Zhi A Wang [Thu, 26 Aug 2021 14:38:34 +0000 (14:38 +0000)]
drm/i915/gvt: fix the usage of ww lock in gvt scheduler.
As the APIs related to ww lock in i915 was changed recently, the usage of
ww lock in GVT-g scheduler needs to be changed accrodingly. We noticed a
deadlock when GVT-g scheduler submits the workload to i915. After some
investigation, it seems the way of how to use ww lock APIs has been
changed. Releasing a ww now requires a explicit i915_gem_ww_ctx_fini().
Joonas Lahtinen [Wed, 27 Oct 2021 09:32:55 +0000 (12:32 +0300)]
drm/i915: Revert 'guc_id' from i915_request tracepoint
Avoid adding backend specific data to the tracepoints outside of
the LOW_LEVEL_TRACEPOINTS kernel config protection. These bits of
information are bound to change depending on the selected submission
method per platform and are not necessarily possible to maintain in
the future.
Fixes: dbf9da8d55ef ("drm/i915/guc: Add trace point for GuC submit") Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: John Harrison <john.c.harrison@intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Matt Roper <matthew.d.roper@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211027093255.66489-1-joonas.lahtinen@linux.intel.com
(cherry picked from commit 64512a66b67e6546e2db15192b3603cd6d58b75c) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
It's not graceful, instead it leads to boot time warning splats in the
case it is supposed to handle gracefully. Apparently the BIOS/GOP
enabling the port we end up skipping leads to state readout
problems. Back to the drawing board.
References: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21255/bat-adlp-4/boot0.txt Fixes: 05734ca2a8f7 ("drm/i915/bios: gracefully disable dual eDP for now") Cc: José Roberto de Souza <jose.souza@intel.com> Cc: Uma Shankar <uma.shankar@intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Swati Sharma <swati2.sharma@intel.com> Reviewed-by: José Roberto de Souza <jose.souza@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211019114334.24643-1-jani.nikula@intel.com
(cherry picked from commit 171c555c2c2664a250cd35680c34c31a6d233dc7) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Imre Deak [Mon, 18 Oct 2021 09:41:51 +0000 (12:41 +0300)]
drm/i915/dp: Ensure max link params are always valid
Atm until the DPCD for a connector is read the max link rate and lane
count params are invalid. If the connector is modeset, in
intel_dp_compute_config(), intel_dp_common_len_rate_limit(max_link_rate)
will return 0, leading to a intel_dp->common_rates[-1] access.
Fix the above by making sure the max link params are always valid.
The above access leads to an undefined behaviour by definition, though
not causing a user visible problem to my best knowledge, see the previous
patch why. Nevertheless it is an undefined behaviour and it triggers a
BUG() in CONFIG_UBSAN builds, hence CC:stable.
Imre Deak [Mon, 18 Oct 2021 14:34:17 +0000 (17:34 +0300)]
drm/i915/dp: Ensure sink rate values are always valid
Atm, there are no sink rate values set for DP (vs. eDP) sinks until the
DPCD capabilities are successfully read from the sink. During this time
intel_dp->num_common_rates is 0 which can lead to a
intel_dp->common_rates[-1] (*)
access, which is an undefined behaviour, in the following cases:
- In intel_dp_sync_state(), if the encoder is enabled without a sink
connected to the encoder's connector (BIOS enabled a monitor, but the
user unplugged the monitor until the driver loaded).
- In intel_dp_sync_state() if the encoder is enabled with a sink
connected, but for some reason the DPCD read has failed.
- In intel_dp_compute_link_config() if modesetting a connector without
a sink connected on it.
- In intel_dp_compute_link_config() if modesetting a connector with a
a sink connected on it, but before probing the connector first.
To avoid the (*) access in all the above cases, make sure that the sink
rate table - and hence the common rate table - is always valid, by
setting a default minimum sink rate when registering the connector
before anything could use it.
I also considered setting all the DP link rates by default, so that
modesetting with higher resolution modes also succeeds in the last two
cases above. However in case a sink is not connected that would stop
working after the first modeset, due to the LT fallback logic. So this
would need more work, beyond the scope of this fix.
