While implementing writeback support, odd behavior of WBDELAYCOUNT was
observed with the combination of WB capture and HDMI. The result of the
debugging was that the HDMI sync polarities are not set correctly.
The current code sets the sync polarities going from HDMI WP to DISPC
according to the video mode used, which seems to work normally fine, but
causes problems with WB as WB expects the syncs to be active-high.
This patch changes the HDMI sync polarities so that the DISPC always
gets active-high syncs from HDMI WP, and the HDMI core gets sync
polarities according to the used video mode.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Add support for render nodes in omap driver and allow required
ioctls to be accessible via render nodes.
This enables unprivileged clients to allocate resources like GEM buffers
for rendering their content into. Mode setting (KMS ioctls) is not
allowed using render nodes. These buffers are then shared with
a previleged process (e.g compositor) that has mode setting access.
An example of this use case is Android where the hardware composer is
the only master and has mode setting access. Every other client then
uses render node(e.g /dev/dri/renderD128 to allocate and use its buffers.
Signed-off-by: Hemant Hariyani <hemanthariyani@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
The clock source selection for the LCD outputs is too hardcoded at the
moment. For example, LCD3 is set to use PLL2_1, and PLL2 doesn't exist
on DRA72x SoCs.
There are quite many ways to configure the clocks, even using HDMI PLL
for LCD outputs, but enabling full configuration of the clocks is rather
tricky.
This patch improves the situation a bit by checking if the PLL about to
be used exists, and if not, tries another one.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
mode_config's min_width and min_height are both set to 32, which is
overly restrictive.
The real limits depend on whether we're configuring a crtc or a plane,
but a limit of 8x2 is safe for both cases.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The driver only uses even dividers for hsdiv when pclk >= 100MHz, as odd
dividers can create uneven duty cycle. However, while this holds true
for some dividers like DISPC's LCK and PCK dividers, it is not actually
true for hsdiv.
hsdiv always produces even duty cycle, so the constraint can be removed.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The current driver doesn't expose any of the CRTC HW properties like
background color or transparency key, and sets them at CRTC enable time.
Refactor this into a separate function and call that function from
omap_crtc_atomic_flush(). This is the behavior we want when the
properties can be configured, so this patch makes it easier to add
patches later which implement those properties.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
DRA7 errata i886 (FPDLink PLL Unlocks With Certain SoC PLL M/N Values)
says that FPDLink is sensitive to jitter on the vout clock, and that low
PLL M and N values result in more jitter than high M and N values.
This patch implements a workaround for the problem by changing the PLL
setup to search for clocks starting from high M and N values, instead of
low values. This should not cause any functional change, and only
reduces the jitter.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
drm-misc for 4.12:
Core:
- Removed some fb subsampling dimension checks from core (Ville)
- Some MST slot cleanup (Dhinakaran)
- Extracted drm_debugfs.h & drm_ioctl.h from drmP.h (Daniel)
- Added drm_atomic_helper_shutdown() to compliment suspend/resume counterparts
(Daniel)
- Pipe context through legacy modeset to remove legacy_backoff nasties (Daniel)
- Cleanups around vblank as well as allowing lockless counter reads (Chris W.)
- VGA Switcheroo added to MAINTAINERS with Lukas Wunner as reviewer (Lukas)
Drivers:
- Enhancements to rockchip driver probe (Jeffy) and dsi (Chris Z.)
- Thunderbolt external GPU awareness added (Lukas)
* tag 'drm-misc-next-2017-03-31' of git://anongit.freedesktop.org/git/drm-misc: (63 commits)
apple-gmux: Don't switch external DP port on 2011+ MacBook Pros
drm/nouveau: Don't register Thunderbolt eGPU with vga_switcheroo
drm/amdgpu: Don't register Thunderbolt eGPU with vga_switcheroo
drm/radeon: Don't register Thunderbolt eGPU with vga_switcheroo
PCI: Recognize Thunderbolt devices
MAINTAINERS: Add Lukas Wunner as reviewer for vga_switcheroo
drm: Fix locking gotcha in page_flip ioctl
drm: Clarify the role of plane_state argument to drm_simple update().
drm: Clear e after kfree in drm_mode_page_flip_ioctl
drm: Convert cmpxchg(bool) back to a two step operation
drm/bridge: ti-tfp410: support hpd via gpio
drm: use .hword to represent 16-bit numbers
Revert unrelated part of "drm: simplify the locking in the GETCRTC ioctl"
drm: Fixup failure paths in drm_atomic_helper_set_config
drm: Peek at the current counter/timestamp for vblank queries
drm: Refactor vblank sequence number comparison
drm: vblank cannot be enabled if dev->irq_enabled is false
drm: Mark up accesses of vblank->enabled outside of its spinlock
drm: Make the decision to keep vblank irq enabled earlier
drm/atomic-helper: Remove the backoff hack from set_config
...
