5.15-rc1 crashes with blank screen when booting up on two ThinkPads
using i915. Bisections converge convincingly, but arrive at different
and suprising "culprits", none of them the actual culprit.
netconsole (with init_netconsole() hacked to call i915_init() when
logging has started, instead of by module_init()) tells the story:
kernel BUG at drivers/gpu/drm/i915/i915_sw_fence.c:245!
with RSI: ffffffff814d408b pointing to sw_fence_dummy_notify().
I've been building with CONFIG_CC_OPTIMIZE_FOR_SIZE=y, and that
function needs to be 4-byte aligned.
Fixes: 62eaf0ae21 ("drm/i915/guc: Support request cancellation")
Signed-off-by: Hugh Dickins <hughd@google.com>
Tested-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the s2idle stress test sdma resume fail occasionally,in the
failed case GPU is in the gfxoff state.This issue may introduce
by firmware miss handle doorbell S/R and now temporary fix the issue
by forcing exit gfxoff for sdma resume.
Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
On GFX9+, format modifiers are always enabled and ensure the
frame-buffers can be scanned out at ADDFB2 time.
On GFX8-, format modifiers are not supported and no other check
is performed. This means ADDFB2 IOCTLs will succeed even if the
tiling isn't supported for scan-out, and will result in garbage
displayed on screen [1].
Fix this by adding a check for tiling flags for GFX8 and older.
The check is taken from radeonsi in Mesa (see how is_displayable
is populated in gfx6_compute_surface).
Changes in v2: use drm_WARN_ONCE instead of drm_WARN (Michel)
[1]: https://github.com/swaywm/wlroots/issues/3185
Signed-off-by: Simon Ser <contact@emersion.fr>
Acked-by: Michel Dänzer <mdaenzer@redhat.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Harry Wentland <hwentlan@amd.com>
Cc: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Cc: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
[Why]
ASSR is dependent on Signed PSP Verstage to enable Content
Protection for eDP panels. Unsigned PSP verstage is used
during development phase causing ASSR to FAIL.
As a result, link training is performed with
DP_PANEL_MODE_DEFAULT instead of DP_PANEL_MODE_EDP for
eDP panels that causes display flicker on some panels.
[How]
- Do not change panel mode, if ASSR is disabled
- Just report and continue to perform eDP link training
with right settings further.
Signed-off-by: Praful Swarnakar <Praful.Swarnakar@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Currently we blow up in trace_dma_fence_init, when calling into
get_driver_name or get_timeline_name, since both the engine and context
might be NULL(or contain some garbage address) in the case of newly
allocated slab objects via the request ctor. Note that we also use
SLAB_TYPESAFE_BY_RCU here, which allows requests to be immediately
freed, but delay freeing the underlying page by an RCU grace period.
With this scheme requests can be re-allocated, at the same time as they
are also being read by some lockless RCU lookup mechanism.
In the ctor case, which is only called for new slab objects(i.e allocate
new page and call the ctor for each object) it's safe to reset the
context/engine prior to calling into dma_fence_init, since we can be
certain that no one is doing an RCU lookup which might depend on peeking
at the engine/context, like in active_engine(), since the object can't
yet be externally visible.
In the recycled case(which might also be externally visible) the request
refcount always transitions from 0->1 after we set the context/engine
etc, which should ensure it's valid to dereference the engine for
example, when doing an RCU list-walk, so long as we can also increment
the refcount first. If the refcount is already zero, then the request is
considered complete/released. If it's non-zero, then the request might
be in the process of being re-allocated, or potentially still in flight,
however after successfully incrementing the refcount, it's possible to
carefully inspect the request state, to determine if the request is
still what we were looking for. Note that all externally visible
requests returned to the cache must have zero refcount.
One possible fix then is to move dma_fence_init out from the request
ctor. Originally this was how it was done, but it was moved in:
commit 855e39e65c
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon Feb 3 09:41:48 2020 +0000
drm/i915: Initialise basic fence before acquiring seqno
where it looks like intel_timeline_get_seqno() relied on some of the
rq->fence state, but that is no longer the case since:
commit 12ca695d2c
Author: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Date: Tue Mar 23 16:49:50 2021 +0100
drm/i915: Do not share hwsp across contexts any more, v8.
intel_timeline_get_seqno() could also be cleaned up slightly by dropping
the request argument.
Moving dma_fence_init back out of the ctor, should ensure we have enough
of the request initialised in case of trace_dma_fence_init.