As I mentioned in the previous patch, I don't think the issue this patch
fixes is user visible, however it is an undefined behaviour by
definition and triggers a BUG() in CONFIG_UBSAN builds, hence CC:stable.
v2: Clear the default sink rates, before initializing these for eDP.
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4297
References: https://gitlab.freedesktop.org/drm/intel/-/issues/4298 Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: Imre Deak <imre.deak@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Acked-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211018143417.1452632-1-imre.deak@intel.com
(cherry picked from commit 3f61ef9777c0ab0f03f4af0ed6fd3e5250537a8d) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Simon Ser [Thu, 14 Oct 2021 15:35:13 +0000 (15:35 +0000)]
amd/display: remove ChromeOS workaround
This reverts commits ddab8bd788f5 ("drm/amd/display: Fix two cursor duplication
when using overlay") and e7d9560aeae5 ("Revert "drm/amd/display: Fix overlay
validation by considering cursors"").
tl;dr ChromeOS uses the atomic interface for everything except the cursor. This
is incorrect and forces amdgpu to disable some hardware features. Let's revert
the ChromeOS-specific workaround in mainline and allow the Chrome team to keep
it internally in their own tree.
See [1] for more details. This patch is an alternative to [2], which added
ChromeOS detection.
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> Cc: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Cc: Sean Paul <seanpaul@chromium.org> Fixes: ddab8bd788f5 ("drm/amd/display: Fix two cursor duplication when using overlay") Fixes: e7d9560aeae5 ("Revert "drm/amd/display: Fix overlay validation by considering cursors"") Reviewed-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lijo Lazar [Fri, 22 Oct 2021 07:35:25 +0000 (15:35 +0800)]
drm/amd/pm: Disable fan control if not supported
On arcturus, not all platforms use PMFW based fan control. On such
ASICs fan control by PMFW will be disabled in PPTable. Disable hwmon
knobs for fan control also as it is not possible to report or control
fan speed on such platforms through driver.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 19 Oct 2021 15:28:18 +0000 (11:28 -0400)]
drm/amdgpu/smu11.0: add missing IP version check
Add missing check in smu_v11_0_init_display_count(),
Fixes: af3b89d3a639d5 ("drm/amdgpu/smu11.0: convert to IP version checking") Reviewed-by: Guchun Chen <guchun.chen@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Philip Yang [Wed, 20 Oct 2021 21:26:49 +0000 (17:26 -0400)]
drm/amdkfd: debug message to count successfully migrated pages
Not all migrate.cpages returned from migrate_vma_setup can be migrated,
for example non anonymous page, or out of device memory. So after
migrate_vma_pages returns, add debug message to count pages are
successfully migrated which has MIGRATE_PFN_VALID and
MIGRATE_PFN_MIGRATE flag set.
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>
Philip Yang [Fri, 15 Oct 2021 19:57:52 +0000 (15:57 -0400)]
drm/amdkfd: clarify the origin of cpages returned by migration functions
cpages is only updated by migrate_vma_setup. So capture its value at
that point to clarify the significance of the number. The next patch
will add counting of actually migrated pages after migrate_vma_pages for
debug purposes.
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>
Huang Rui [Tue, 19 Oct 2021 06:26:39 +0000 (14:26 +0800)]
drm/amdgpu: remove grbm cam index/data operations for gfx v10
PSP firmware will be responsible for applying the GRBM CAM remapping in
the production. And the GRBM_CAM_INDEX / GRBM_CAM_DATA registers will be
protected by PSP under security policy. So remove it according to the
new security policy.
Signed-off-by: Huang Rui <ray.huang@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
- Fix blank screen booting crashes when CONFIG_CC_OPTIMIZE_FOR_SIZE=y (Hugh)
- Add support for multi-LRC submission in the GuC backend (Matt B)
- Add extra cache flushing before making pages userspace visible (Matt A, Thomas)
- Mark internal GPU object pages dirty so they will be flushed properly (Matt A)
- Move remaining debugfs interfaces i915_wedged/i915_forcewake_user into gt (Andi)
- Replace the unconditional clflushes with drm_clflush_virt_range() (Ville)
- Remove IS_ACTIVE macro completely (Lucas)
- Improve kerneldocs for cache_dirty (Matt A)
- Add missing includes (Lucas)
- Selftest improvements (Matt R, Ran, Matt A)
Dave Airlie [Thu, 21 Oct 2021 19:49:21 +0000 (05:49 +1000)]
Merge tag 'drm-intel-next-2021-10-15' of git://anongit.freedesktop.org/drm/drm-intel into drm-next
UAPI Changes:
- No Functional change, but a clarification around I915_TILING values (Matt).