The scanout surface size is the smaller of max texture size and
max STDU size.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Most of the display servers today use a single surface to represent the
entire desktop even if it's stretched across multiple screens.
For vmwgfx with STDU, the maximum surface size is limited to the
maximum texture size on the host. On a 2D VM, this limits our
ability to support configurations with more than one 4K monitor.
To get past this limitation, we will now allow using a large DMA buf
as the framebuffer, and take care of blitting contents from this DMA buf
to the display buffer.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
The callbacks we need to provide to many resources are very similar, so
provide a simple resource type with a number of helpers for these
callbacks.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
This reverts commit 2d8e60e8b0 ("drm/vmwgfx: Replace numeric
parameter like 0444 with macro")
The commit belongs to the series of 1285 patches sent to LKML on
2016-08-02, it changes the representation of file permissions from the
octal value "0600" to "S_IRUSR | S_IWUSR".
The general consensus was that the changes does not increase
readability, quite the opposite; 0600 is easier to parse mentally than
S_IRUSR | S_IWUSR.
It also causes argument inconsistency, due to commit 04319d89fb
("drm/vmwgfx: Add an option to change assumed FB bpp") that added
another call to module_param_named() where the permissions are written
as 0600.
Signed-off-by: Øyvind A. Holm <sunny@sunbase.org>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
vmw_ldu_crtc_helper_commit() is not called if
drm_atomic_crtc_needs_modeset() decides nothing related to CRTC timing has
changed.
So a better place for this code is in vmw_ldu_primary_plane_atomic_update()
since we will need to update ld->fb every time the FB is updated.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Pinning fbdev's FB at the start of VRAM prevents X from pinning
its FB. Since for ldu, the fb would be pinned anyway during a
mode set, just skip pinning it in fbdev.
This is not the best solution, but since ldu is not used much
anymore, it seems like a reasonable workaround.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
We can no longer make the assumption that vmw_stdu_update_st() will
be called when there's a valid display surface attached. So
instead of using display_srf for width and height, make a record of
these paremeters when the screen target is first defined.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Now that the legacy path has been tested, turn on the
DRIVER_ATOMIC flag so user mode driver can start going through
the Atomic path.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Switch over to internal atomic API. This completes the atomic
internal atomic switch for all the Display Units.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Switch over to using internal atomic API for mode set.
This removes the legacy set_config API, replacing it with
drm_atomic_helper_set_config(). The DRM helper will use various
vmwgfx-specific atomic functions to set a mode.
DRIVER_ATOMIC capability flag is not yet set, so the user mode
will still use the legacy mode set IOCTL.
v2:
* Avoid a clash between page-flip pinning and setcrtc pinning, modify
the page-flip code to use the page-flip helper and the atomic callbacks.
To enable this, we will need to add a wrapper around atomic_commit.
* Add vmw_kms_set_config() to work around vmwgfx xorg driver bug
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
1. When unsetting a mode, num_connector should be set to zero
2. The pixel_format field needs to be initialized as newer DRM internal
functions checks this field
3. Take the drm_modeset_lock_all() because vmw_fb_kms_detach() can
change current mode
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
This connects the main state object check and commit function.
v2
* Use drm_atomic_helper_commit() rather than a vmwgfx-specific one
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
These helpers won't be called until we flip on the atomic support
flag or set drm_crtc_funcs->set_config to using the atomic
helper.
v2
Use drm_atomic_helper_best_encoder() rather than a vmwgfx-specific one
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Refactor previous FB and cursor plane update code into their
atomic counterparts: check, update, prepare, cleanup, and disable.
These helpers won't be called until we flip on the atomic support
flag or set drm_crtc_funcs->set_config to using the atomic
helper.
v2:
* Removed unnecessary pinning of cursor surface
* Added a few function headers
v3:
* Set clip region equal to the destination region
* Fixed surface pinning policy
* Enable SVGA mode in vmw_sou_primary_plane_prepare_fb
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Atomic mode set requires us to refactor existing vmw_stdu_crtc_set_config
code into sections that check the validity of the new mode, and sections
that actually program the hardware state.
vmw_du_crtc_atomic_check() takes CRTC-related checking code. In a later
patch, vmw_du_primary_plane_atomic_check() will take framebuffer-related
checking code.