Functionally this should be the same, and is effectively what we were
already open coding before, except now we also assign the fence->lock
and fence->ops, but since these are invariant for recycled
requests(which might be externally visible), and will therefore already
hold the same value, it shouldn't matter.
An alternative fix, since we don't yet have a fully initialised request
when in the ctor, is just setting the context/engine as NULL, but this
does require adding some extra handling in get_driver_name etc.
v2(Daniel):
- Try to make the commit message less confusing
Fixes: 855e39e65c ("drm/i915: Initialise basic fence before acquiring seqno")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Michael Mason <michael.w.mason@intel.com>
Cc: Daniel Vetter <daniel@ffwll.ch>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210921134202.3803151-1-matthew.auld@intel.com
(cherry picked from commit be988eaee1)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Nested grids in grid-table cells are not specified as proper ReST
constructs.
Commit 572f2a5cd9 ("drm/i915/guc: Update firmware to v62.0.0")
added a couple of kerneldoc tables of the form:
+---+-------+------------------------------------------------------+
| 1 | 31:0 | +------------------------------------------------+ |
+---+-------+ | | |
|...| | | Embedded `HXG Message`_ | |
+---+-------+ | | |
| n | 31:0 | +------------------------------------------------+ |
+---+-------+------------------------------------------------------+
For "make htmldocs", they happen to work as one might expect,
but they are incompatible with "make latexdocs" and "make pdfdocs",
and cause the generated gpu.tex file to become incomplete and
unbuildable by xelatex.
Restore the compatibility by removing those nested grids in the tables.
Size comparison of generated gpu.tex:
Sphinx 2.4.4 Sphinx 4.2.0
v5.14: 3238686 3841631
v5.15-rc1: 376270 432729
with this fix: 3377846 3998095
Fixes: 572f2a5cd9 ("drm/i915/guc: Update firmware to v62.0.0")
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Akira Yokosawa <akiyks@gmail.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/4a227569-074f-c501-58bb-d0d8f60a8ae9@gmail.com
(cherry picked from commit 017792a041)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Device manager releases device-specific resources when a driver
disconnects from a device, devm_memunmap_pages and
devm_release_mem_region calls in svm_migrate_fini are redundant.
It causes below warning trace after patch "drm/amdgpu: Split
amdgpu_device_fini into early and late", so remove function
svm_migrate_fini.
BUG: https://gitlab.freedesktop.org/drm/amd/-/issues/1718
WARNING: CPU: 1 PID: 3646 at drivers/base/devres.c:795
devm_release_action+0x51/0x60
Call Trace:
? memunmap_pages+0x360/0x360
svm_migrate_fini+0x2d/0x60 [amdgpu]
kgd2kfd_device_exit+0x23/0xa0 [amdgpu]
amdgpu_amdkfd_device_fini_sw+0x1d/0x30 [amdgpu]
amdgpu_device_fini_sw+0x45/0x290 [amdgpu]
amdgpu_driver_release_kms+0x12/0x30 [amdgpu]
drm_dev_release+0x20/0x40 [drm]
release_nodes+0x196/0x1e0
device_release_driver_internal+0x104/0x1d0
driver_detach+0x47/0x90
bus_remove_driver+0x7a/0xd0
pci_unregister_driver+0x3d/0x90
amdgpu_exit+0x11/0x20 [amdgpu]
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If svm migration init failed to create pgmap for device memory, set
pgmap type to 0 to disable device SVM support capability.
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Update the current state as boot state during dpm initialization.
During the subsequent initialization, set_power_state gets called to
transition to the final power state. set_power_state refers to values
from the current state and without current state populated, it could
result in NULL pointer dereference.
For ex: on platforms where PCI speed change is supported through ACPI
ATCS method, the link speed of current state needs to be queried before
deciding on changing to final power state's link speed. The logic to query
ATCS-support was broken on certain platforms. The issue became visible
when broken ATCS-support logic got fixed with commit
f9b7f3703f ("drm/amdgpu/acpi: make ATPX/ATCS structures global (v2)").
Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1698
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
For xnack off, restore work dma unmap previous system memory page, and
dma map the updated system memory page to update GPU mapping, this is
not dma mapping leaking, remove the WARN_ONCE for dma mapping leaking.
prange->dma_addr store the VRAM page pfn after the range migrated to
VRAM, should not dma unmap VRAM page when updating GPU mapping or
remove prange. Add helper svm_is_valid_dma_mapping_addr to check VRAM
page and error cases.