Driver Changes:
- Changes around async flip VT-d w/a (Ville)
- Delete bogus NULL check in intel_ddi_encoder_destroy (Dan)
- DP link training improvements and DP per-lane driver settings (Ville)
- Free the returned object of acpi_evaluate_dsm (Zenghui)
- Fixes and improvements around DP's UHBR and MST (Jani)
- refactor plane config + pin out (Dave)
- remove unused include in intel_dsi_vbt.c (Lucas)
- some code clean up (Lucas, Jani)
- gracefully disable dual eDP (Jani)
- Remove memory frequency calculation (Jose)
- Fix oops on platforms w/o hpd support (Ville)
- Clean up PXP Kconfig info (Rodrigo)
Matthew Auld [Mon, 18 Oct 2021 17:45:07 +0000 (18:45 +0100)]
drm/i915: mark up internal objects with start_cpu_write
While the pages can't be swapped out, they can be discarded by the shrinker.
Normally such objects are marked with __I915_MADV_PURGED, which can't be
unset, and therefore requires a new object. For kernel internal objects
this is not true, since the madv hint is reset for our special volatile
objects, such that we can re-acquire new pages, if so desired, without
needing a new object. As a result we should probably be paranoid here
and put the object back into the CPU domain when discarding the pages,
and also correctly set cache_dirty, if required.
Matthew Auld [Mon, 18 Oct 2021 17:45:05 +0000 (18:45 +0100)]
drm/i915/shmem: ensure flush during swap-in on non-LLC
On non-LLC platforms, force the flush-on-acquire if this is ever
swapped-in. Our async flush path is not trust worthy enough yet(and
happens in the wrong order), and with some tricks it's conceivable for
userspace to change the cache-level to I915_CACHE_NONE after the pages
are swapped-in, and since execbuf binds the object before doing the
async flush, there is a potential race window.
Matthew Auld [Mon, 18 Oct 2021 17:45:04 +0000 (18:45 +0100)]
drm/i915/userptr: add paranoid flush-on-acquire
Even though userptr objects are always coherent with the GPU, with no
way for userspace to change this with the set_caching ioctl, even on
non-LLC platforms, there is still the 'Bypass LCC' mocs setting, which
might permit reading the contents of main memory directly.
Matthew Auld [Mon, 18 Oct 2021 17:45:03 +0000 (18:45 +0100)]
drm/i915/dmabuf: add paranoid flush-on-acquire
As pointed out by Thomas, we likely need to flush the pages here if the
GPU can read the page contents directly from main memory. Underneath we
don't know what the sg_table is pointing to, so just add a
wbinvd_on_all_cpus() here, for now.
Matthew Auld [Mon, 18 Oct 2021 17:45:01 +0000 (18:45 +0100)]
drm/i915: mark userptr objects as ALLOC_USER
These are userspace objects, so mark them as such. In a later patch it's
useful to determine how paranoid we need to be when managing cache
flushes. In theory no functional changes.
Matthew Auld [Mon, 18 Oct 2021 17:45:00 +0000 (18:45 +0100)]
drm/i915: mark dmabuf objects as ALLOC_USER
These are userspace objects, so mark them as such. In a later patch it's
useful to determine how paranoid we need to be when managing cache
flushes. In theory no functional changes.
Alex Deucher [Tue, 12 Oct 2021 16:22:59 +0000 (12:22 -0400)]
drm/amdgpu/gfx10: fix typo in gfx_v10_0_update_gfx_clock_gating()
Check was incorrectly converted to IP version checking.
Fixes: 4b0ad8425498ba ("drm/amdgpu/gfx10: convert to IP version checking") Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 4 Oct 2021 15:33:00 +0000 (11:33 -0400)]
drm/amdgpu/pm: properly handle sclk for profiling modes on vangogh
When selecting between levels in the force performance levels interface
sclk (gfxclk) was not set correctly for all levels. Select the proper
sclk settings for all levels.
Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1726 Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch introduced a couple of dmesg warnings, this is not a valid
approach anymore. For this reason, we are reverting this patch, and we
need to revert the workaround patch.
v2: squash in unused variable removal (Alex)
Cc: Hanghong Ma <hanghong.ma@amd.com> Cc: Mark Broadworth <mark.broadworth@amd.com> Signed-off-by: Agustin Gutierrez <agustin.gutierrez@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aric Cyr [Mon, 4 Oct 2021 03:42:48 +0000 (23:42 -0400)]
drm/amd/display: 3.2.157
This DC patchset brings improvements in multiple areas. In summary, we
highlight:
* Fix some issues such as DP2 problem, prefetch bandwidth calculation
for DCN3.1 and others.
* Increased Z9 latency and removed z10 save after dsc disable.
* Revert a couple of bad changes.
* Added missing PSR state patch.
Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Aric Cyr <aric.cyr@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jake Wang [Fri, 1 Oct 2021 21:14:21 +0000 (17:14 -0400)]
drm/amd/display: Moved dccg init to after bios golden init
[Why]
bios_golden_init will override dccg_init during init_hw.
[How]
Move dccg_init to after bios_golden_init.
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com> Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Jake Wang <haonan.wang2@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jake Wang [Fri, 1 Oct 2021 20:46:49 +0000 (16:46 -0400)]
drm/amd/display: Removed z10 save after dsc disable
[Why & How]
Z10 save is done during PSR and bootup.
DSC disable does not need to save for Z10.
Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Jake Wang <haonan.wang2@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Yang [Thu, 30 Sep 2021 17:46:45 +0000 (13:46 -0400)]
drm/amd/display: increase Z9 latency to workaround underflow in Z9
[Why]
Z9 latency is higher than when we originally tuned the watermark
parameters, causing underflow. Increasing the value until the latency
issues is resolved.
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Eric Yang <Eric.Yang2@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why]
PSR_STATE2b was introduced on DMCUB side, but not on the driver side,
which caused convert_psr_state helper function to return
PSR_STATE_INVALID. That caused visual lagging during state transition.
[how]
Add PSR_STATE2b to dc_psr_state and convert_psr_state
drm/amd/display: Fix prefetch bandwidth calculation for DCN3.1
[Why]
Prefetch BW calculated is lower than the DML reference because of a
porting error that's excluding cursor and row bandwidth from the
pixel data bandwidth.
[How]
Change the dml_max4 to dml_max3 and include cursor and row bandwidth
in the same calculation as the rest of the pixel data during vactive.
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wenjing Liu [Tue, 28 Sep 2021 17:09:14 +0000 (13:09 -0400)]
drm/amd/display: correct apg audio channel enable golden value
Hardware team has recommended to generically hard code this register to
0xFF as part of the effort to eventually remove this control. However
we set it to 0xF instead.
This causes 4 of audio 8ch to be muted.
Reviewed-by: Ariel Bernstein <eric.bernstein@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Wenjing Liu <wenjing.liu@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Michael Strauss [Tue, 21 Sep 2021 17:35:19 +0000 (13:35 -0400)]
drm/amd/display: Clean Up VPG Low Mem Power
[WHAT]
One of the current VPG power on calls is unnecessary
Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Michael Strauss <michael.strauss@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wenjing Liu [Fri, 24 Sep 2021 20:12:34 +0000 (16:12 -0400)]
drm/amd/display: add DP2.0 debug option to set MST_EN for SST stream
[why]
Some DP2.0 RX requires us to set MST_EN even for SST configuration.
We added this debug option so we can configure this temporary workaround
for the RX.
Reviewed-by: George Shen <george.shen@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Wenjing Liu <wenjing.liu@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jimmy Kizito [Wed, 22 Sep 2021 13:50:38 +0000 (09:50 -0400)]
drm/amd/display: Do not skip link training on DP quick hot plug
[Why]
When rapidly plugging and unplugging a DP sink, detection link
training can be mistakenly skipped.
This is due to the hotplug processing occurring before the
encoder assignment logic has had a chance to process the removal
of a stream. The encoder that would be used for detection link
training is then erroneously reported as already in use and
detection link training is skipped.
[How]
During hot plug processing, only determine a link encoder to be
unavailable for a particular link if it has been assigned to a
different link.