These helpers won't be called until we flip on the atomic support
flag or set drm_crtc_funcs->set_config to using the atomic
helper.
v2:
* The state->num_connector is actually the total number of potential
connectors, not just the one associated with the display unit.
The proper one to check is ->connector_mask.
* Add the check to only allow plane state to be the same as crtc state
(Thanks to mlankhorst)
* Make sure to turn on SVGA mode before using VRAM. SVGA mode is
disabled in master_drop if dbdev is not running.
v3:
* Moved dot clock override to crtc_atomic_check
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Add connector handling functions. Start tracking is_implicity in
the connector state. Eventually, this field should be tracked
exclusively in a connector state.
Now that plane and connector states have been created, we can also
activate the code that use CRTC state.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Add plane state handling functions.
We have to keep track of a few plane states so we cannot use the
DRM helper for this.
Created vmw_plane_state along with functions to reset, duplicate,
and destroty it.
v2
* Removed cursor clean up special case
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Daniel Vetter <daniel@ffwll.ch>
Create and Add CRTC state. We currently do not track any properties
or custom states so we can technically use the DRM helpers. Creating
this code just to make potential future additions easier.
Most of the new code will be compiled but not enabled until
plane/connector state handling code is also in place.
This is the first of a series to enable atomic mode set for vmwgfx.
The atomic enabling effort was done in collaboration with Thomas
Hellstrom and the VMware Graphics Team.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Universal support is prerequisite for atomic mode set.
Explicitly create planes for the cursor and the primary FB. With
a functional cursor plane, the DRM will no longer use the legacy
cursor_set2 and cursor_move entry points.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
An external Thunderbolt GPU can neither drive the laptop's panel nor be
powered off by the platform, so there's no point in registering it with
vga_switcheroo. In fact, when the external GPU is runtime suspended,
vga_switcheroo will cut power to the internal discrete GPU, resulting in
a lockup.
Cc: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: http://patchwork.freedesktop.org/patch/msgid/8e733152b13e7c14501ad5af45c1c5c736584111.1489145162.git.lukas@wunner.de
An external Thunderbolt GPU can neither drive the laptop's panel nor be
powered off by the platform, so there's no point in registering it with
vga_switcheroo. In fact, when the external GPU is runtime suspended,
vga_switcheroo will cut power to the internal discrete GPU, resulting in
a lockup. Moreover AMD's Windows driver special-cases Thunderbolt as
well.
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: http://patchwork.freedesktop.org/patch/msgid/701a8e89ce8ac39734736ab779558b6a4042a19e.1489145162.git.lukas@wunner.de
An external Thunderbolt GPU can neither drive the laptop's panel nor be
powered off by the platform, so there's no point in registering it with
vga_switcheroo. In fact, when the external GPU is runtime suspended,
vga_switcheroo will cut power to the internal discrete GPU, resulting in
a lockup. Moreover AMD's Windows driver special-cases Thunderbolt as
well.
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Link: http://patchwork.freedesktop.org/patch/msgid/72d8a9645aece3eff44e116303f0fec8be061c88.1489145162.git.lukas@wunner.de
We want to lock the primary plane, not the cursor (which might be
optional). Real bad case of copy-paste fail, unfortunately our CI
didn't catch that because i915 does have a cursor plane.
Reported-by: Eric Anholt <eric@anholt.net>
Fixes: 29dc0d1de1 ("drm: Roll out acquire context for the page_flip ioctl")
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330204831.8225-1-daniel.vetter@ffwll.ch
Tested-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c:187:2: warning: right shift count >= width of type [enabled by default]
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c:173:2: warning: right shift count >= width of type [enabled by default]
drivers/gpu/drm/amd/amdgpu/vega10_ih.c:106:3: warning: right shift count >= width of type [enabled by default]
v2: Add a space between "&" and "0xff"
Reported by: kbuild-all@01.org
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Like the atomic update hook it's wrapping, the plane_state is the old
one, and the new one is in plane->state. Both msxfb and tinydrm use
it correctly, but I mistook it for the new state in pl111 due to its
naming.
Signed-off-by: Eric Anholt <eric@anholt.net>
Link: http://patchwork.freedesktop.org/patch/msgid/20170320233615.5242-3-eric@anholt.net
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Not sure what the original intention was here, but returning a random piece of
kernel memory to userspace because we didn't set the value at all is clearly
not a good idea.
This patch disallows reading the register and returns
a proper error code instead.
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Tom St Denis <tom.stdenis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Remove duplicate mmCP_CPF_BUSY_STAT from the allowed registers.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
With the explicit retry loop static analyzers get confused by the
control flow and believe that e could be accessed after kfree. That's
not possible, but it's non-obvious, so let's clear it to NULL.
