When the KMD sends a CLIENT_RESET request to GuC (as part of the
suspend sequence), GuC will mark the CTB buffer as 'UNUSED'. If the
KMD then checked the CTB queue, it would see a non-zero status value
and report the buffer as corrupted.
Technically, no G2H messages should be received once the CLIENT_RESET
has been sent. However, if a context was outstanding on an engine then
it would get reset and a reset notification would be sent. So, don't
actually treat UNUSED as a catastrophic error. Just flag it up as
unexpected and keep going.
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728024225.2363663-7-John.C.Harrison@Intel.com
Having semaphores results in different behavior when a dependent request
is cancelled. In the case of semaphores the request could be on the HW
and complete successfully while without the request is held in the
driver and the error from the dependent request is propagated. Fix
live_preempt_cancel to take this behavior into account.
Also update live_preempt_cancel to use new function intel_context_ban
rather than intel_context_set_banned.
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728024225.2363663-3-John.C.Harrison@Intel.com
Pull powerpc fixes from Michael Ellerman:
- Re-enable the new amdgpu display engine for powerpc, as long as the
compiler is correctly configured.
- Disable stack variable initialisation in prom_init to fix GCC 12
allmodconfig.
Thanks to Dan Horák and Sudip Mukherjee.
* tag 'powerpc-5.19-6' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
drm/amdgpu: Re-enable DCN for 64-bit powerpc
powerpc/64s: Disable stack variable initialisation for prom_init
Move the PIXPLLC code into per-model source files and wire it up
with per-model callbacks. No functional changes.
The PIXPLLC pixel-clock is part of the CRTC, but really separate
hardware that varies with each model of the G200. Move the PIXPLLC
code for each model into the per-model source file and call it from
CRTC helpers via device functions.
This allows to remove struct mgag200_pll and the related code. The
new callbacks behave like the CRTC's atomic_check and atomic_enable
functions.
v3:
* clean up style
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-12-tzimmermann@suse.de
Move the BMC-related code into its own file and wire it up with device
callbacks.
While programming a new display mode, G200EW3 and G200WB have to de-
synchronize with the BMC. Synchronization is done via VIDRST pins
and controlled via VRSTEN and HRSTEN bits. Move the BMC code behind
a serviceable interface and call it from the CRTC's enable and
disable functions.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-11-tzimmermann@suse.de
Drop simple-KMS in favor of regular atomic helpers. Makes the code
more modular and hence better to adapt to per-model requirements.
The simple-KMS helpers provide few extra features, so the patch is
mostly about open-coding what simple-KMS does. The simple-KMS helpers
do mix up plane and CRTC state. Changing to regular atomic helpers
requires to split some of the simple-pipe functions into per-plane
and per-CRTC code
No functional changes.
v3:
* always run drm_atomic_helper_check_plane_state()
* clean up style
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-8-tzimmermann@suse.de
Store the primary plane's color format in the CRTC state and use
it for programming the CRTC's gamma LUTs.
Gamma tables (i.e., color management) are provided by the CRTC, but
depend in the primary plane's color format. Store the format in the
CRTC state and use it. This has not been an issue with simple-KMS
helpers, which mix-up plane and CRTC state to some extent. For using
regular atomic helpers, it's necessary to distinguish between the two.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-6-tzimmermann@suse.de
Hold I/O-register lock in atomic_commit_tail to protect all pipeline
updates at once. Protects against concurrent I/O access in get-modes
helper.
Complex modesetting operations involve mode changes, plane updates and
possibly BMC updates. Make all this atomic wrt to reading display modes
via EDID. It's not so much an issue with simple-KMS helpers, but will
become necessary for using regular atomic helpers.
v4:
* remove empty line
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-5-tzimmermann@suse.de
Provide an init function for each model's DAC registers. Remove
the shared helper.
The code for initializing the DAC registers consisted of a large
table of default value, plus many exceptions for the various G200
models. Providing a per-model implementation makes if more readable.
At some point, some of the initialization should probably move into
the modesetting code.
v2:
* don't duplicate DAC values unnecessarily (Sam, Jocelyn)
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220728124103.30159-3-tzimmermann@suse.de
The igt_check_drm_format_min_pitch() function had a lot of
KUNIT_EXPECT_* calls, all of which ended up allocating and initializing
various test assertion structures on the stack.
This behavior was producing -Wframe-larger-than warnings on PowerPC, i386,
and MIPS architectures, such as:
drivers/gpu/drm/tests/drm_format_test.c: In function 'igt_check_drm_format_min_pitch':
drivers/gpu/drm/tests/drm_format_test.c:271:1: error: the frame size of
3712 bytes is larger than 2048 bytes
So, the igt_check_drm_format_min_pitch() test case was split into three
smaller functions: one testing single plane formats, one testing
multi-planar formats, and the other testing tiled formats.
Fixes: 0421bb0baa ("drm: selftest: convert drm_format selftest to KUnit")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Melissa Wen <melissa.srw@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220729124726.748221-1-mairacanal@riseup.net
Users may request that pages from an OpenCL SVM allocation be migrated
to the GPU with clEnqueueSVMMigrateMem(). In Nouveau this will call into
nouveau_dmem_migrate_vma() to do the migration. If the total range to be
migrated exceeds SG_MAX_SINGLE_ALLOC the pages will be migrated in
chunks of size SG_MAX_SINGLE_ALLOC. However a typo in updating the
starting address means that only the first chunk will get migrated.
Fix the calculation so that the entire range will get migrated if
possible.
Signed-off-by: Alistair Popple <apopple@nvidia.com>
Fixes: e3d8b08904 ("drm/nouveau/svm: map pages after migration")
Reviewed-by: Ralph Campbell <rcampbell@nvidia.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220720062745.960701-1-apopple@nvidia.com
Cc: <stable@vger.kernel.org> # v5.8+
mes self test rely on vm mapping, move it after
drm sched re-started so that vm mapping can work
during gpu reset.
Signed-off-by: Jack Xiao <Jack.Xiao@amd.com>
Acked-and-tested-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This extra call trace dump comes out in every gpu reset.
And it gives people a wrong impression that something
went wrong. Although actually there was not.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
All references to struct freesync_context were removed, so remove the
struct freesync_context itself and its entry on struct dc_stream_state.
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The kgd_mem memory allocated in amdgpu_amdkfd_gpuvm_import_dmabuf()
is not freed properly.
Explicitly free it in amdgpu_amdkfd_gpuvm_free_memory_of_gpu()
under condition "mem->bo->kfd_bo != mem".
Suggested-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
Unified memory with xnack off should be tracked, as userptr mappings
and legacy allocations do. To avoid oversuscribe system memory when
xnack off.
[How]
Exposing functions reserve_mem_limit and unreserve_mem_limit to SVM
API and call them on every prange creation and free.
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The "ret" variable needs to be signed for the error handling to work.
Fixes: 2ca97adccd ("drm/amd/display: Add Synaptics Fifo Reset Workaround")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