Reviewed-by: Meenakshikumar Somasundaram <Meenakshikumar.Somasundaram@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Jimmy Kizito <Jimmy.Kizito@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jimmy Kizito [Tue, 21 Sep 2021 14:03:05 +0000 (10:03 -0400)]
drm/amd/display: Clear encoder assignment for copied streams
[Why]
When copying a stream, the encoder assigned to it is copied too.
Encoder assignment should only happen when executing the encoder
assignment function link_encs_assign().
[How]
Clear the link encoder pointer for copied stream.
Reviewed-by: Meenakshikumar Somasundaram <Meenakshikumar.Somasundaram@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Jimmy Kizito <Jimmy.Kizito@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jake Wang [Wed, 22 Sep 2021 18:22:59 +0000 (14:22 -0400)]
drm/amd/display: Disable dpp root clock when not being used
[Why & How]
Disable root clock for dpp when not being used.
Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Agustin Gutierrez Sanchez <agustin.gutierrez@amd.com> Signed-off-by: Jake Wang <haonan.wang2@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Roman Li [Fri, 15 Oct 2021 17:16:31 +0000 (13:16 -0400)]
drm/amd/display: Fully switch to dmub for all dcn21 asics
[Why]
On renoir usb-c port stops functioning on resume after f/w update.
New dmub firmware caused regression due to conflict with dmcu.
With new dmub f/w dmcu is superseded and should be disabled.
[How]
- Disable dmcu for all dcn21.
Check dmesg for dmub f/w version.
The old firmware (before regression):
[drm] DMUB hardware initialized: version=0x00000001
All other versions require this patch for renoir.
Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1735 Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Harry Wentland <harry.wentland@amd.com> Tested-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Roman Li <Roman.Li@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
YuBiao Wang [Tue, 19 Oct 2021 21:16:23 +0000 (17:16 -0400)]
drm/amd/amdgpu: Do irq_fini_hw after ip_fini_early
[Why]
drm_irq_uninstall is called in irq_fini_hw so that irq is disabled in sw
stage. SMU (and maybe other IP blocks) fini_hw will call irq_put for
cleanup and the whole cleanup process will be skipped because of
drm->irq_enable = false.
[How]
Move ip_fini_early before irq_fini_hw.
Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com> Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jonathan Kim [Thu, 14 Oct 2021 16:34:27 +0000 (12:34 -0400)]
drm/amdkfd: map gpu hive id to xgmi connected cpu
ROCr needs to be able to identify all devices that have direct access to
fine grain memory, which should include CPUs that are connected to GPUs
over xGMI. The GPU hive ID can be mapped onto the CPU hive ID since the
CPU is part of the hive.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lang Yu [Mon, 11 Oct 2021 05:57:25 +0000 (13:57 +0800)]
drm/amdkfd: Fix an inappropriate error handling in allloc memory of gpu
We should unreference a gem object instead of an amdgpu bo here.
Fixes: fd9a9f8801de ("drm/amdgpu: Use GEM obj reference for KFD BOs") 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>
Ville Syrjälä [Thu, 14 Oct 2021 09:09:39 +0000 (12:09 +0300)]
drm/i915: Convert unconditional clflush to drm_clflush_virt_range()
This one is apparently a "clflush for good measure", so bit more
justification (if you can call it that) than some of the others.
Convert to drm_clflush_virt_range() again so that machines without
clflush will survive the ordeal.
Cc: stable@vger.kernel.org Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Thomas Hellström <thomas.hellstrom@intel.com> #v1 Fixes: 12ca695d2c1e ("drm/i915: Do not share hwsp across contexts any more, v8.") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211014090941.12159-3-ville.syrjala@linux.intel.com Reviewed-by: Dave Airlie <airlied@redhat.com>
Ville Syrjälä [Thu, 14 Oct 2021 09:09:38 +0000 (12:09 +0300)]
drm/i915: Replace the unconditional clflush with drm_clflush_virt_range()
Not all machines have clflush, so don't go assuming they do.
Not really sure why the clflush is even here since hwsp
is supposed to get snooped I thought.
Although in my case we're talking about a i830 machine where
render/blitter snooping is definitely busted. But it might
work for the hswp perhaps. Haven't really reverse engineered
that one fully.