We already cleared e = NULL at the top of the function, so this is all
in line.
Cc: Julia Lawall <julia.lawall@lip6.fr>
Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Fixes: 29dc0d1de1 ("drm: Roll out acquire context for the page_flip ioctl")
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Sean Paul <seanpaul@chromium.org>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330133253.29500-1-daniel.vetter@ffwll.ch
ARM v6 (at least) only allows cmpxchg on 32bit variables which doesn't
always include the bool type.
drivers/built-in.o: In function `vblank_disable_and_save':
imx-ocotp.c:(.text+0xb45e8): undefined reference to `__bad_cmpxchg'
Makefile:986: recipe for target 'vmlinux' failed
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Reported-by: Leonard Crestez <leonard.crestez@nxp.com>
Fixes: 43dc7fe2b2 ("drm: Mark up accesses of vblank->enabled outside of its spinlock")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170330140832.32377-1-chris@chris-wilson.co.uk
On some boards the hpd pin of a hdmi connector is wired up to a gpio
pin. Since in the DRM world the tfp410 driver is responsible for
handling the connector, add support for hpd gpios in this very driver.
Reviewed-by: Jyri Sarha <jsarha@ti.com>
Signed-off-by: Christopher Spinrath <christopher.spinrath@rwth-aachen.de>
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Link: http://patchwork.freedesktop.org/patch/msgid/2e47786ab3d04078ae70d0c4064f7c4d@rwthex-s1-b.rwth-ad.de
v2 of the commit 2c77bb29d3 ("drm: simplify the locking in the GETCRTC ioctl")
accidentally introduced a unrelated change in intel_display.c, revert the
unrelated change.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Fixes: 2c77bb29d3 ("drm: simplify the locking in the GETCRTC ioctl")
Reported-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/6be47261-475f-c190-af56-c136677246d9@linux.intel.com
I've screwed this up when removing the legacy backoff hack.
Fixes: 38b6441e4e ("drm/atomic-helper: Remove the backoff hack from set_config")
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Sean Paul <seanpaul@chromium.org>
Cc: David Airlie <airlied@linux.ie>
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20170329174136.10330-1-daniel.vetter@ffwll.ch
SRIOV can support for loading ucode with PSP block, enable it.
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Disable pp block if device is vf.
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Monk Liu <Monk.Liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Confirm if sys driver and sOS are already been loaded through sOS
sign register, skip loading sys driver and sOS if finding the sign.
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On gfx9 hardware the value is not wrapped and is a 64-bit value. So
we reduce it modulo the ring size.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
(v2) use buf_mask instead of computing on the fly
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fix the start/end address calculation for address ranges that span
multiple page directories in amdgpu_vm_alloc_levels.
Add error messages if page tables aren't found. Otherwise the page
table update would just fail silently.
v2:
* Change WARN_ON to WARN_ON_ONCE
* Move masking of high address bits to caller
* Add range-check for "from" and "to"
v3:
* Replace WARN_ON_ONCE in get_pt with pr_err in caller
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
adev->family is not initialized yet when amdgpu_get_block_size is
called. Use adev->asic_type instead.
Minimum VM size is 512GB, not 256GB, for a single page table entry
in the root page table.
gmc_v9_0_vm_init is called after adev->vm_manager.max_pfn is
initialized. Move the minimum VM-size enforcement ahead of max_pfn
initializtion. Cast to 64-bit before the left-shift.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
With 4-level page tables the maximum VM size is 256TB. That's 64G
pages, which can't be represented in 32-bit.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
That register is marked deprecated, reading it results in a bus error.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
v2: fix for all sdma engines
v3: squash in fix for SI/CI
Signed-off-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1. security firmware loading has moved to sw init, so this code
is useless.
2. it seems that driver could not call request_firmware on
kernel 2.6, when S3 resume. for request firmware depends on
userspace, at this time, userspace is freeze.
Signed-off-by: Jim Qu <Jim.Qu@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When VRAM pressue and trigger huge evictions there is performance drop,
this patch fix it.
Signed-off-by: Roger.He <Hongbo.He@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Clean up the VM fault message format and use rate-limiting similar
to other ASICs.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The RB harvest registers are not necessary, the driver already
exposes this info via the info ioctl. GB_BACKEND_MAP has
been deprecated since SI and is not relevant to the RB mapping.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Required for SR-IOV and saves MMIO transactions.
v2: drop cached RB harvest registers
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We check the mem config register to make sure it's been
programmed by the vbios to determine if we need to post
so we check for a non-0 value. However, when the asic
comes out of reset, we may see all ones here, so check
for that too.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This used by DAL ISR logic for VBLANK handling.
