),
power_domains->domain_use_count[domain]);
}
}
/**
* intel_power_domains_verify_state - verify the HW/SW state for all power wells
* @dev_priv: i915 device instance
*
* Verify if the reference count of each power well matches its HW enabled
* state and the total refcount of the domains it belongs to. This must be
* called after modeset HW state sanitization, which is responsible for
* acquiring reference counts for any power wells in use and disabling the
* ones left on by BIOS but not required by any active output.
*/
void intel_power_domains_verify_state(struct drm_i915_private *dev_priv)
{
struct i915_power_domains *power_domains = &dev_priv->power_domains;
struct i915_power_well *power_well;
bool dump_domain_info;
mutex_lock(&power_domains->lock);
dump_domain_info = false;
for_each_power_well(dev_priv, power_well) {
enum intel_display_power_domain domain;
int domains_count;
bool enabled;
/*
* Power wells not belonging to any domain (like the MISC_IO
* and PW1 power wells) are under FW control, so ignore them,
* since their state can change asynchronously.
*/
if (!power_well->domains)
continue;
enabled = power_well->ops->is_enabled(dev_priv, power_well);
if ((power_well->count || power_well->always_on) != enabled)
DRM_ERROR("power well %s state mismatch (refcount %d/enabled %d)",
power_well->name, power_well->count, enabled);
domains_count = 0;
for_each_power_domain(domain, power_well->domains)
domains_count += power_domains->domain_use_count[domain];
if (power_well->count != domains_count) {
DRM_ERROR("power well %s refcount/domain refcount mismatch "
"(refcount %d/domains refcount %d)\n",
power_well->name, power_well->count,
domains_count);
dump_domain_info = true;
}
}
if (dump_domain_info) {
static bool dumped;
if (!dumped) {
intel_power_domains_dump_info(dev_priv);
dumped = true;
}
}
mutex_unlock(&power_domains->lock);
}
/**
* intel_runtime_pm_get - grab a runtime pm reference
* @dev_priv: i915 device instance
*
* This function grabs a device-level runtime pm reference (mostly used for GEM
* code to ensure the GTT or GT is on) and ensures that it is powered up.
*
* Any runtime pm reference obtained by this function must have a symmetric
* call to intel_runtime_pm_put() to release the reference again.
*/
void intel_runtime_pm_get(struct drm_i915_private *dev_priv)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
struct device *kdev = &pdev->dev;
int ret;
ret = pm_runtime_get_sync(kdev);
WARN_ONCE(ret < 0, "pm_runtime_get_sync() failed: %d\n", ret);
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
atomic_inc(&dev_priv->runtime_pm.wakeref_count);
assert_rpm_wakelock_held(dev_priv);
}
/**
* intel_runtime_pm_get_if_in_use - grab a runtime pm reference if device in use
* @dev_priv: i915 device instance
*
* This function grabs a device-level runtime pm reference if the device is
* already in use and ensures that it is powered up.
*
* Any runtime pm reference obtained by this function must have a symmetric
* call to intel_runtime_pm_put() to release the reference again.
*/
bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
struct device *kdev = &pdev->dev;
if (IS_ENABLED(CONFIG_PM)) {
int ret = pm_runtime_get_if_in_use(kdev);
/*
* In cases runtime PM is disabled by the RPM core and we get
* an -EINVAL return value we are not supposed to call this
* function, since the power state is undefined. This applies
* atm to the late/early system suspend/resume handlers.
*/
WARN_ONCE(ret < 0,
"pm_runtime_get_if_in_use() failed: %d\n", ret);
if (ret <= 0)
return false;
}
atomic_inc(&dev_priv->runtime_pm.wakeref_count);
assert_rpm_wakelock_held(dev_priv);
return true;
}
/**
* intel_runtime_pm_get_noresume - grab a runtime pm reference
* @dev_priv: i915 device instance
*
* This function grabs a device-level runtime pm reference (mostly used for GEM
* code to ensure the GTT or GT is on).
*
* It will _not_ power up the device but instead only check that it's powered
* on. Therefore it is only valid to call this functions from contexts where
* the device is known to be powered up and where trying to power it up would
* result in hilarity and deadlocks. That pretty much means only the system
* suspend/resume code where this is used to grab runtime pm references for
* delayed setup down in work items.
*
* Any runtime pm reference obtained by this function must have a symmetric
* call to intel_runtime_pm_put() to release the reference again.
*/
void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
struct device *kdev = &pdev->dev;
assert_rpm_wakelock_held(dev_priv);
pm_runtime_get_noresume(kdev);
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
atomic_inc(&dev_priv->runtime_pm.wakeref_count);
}
/**
* intel_runtime_pm_put - release a runtime pm reference
* @dev_priv: i915 device instance
*
* This function drops the device-level runtime pm reference obtained by
* intel_runtime_pm_get() and might power down the corresponding
* hardware block right away if this is the last reference.
*/
void intel_runtime_pm_put(struct drm_i915_private *dev_priv)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
struct device *kdev = &pdev->dev;
assert_rpm_wakelock_held(dev_priv);
atomic_dec(&dev_priv->runtime_pm.wakeref_count);
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
pm_runtime_mark_last_busy(kdev);
pm_runtime_put_autosuspend(kdev);
}
/**
* intel_runtime_pm_enable - enable runtime pm
* @dev_priv: i915 device instance
*
* This function enables runtime pm at the end of the driver load sequence.
*
* Note that this function does currently not enable runtime pm for the
* subordinate display power domains. That is only done on the first modeset
* using intel_display_set_init_power().
*/
void intel_runtime_pm_enable(struct drm_i915_private *dev_priv)
{
struct pci_dev *pdev = dev_priv->drm.pdev;
struct device *kdev = &pdev->dev;
pm_runtime_set_autosuspend_delay(kdev, 10000); /* 10s */
pm_runtime_mark_last_busy(kdev);
/*
* Take a permanent reference to disable the RPM functionality and drop
* it only when unloading the driver. Use the low level get/put helpers,
* so the driver's own RPM reference tracking asserts also work on
* platforms without RPM support.
*/
if (!HAS_RUNTIME_PM(dev_priv)) {
int ret;
pm_runtime_dont_use_autosuspend(kdev);
ret = pm_runtime_get_sync(kdev);
WARN(ret < 0, "pm_runtime_get_sync() failed: %d\n", ret);
} else {
pm_runtime_use_autosuspend(kdev);
}
/*
* The core calls the driver load handler with an RPM reference held.
* We drop that here and will reacquire it during unloading in
* intel_power_domains_fini().
*/
pm_runtime_put_autosuspend(kdev);
}