Cc: stable@vger.kernel.org Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Fixes: b436a5f8b6c8 ("drm/i915/gt: Track all timelines created using the HWSP") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20211014090941.12159-2-ville.syrjala@linux.intel.com Reviewed-by: Dave Airlie <airlied@redhat.com>
Rodrigo Vivi [Thu, 14 Oct 2021 17:12:54 +0000 (13:12 -0400)]
drm/i915: Clean up PXP Kconfig info.
During the review I focused on stop the using of the "+"
to reference the newer platforms, but I forgot that we are
in a process of making things more clear and differentiate
graphics and display versions. So, let me to clean up this
a bit. Also, we don't need any version mentioned in the
config menu entry, only in the help.
Matthew Brost [Thu, 14 Oct 2021 17:20:03 +0000 (10:20 -0700)]
drm/i915: Update I915_GEM_BUSY IOCTL to understand composite fences
Parallel submission create composite fences (dma_fence_array) for excl /
shared slots in objects. The I915_GEM_BUSY IOCTL checks these slots to
determine the busyness of the object. Prior to patch it only check if
the fence in the slot was a i915_request. Update the check to understand
composite fences and correctly report the busyness.
Matthew Brost [Thu, 14 Oct 2021 17:20:02 +0000 (10:20 -0700)]
drm/i915: Make request conflict tracking understand parallel submits
If an object in the excl or shared slot is a composite fence from a
parallel submit and the current request in the conflict tracking is from
the same parallel context there is no need to enforce ordering as the
ordering is already implicit. Make the request conflict tracking
understand this by comparing a parallel submit's parent context and
skipping conflict insertion if the values match.
Matthew Brost [Thu, 14 Oct 2021 17:20:01 +0000 (10:20 -0700)]
drm/i915/guc: Handle errors in multi-lrc requests
If an error occurs in the front end when multi-lrc requests are getting
generated we need to skip these in the backend but we still need to
emit the breadcrumbs seqno. An issues arises because with multi-lrc
breadcrumbs there is a handshake between the parent and children to make
forward progress. If all the requests are not present this handshake
doesn't work. To work around this, if multi-lrc request has an error we
skip the handshake but still emit the breadcrumbs seqno.
v2:
(John Harrison)
- Add comment explaining the skipping of the handshake logic
- Fix typos in the commit message
v3:
(John Harrison)
- Fix up some comments about the math to NOP the ring
Matthew Brost [Thu, 14 Oct 2021 17:20:00 +0000 (10:20 -0700)]
drm/i915: Multi-BB execbuf
Allow multiple batch buffers to be submitted in a single execbuf IOCTL
after a context has been configured with the 'set_parallel' extension.
The number batches is implicit based on the contexts configuration.
This is implemented with a series of loops. First a loop is used to find
all the batches, a loop to pin all the HW contexts, a loop to create all
the requests, a loop to submit (emit BB start, etc...) all the requests,
a loop to tie the requests to the VMAs they touch, and finally a loop to
commit the requests to the backend.
A composite fence is also created for the generated requests to return
to the user and to stick in dma resv slots.
No behavior from the existing IOCTL should be changed aside from when
throttling because the ring for a context is full. In this situation,
i915 will now wait while holding the object locks. This change was done
because the code is much simpler to wait while holding the locks and we
believe there isn't a huge benefit of dropping these locks. If this
proves false we can restructure the code to drop the locks during the
wait.
IGT: https://patchwork.freedesktop.org/patch/447008/?series=93071&rev=1
media UMD: https://github.com/intel/media-driver/pull/1252
v2:
(Matthew Brost)
- Return proper error value if i915_request_create fails
v3:
(John Harrison)
- Add comment explaining create / add order loops + locking
- Update commit message explaining different in IOCTL behavior
- Line wrap some comments
- eb_add_request returns void
- Return -EINVAL rather triggering BUG_ON if cmd parser used
(Checkpatch)
- Check eb->batch_len[*current_batch]
v4:
(CI)
- Set batch len if passed if via execbuf args
- Call __i915_request_skip after __i915_request_commit
(Kernel test robot)
- Initialize rq to NULL in eb_pin_timeline
v5:
(John Harrison)
- Fix typo in comments near bb order loops
Matthew Brost [Thu, 14 Oct 2021 17:19:59 +0000 (10:19 -0700)]
drm/i915/guc: Implement no mid batch preemption for multi-lrc
For some users of multi-lrc, e.g. split frame, it isn't safe to preempt
mid BB. To safely enable preemption at the BB boundary, a handshake
between parent and child is needed, syncing the set of BBs at the
beginning and end of each batch. This is implemented via custom
emit_bb_start & emit_fini_breadcrumb functions and enabled by default if
a context is configured by set parallel extension.