Mask out SVM_RANGE_VRAM_DOMAIN flag in dma_addr before calling amdgpu vm
update to avoid BUG_ON(*addr & 0xFFFF00000000003FULL), and set it again
immediately after. This flag is used to know the type of page later to
dma unmapping system memory page.
Fixes: 1d5dbfe6c0 ("drm/amdkfd: classify and map mixed svm range pages in GPU")
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
SVM range may includes multiple VMAs with different vm_flags, if prange
page index is the last page of the VMA offset + npages, update GPU
mapping to create GPU page table with same VMA access permission.
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Using an empty macro expansion as a conditional expression
produces a W=1 warning:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dce_aux.c: In function 'dce_aux_transfer_with_retries':
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dce_aux.c:775:156: error: suggest braces around empty body in an 'if' statement [-Werror=empty-body]
775 | "dce_aux_transfer_with_retries: AUX_RET_SUCCESS: AUX_TRANSACTION_REPLY_I2C_OVER_AUX_DEFER");
| ^
drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dce_aux.c:783:155: error: suggest braces around empty body in an 'if' statement [-Werror=empty-body]
783 | "dce_aux_transfer_with_retries: AUX_RET_SUCCESS: AUX_TRANSACTION_REPLY_I2C_OVER_AUX_NACK");
| ^
Expand it to "do { } while (0)" instead to make the expression
more robust and avoid the warning.
Fixes: 56aca23093 ("drm/amd/display: Add AUX I2C tracing.")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
nvkm test builds fail with the following error.
drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c: In function 'nvkm_control_mthd_pstate_info':
drivers/gpu/drm/nouveau/nvkm/engine/device/ctrl.c:60:35: error: overflow in conversion from 'int' to '__s8' {aka 'signed char'} changes value from '-251' to '5'
The code builds on most architectures, but fails on parisc where ENOSYS
is defined as 251.
Replace the error code with -ENODEV (-19). The actual error code does
not really matter and is not passed to userspace - it just has to be
negative.
Fixes: 7238eca4cf ("drm/nouveau: expose pstate selection per-power source in sysfs")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull drm fixes from Dave Airlie:
"Slightly busier than usual rc2, but mostly scattered amdgpu fixes,
some i915 and etnaviv resolves an MMU/runtime PM blowup.
amdgpu:
- UBSAN fix
- Powerplay table update fix
- Fix use after free in BO moves
- Debugfs init fixes
- vblank workqueue fixes for headless devices
- FPU fixes
- sysfs_emit fixes
- SMU updates for cyan skillfish
- Backlight fixes when DMCU is not initialized
- DP MST fixes
- HDCP compliance fix
- Link training fix
- Runtime pm fix
- Panel orientation fixes
- Display GPUVM fix for yellow carp
- Add missing license
amdkfd:
- Drop PCI atomics requirement if proper firmware is available
- Suspend/resume fixes for IOMMUv2 cases
radeon:
- AGP fix
i915:
- Propagate DP link training error returns
- Use max link params for eDP 1.3 and earlier
- Build warning fixes
- Gem selftest fixes
- Ensure wakeref is held before hardware access
etnaviv:
- MMU context vs runtime PM fix"
* tag 'drm-fixes-2021-09-17' of git://anongit.freedesktop.org/drm/drm: (44 commits)
drm/amdgpu/display: add a proper license to dc_link_dp.c
drm/amd/display: Fix white screen page fault for gpuvm
amd/display: enable panel orientation quirks
drm/amdgpu: Demote TMZ unsupported log message from warning to info
drm/amdgpu: Drop inline from amdgpu_ras_eeprom_max_record_count
drm/amd/pm: fix runpm hang when amdgpu loaded prior to sound driver
drm/radeon: pass drm dev radeon_agp_head_init directly
drm/amdgpu: move iommu_resume before ip init/resume
drm/amdgpu: add amdgpu_amdkfd_resume_iommu
drm/amdkfd: separate kfd_iommu_resume from kfd_resume
drm/amd/display: Link training retry fix for abort case
drm/amd/display: Fix unstable HPCP compliance on Chrome Barcelo
drm/amd/display: dsc mst 2 4K displays go dark with 2 lane HBR3
drm/amd/display: Get backlight from PWM if DMCU is not initialized
drm/amdkfd: make needs_pcie_atomics FW-version dependent
drm/amdgpu: add manual sclk/vddc setting support for cyan skilfish(v3)
drm/amdgpu: add some pptable funcs for cyan skilfish(v3)
drm/amdgpu: update SMU driver interface for cyan skilfish(v3)
drm/amdgpu: update SMU PPSMC for cyan skilfish
drm/amdgpu: fix sysfs_emit/sysfs_emit_at warnings(v2)
...
