There was a change(below) target for such issue:
d82e2c249c ("drm/amdgpu: Fix crash on device remove/driver unload")
But the fix for VI ASICs was missing there. This is a supplement for
that.
Fixes: d82e2c249c ("drm/amdgpu: Fix crash on device remove/driver unload")
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Properly handle SI DC support when CONFIG_DRM_AMD_DC_SI is not
set.
Fixes: f7f12b2582 ("drm/amdgpu: default to true in amdgpu_device_asic_has_dc_support")
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If a kfd_bo was shared (e.g. a dmabuf export), the original kfd_bo may be
freed when the amdgpu_bo still lives on. Free the kfd_bo struct in the
release_notify callback then the amdgpu_bo is freed.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-By: Ramesh Errabolu <Ramesh.Errabolu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When kfd need to be reset, sent command to HWS might cause hang and get unnecessary timeout.
This change try not to touch HW in pre_reset and keep queues to be in the evicted state
when the reset is done, so they are not put back on the runlist. These queues will be destroied
on process termination.
Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
As part of the ib padding process, accessing the RLC_SPM_* register may
trigger gfx hang. Since gfxoff may be already kicked during the whole period.
To address that, we manually toggle gfx on/off around the RLC_SPM_*
register access.
This can resolve the gfx hang issue observed on running Talos with RDP launched
in parallel.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Previously there was a check based on chip # for chips that aligned to
>=CHIP_NAVI10 to have RLC stopped as part of DPMS check. This was because
of gfxclk being controlled by RLC in the newer designs.
As part of IP version checking though, this got changed to match IP
version for SMU. Because Renoir designs also include smu11 that meant
that even GFX9 started to stop RLC earlier.
Adjust to match GFX IP version instead of SMU IP version to restore the
previous behavior.
Fixes: a8967967f6 ("drm/amdgpu/amdgpu_smu: convert to IP version checking")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
csb bo is not unpinned in gfx 9. It will lead to pin_count leak on
driver unload.
[How]
Call bo_free_kernel corresponding to bo_create_kernel in
gfx_rlc_init_csb. This will also unify the code path with other gfx
versions.
Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
For Vega10, disabling gart of gfxhub could mess up KIQ and PSP
under sriov mode, and lead to DMAR on host side.
[How]
Skip writing GMC registers under sriov.
Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
BOs need to be reserved before they are added or removed, so ensure that
they are reserved during kfd_mem_attach and kfd_mem_detach
Signed-off-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]:
When we call hmm_range_fault to map memory after a migration, we don't
expect memory to be migrated again as a result of hmm_range_fault. The
driver ensures that all memory is in GPU-accessible locations so that
no migration should be needed. However, there is one corner case where
hmm_range_fault can unexpectedly cause a migration from DEVICE_PRIVATE
back to system memory due to a write-fault when a system memory page in
the same range was mapped read-only (e.g. COW). Ranges with individual
pages in different locations are usually the result of failed page
migrations (e.g. page lock contention). The unexpected migration back
to system memory causes a deadlock from recursive locking in our
driver.
[How]:
Creating a task reference new member under svm_range_list struct.
Setting this with "current" reference, right before the hmm_range_fault
is called. This member is checked against "current" reference at
svm_migrate_to_ram callback function. If equal, the migration will be
ignored.
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>
There is no reason to allow for partial buffers from userspace in our
debugfs. In this particular case callers will zero out the wr_buf but if
callers in the future don't do that we might be looking at corrupt data.
Linus puts it better than I can in
https://lkml.org/lkml/2021/10/26/993
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
During Z10, root clock gating and memory low power registers needs to
to be restored if optimization is enabled in driver.
[How]
Added new DMUB boot option for root clock gating and memory low power.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Jake Wang <haonan.wang2@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
HPO is only used for DP2.0. HPO HW control should be
disable when not being used to save power.
[How]
Shutdown HPO HW control during init hw.
Shutdown HPO HW control during stream disable.
Enable HPO HW control during stream enable if DP2.0.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Jake Wang <haonan.wang2@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
State can be cleared without removing individual streams (by
calling dc_remove_stream_from_ctx()). This can leave the
encoder assignment module in an incoherent state and cause
future assignments to be incorrect.
[How]
Clear encoder assignments when committing 0 streams or
re-initializing hardware.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Jimmy Kizito <Jimmy.Kizito@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
In scenario when 1 display connected with pipe split (2 pipes in use)
and 3 new displays simultaneously hotplugged via MST hub (4 pipes in use),
mpcc may get reprogram to other vtg, remaining busy.
In this case waiting for mpcc idle timeouts with error like this:
[drm] REG_WAIT timeout 1us * 100000 tries - mpc2_assert_idle_mpcc
RIP: 0010:mpc2_assert_mpcc_idle_before_connect
Call Trace:
dcn20_update_mpcc
dcn20_program_front_end_for_ctx
dc_commit_state
amdgpu_dm_atomic_commit_tail
...