Lastly, this patch updates the process descriptor to the correct size as
the memory used in the handshake is directly after the process
descriptor.
v2:
(John Harrison)
- Fix a few comments wording
- Add struture for parent page layout
v3:
(John Harrison)
- A structure for sync semaphore
- Use offsetof to calc address
- Update commit message
v4:
(John Harrison)
- Fix typos in comment explaining memory map of scratch page
Matthew Brost [Thu, 14 Oct 2021 17:19:56 +0000 (10:19 -0700)]
drm/i915/guc: Connect UAPI to GuC multi-lrc interface
Introduce 'set parallel submit' extension to connect UAPI to GuC
multi-lrc interface. Kernel doc in new uAPI should explain it all.
IGT: https://patchwork.freedesktop.org/patch/447008/?series=93071&rev=1
media UMD: https://github.com/intel/media-driver/pull/1252
v2:
(Daniel Vetter)
- Add IGT link and placeholder for media UMD link
v3:
(Kernel test robot)
- Fix warning in unpin engines call
(John Harrison)
- Reword a bunch of the kernel doc
v4:
(John Harrison)
- Add comment why perma-pin is done after setting gem context
- Update some comments / docs for proto contexts
v5:
(John Harrison)
- Rework perma-pin comment
- Add BUG_IN if context is pinned when setting gem context
Matthew Brost [Thu, 14 Oct 2021 17:19:54 +0000 (10:19 -0700)]
drm/i915/guc: Implement multi-lrc reset
Update context and full GPU reset to work with multi-lrc. The idea is
parent context tracks all the active requests inflight for itself and
its children. The parent context owns the reset replaying / canceling
requests as needed.
v2:
(John Harrison)
- Simply loop in find active request
- Add comments to find ative request / reset loop
v3:
(John Harrison)
- s/its'/its/g
- Fix comment when searching for active request
- Reorder if state in __guc_reset_context
v4:
(Kernel test robot)
- Delete unused is_multi_lrc function
Matthew Brost [Thu, 14 Oct 2021 17:19:53 +0000 (10:19 -0700)]
drm/i915/guc: Insert submit fences between requests in parent-child relationship
The GuC must receive requests in the order submitted for contexts in a
parent-child relationship to function correctly. To ensure this, insert
a submit fence between the current request and last request submitted
for requests / contexts in a parent child relationship. This is
conceptually similar to a single timeline.
Matthew Brost [Thu, 14 Oct 2021 17:19:52 +0000 (10:19 -0700)]
drm/i915/guc: Implement multi-lrc submission
Implement multi-lrc submission via a single workqueue entry and single
H2G. The workqueue entry contains an updated tail value for each
request, of all the contexts in the multi-lrc submission, and updates
these values simultaneously. As such, the tasklet and bypass path have
been updated to coalesce requests into a single submission.
v2:
(John Harrison)
- s/wqe/wqi
- Use FIELD_PREP macros
- Add GEM_BUG_ONs ensures length fits within field
- Add comment / white space to intel_guc_write_barrier
(Kernel test robot)
- Make need_tasklet a static function
v3:
(Docs)
- A comment for submission_stall_reason
v4:
(Kernel test robot)
- Initialize return value in bypass tasklt submit function
(John Harrison)
- Add comment near work queue defs
- Add BUILD_BUG_ON to ensure WQ_SIZE is a power of 2
- Update write_barrier comment to talk about work queue
v5:
(John Harrison)
- Fix typo in work queue comment
Parallel contexts are perma-pinned by the upper layers which makes the
backend implementation rather simple. The parent pins the guc_id and
children increment the parent's pin count on pin to ensure all the
contexts are unpinned before we disable scheduling with the GuC / or
deregister the context.
projects / thirdparty / linux.git / log