With the new static annotation, the compiler warns when the functions
are actually unused:
drivers/gpu/drm/rockchip/cdn-dp-core.c:1123:12: error: 'cdn_dp_resume' defined but not used [-Werror=unused-function]
1123 | static int cdn_dp_resume(struct device *dev)
| ^~~~~~~~~~~~~
Mark them __maybe_unused to suppress that warning as well.
[ Not so 'new' static annotations any more, and I removed the part of
the patch that added __maybe_unused to cdn_dp_suspend(), because it's
used by the shutdown/remove code.
So only the resume function ends up possibly unused if CONFIG_PM isn't
set - Linus ]
Fixes: 7c49abb4c2 ("drm/rockchip: cdn-dp-core: Make cdn_dp_core_suspend/resume static")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The memory allocated for a DMA fence could be leaked if the code failed
to allocate the waiter object. Make sure to release the fence allocation
on failure.
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
The DEFINE_SPINLOCK macro creates a global spinlock symbol that is visible
to the whole kernel. This is unintended in the code, fix it.
Fixes: 687db2207b ("gpu: host1x: Add DMA fence implementation")
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
All jobs are failing on Tegra20 because it doesn't use IOMMU and mapping
function uses size of mapping that is zero instead of BO size, fix it.
Fixes: d7c591bc1a ("drm/tegra: Implement new UAPI")
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
[Why]
The "base_addr_is_mc_addr" field was added for dcn3.1 support but
pa_config was never updated to set it to false.
Uninitialized memory causes it to be set to true which results in
address mistranslation and white screen.
[How]
Use memset to ensure all fields are initialized to 0 by default.
Fixes: 64b1d0e8d5 ("drm/amd/display: Add DCN3.1 HWSEQ")
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
This patch allows panel orientation quirks from DRM core to be
used. They attach a DRM connector property "panel orientation"
which indicates in which direction the panel has been mounted.
Some machines have the internal screen mounted with a rotation.
Since the panel orientation quirks need the native mode from the
EDID, check for it in amdgpu_dm_connector_ddc_get_modes.
Signed-off-by: Simon Ser <contact@emersion.fr>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Harry Wentland <hwentlan@amd.com>
Cc: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
As the user cannot do anything about the unsupported Trusted Memory Zone
(TMZ) feature, do not warn about it, but make it informational, so
demote the log level from warning to info.
Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This was unusual; normally, inline functions are declared static as
well, and defined in a header file if used by multiple compilation
units. The latter would be more involved in this case, so just drop
the inline declaration for now.
Fixes compile failure building for ppc64le on RHEL 8:
In file included from ../drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h:32,
from ../drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:33:
../drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c: In function ‘amdgpu_ras_recovery_init’:
../drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h:90:17: error: inlining failed in call
to ‘always_inline’ ‘amdgpu_ras_eeprom_max_record_count’: function body not available
90 | inline uint32_t amdgpu_ras_eeprom_max_record_count(void);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:1985:34: note: called from here
1985 | max_eeprom_records_len = amdgpu_ras_eeprom_max_record_count();
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fixes: c84d46707e "drm/amdgpu: validate bad page threshold in ras(v3)"
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Current RUNPM mechanism relies on PMFW to master the timing for BACO
in/exit. And that needs cooperation from sound driver for dstate
change notification for function 1(audio). Otherwise(on sound driver
missing), BACO cannot be kicked in correctly and hang will be observed
on RUNPM exit.
By switching back to legacy message way on sound driver missing,
we are able to fix the runpm hang observed for the scenario below:
amdgpu driver loaded -> runpm suspend kicked -> sound driver loaded
Signed-off-by: Evan Quan <evan.quan@amd.com>
Reported-and-tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
[Why]
If link training is aborted, it shall be retried if sink is present.
[How]
Check hpd status to find out whether sink is present or not. If sink is
present, then link training shall be tried again with same settings.
Otherwise, link training shall be aborted.
Reviewed-by: Jimmy Kizito <Jimmy.Kizito@amd.com>
Acked-by: Mikita Lipski <mikita.lipski@amd.com>
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Intermittently, there presents two occurrences of 0 stream
commits in a single HPD event. Current HDCP sequence does
not consider such scenerio, and will thus disable HDCP.
[How]
Add condition check to include stream remove and re-enable
case for HDCP enable.
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Acked-by: Mikita Lipski <mikita.lipski@amd.com>
Signed-off-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