Signed-off-by: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1) Adapt to vulkan:
Now use double SWITCH BUFFER to replace the 128 nops w/a,
because when vulkan introduced, umd can insert 7 ~ 16 IBs
per submit which makes 256 DW size cannot hold the whole
DMAframe (if we still insert those 128 nops), CP team suggests
use double SWITCH_BUFFERs, instead of tricky 128 NOPs w/a.
2) To fix the CE VM fault issue when MCBP introduced:
Need one more COND_EXEC wrapping IB part (original one us
for VM switch part).
this change can fix vm fault issue caused by below scenario
without this change:
>CE passed original COND_EXEC (no MCBP issued this moment),
proceed as normal.
>DE catch up to this COND_EXEC, but this time MCBP issued,
thus DE treats all following packages as NOP. The following
VM switch packages now looks just as NOP to DE, so DE
dosen't do VM flush at all.
>Now CE proceeds to the first IBc, and triggers VM fault,
because DE didn't do VM flush for this DMAframe.
3) change estimated alloc size for gfx9.
with new DMAframe scheme, we need modify emit_frame_size
for gfx9
4) No need to insert 128 nops after gfx8 vm flush anymore
because there was double SWITCH_BUFFER append to vm flush,
and for gfx7 we already use double SWITCH_BUFFER following
after vm_flush so no change needed for it.
5) Change emit_frame_size for gfx8
v2: squash in BUG removal from Monk
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
those MC_VM registers won't be programed by VBIOS in VF
so driver is responsible to programe them.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
ring_write_multiple should use buf_mask instead of ptr_mask
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
for SRIOV we cannot use access register when in IRQ routine
with regular KIQ method
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
no CG for SRIOV on SOC15
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
no hw_fini for SRIOV, otherwise other VF will be affected
no CG for SRIOV
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
must set minor_update.enable before write smaller value
to wptr/doorbell, so for sriov we need set that register
bit in hw_init period.
this could fix the SDMA ring test fail after guest reboot
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
ring->buf_mask need be set prior to ring_clear_ring invoke
and fix ring_clear_ring as well which should use buf_mask
instead of ptr_mask
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
this must be invoked during early init
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Apply the new IB during IB emit for SRIOV with MCBP
v2: agd: use define instead of magic number
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
set bit 21 of IB.control filed to actually enable
MCBP for SRIOV
v2:
add flag for preemption enable bit for soc15 and use
this flag instead of hardcode.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
when MCBP enabled for gfx8, the cond_exec must also
be implemented, otherwise there will be odds to meet
cross engine (ce and me) deadlock when world switch
happens.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1,the check is only appliable for SRIOV GFX engine.
2,use chunk_ib instead of ib.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Ken Wang <Qingqing.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
to prevent submit two or more IBs with PREEMPT flags.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
should use chunk_ib instead of ib, otherwise the logic
is incorrect.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Ken Wang <Qingqing.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Also, add the fence of the clear operations to the BO to ensure that
the underlying memory can only be re-used after all PTEs pointing to
it have been cleared.
This avoids the following sequence of events that could be triggered
by user space:
1. Submit a CS that accesses some BO _without_ adding that BO to the
buffer list.
2. Free that BO.
3. Some other task re-uses the memory underlying the BO.
4. The CS is submitted to the hardware and accesses memory that is
now already in use by somebody else.
By clearing the page tables immediately in step 2, a GPU VM fault will
be triggered in step 4 instead of wild memory accesses.
v2: use amdgpu_bo_fence directly
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allocate and free page directories on demand.
V2:
a. clear entries allocation
b. fix entries index calculation
c. need alloc sub level even parent bo was allocated
Signed-off-by: Christian König <christian.koenig@amd.com> (v1)
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (v1)
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com> (v2)
Acked-by: Alex Deucher <alexander.deucher@amd.com> (v2)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Not the best solution, but good enough for now.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Update all levels of the page directory.
V2:
a. sub level pdes always are written to incorrect place.
b. sub levels need to update regardless of parent updates.
Signed-off-by: Christian König <christian.koenig@amd.com> (V1)
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> (V1)
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com> (V2)
Acked-by: Alex Deucher <alexander.deucher@amd.com> (V2)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Move all levels to the end after command submission.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
All page directory levels should be in place after this.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allows us to get the size for all levels as well.
v2: agd: fix warning
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
No functional change, but the base for multi level page tables.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Needs to be filled with handling.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This way we save passing it through the different functions.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Decribes better what this is used for.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
for power readings.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The ring structure already has what we need.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Avoids passing around additional parameters during setup.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tomi Valkeinen