[How]
Add pipe split change condition to disable dangling plane.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Roman Li <Roman.Li@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
It was found that the system would hang on a dummy pstate when playing
4k60 videos on a 1080p 390Hz monitor.
[HOW]
Properly select the dummy_pstate_latency_ms when firmware assisted
memory clock switching is enabled instead of assuming that the highest
latency would work for every monitor timing.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Felipe Clark <felclark@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
If timing and bpp of displays on mst hub are not changed,
pbn, slot_num for displays should not be changed. Linux
user mode may initiate atomic_check with different display
configuration after set mode finished. This will call to
amdgpu_dm to re-compute payload, slot_num of displays and
saved to dm_connect_state. stream->timing.flags.dsc, pbn,
slot_num are updated to values which may be different from
that were used for set mode. when dsc hub with 3 4k@60hz dp
connected, 3 dsc engines are enabled. timing.flags.dsc = 1.
timing.flags.dsc are changed to 0 due to atomic check. when
dsc hub is unplugged, amdgpu driver check timing.flags.dsc
for last mode set and find out flags.dsc = 0, then does not
disable dsc.
[HOW]
check status of displays on dsc mst hubs. re-compute pbn,
slot_num, timing.flags.dsc only if there is mode, connect
or enable/disable change.
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Hersen Wu <hersenwu@amd.com>
Reviewed-by: Mikita Lipski <Mikita.Lipski@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This better aligns that the caller can make a mistake with the buffer
and -EINVAL should be returned, but if the hardware doesn't support
the feature it should be -EOPNOTSUPP.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This was added by commit bd8dcea93a ("drm/amd/pm: add callbacks to
read/write sysfs file pp_power_profile_mode") but the feature was
deprecated from PMFW. Remove it from the driver.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Previously Renoir compiler gfx target version was forced to Raven.
Update driver side for completeness.
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This is more useful when talking to the SMU team to have the information
in this format, save one less step to manually do it.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If some pages fail to migrate to system memory, don't update
prange->actual_loc = 0. This prevents endless CPU page faults after
partial migration failures due to contested page locks.
Migration to RAM must be complete during migrations from VRAM to VRAM and
during evictions. Implement retry and fail if the migration to RAM fails.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stack and heap pages tend to be shared by many small allocations.
Concurrent access by CPU and GPU is therefore likely, which can lead to
thrashing. Avoid this by setting the preferred location to system memory.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The preferred location should be used as the migration destination
whenever it is accessible by the faulting GPU. System memory is always
accessible. Peer memory is accessible if it's in the same XGMI hive.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aldebaran has different register mask definitions for
regiter MC_VM_XGMI_LFB_CNTL. Use the correct masks
to interpret fields of this register.
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
In advance tdr mode, the real bad job will be resubmitted twice, while
in drm_sched_resubmit_jobs_ext, there's a dma_fence_put, so the bad job
is put one more time than other jobs.
[How]
Adding dma_fence_get before resbumit job in
amdgpu_device_recheck_guilty_jobs and put the fence for normal jobs
Signed-off-by: Jingwen Chen <Jingwen.Chen2@amd.com>
Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
- DPIA MST slot registers are not programmed during payload
allocation and hence MST does not work with DPIA.
- HPD RX interrupts are not handled for DPIA.
[How]
- Added inbox command to program the MST slots whenever
payload allocation happens for DPIA links.
- Added support for handling HPD RX interrupts
Signed-off-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
CVE-2021-42327 was fixed by:
commit f23750b5b3
Author: Thelford Williams <tdwilliamsiv@gmail.com>
Date: Wed Oct 13 16:04:13 2021 -0400
drm/amdgpu: fix out of bounds write
but amdgpu_dm_debugfs.c contains more of the same issue so fix the
remaining ones.
v2:
* Add missing fix in dp_max_bpc_write (Harry Wentland)
Fixes: 918698d5c2 ("drm/amd/display: Return the number of bytes parsed than allocated")
Signed-off-by: Patrik Jakobsson <pjakobsson@suse.de>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DMUB binary is common for both A0 and B0. Hence, driver should
notify FW about the support for DPIA in B0.
[How]
Added dpia_supported bit in dmub_fw_boot_options and will be set
only for B0.
Assign dpia_supported to true before dm_dmub_hw_init
in B0 case.
v2: fix build without CONFIG_DRM_AMD_DC_DCN (Alex)
Signed-off-by: Jude Shih <shenshih@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Notify data from outbox corrupt, the notify type should be 2 (HPD) instead of 0
(No data). We copied the address instead of the value. The memory might be
freed in the end of outbox IRQ
[How]
We should allocate the memory of notify and copy the whole content from outbox to
hpd handle function
Fixes: 88f52b1fff ("drm/amd/display: Support for SET_CONFIG processing with DMUB")
Signed-off-by: Jude Shih <shenshih@amd.com>
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
