linux.git/drivers/gpu, branch v6.6-rc7 Linux kernel source tree Merge tag 'mediatek-drm-fixes-20231017' of https://git.kernel.org/pub/scm/linux/kernel/git/chunkuang.hu/linux into drm-fixes 2023-10-20T04:24:35+00:00 Dave Airlie airlied@redhat.com 2023-10-20T04:23:25+00:00 8b35ce3f7a9699e7580527fe4510d77f2a35f02d Mediatek DRM Fixes - 20231017 1. Correctly free sg_table in gem prime vmap Signed-off-by: Dave Airlie <airlied@redhat.com> From: Chun-Kuang Hu <chunkuang.hu@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20231016233659.3639-1-chunkuang.hu@kernel.org 
Mediatek DRM Fixes - 20231017

1. Correctly free sg_table in gem prime vmap

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20231016233659.3639-1-chunkuang.hu@kernel.org
Merge tag 'drm-intel-fixes-2023-10-19' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes 2023-10-20T04:21:44+00:00 Dave Airlie airlied@redhat.com 2023-10-20T04:21:16+00:00 280bd84f0743f95849dbaadb12e9c0e532398d8d - Fix display issue that was blocking S0ix (Khaled) - Retry gtt fault when out of fence registers (Ville) Signed-off-by: Dave Airlie <airlied@redhat.com> From: Rodrigo Vivi <rodrigo.vivi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/ZTFXbo6M5bWp/hTU@intel.com 
- Fix display issue that was blocking S0ix (Khaled)
- Retry gtt fault when out of fence registers (Ville)

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ZTFXbo6M5bWp/hTU@intel.com
Merge tag 'drm-misc-fixes-2023-10-19' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes 2023-10-20T04:07:58+00:00 Dave Airlie airlied@redhat.com 2023-10-20T04:07:41+00:00 d43c76c8208c1c5e83bcf79c3f08974b231188f3 Short summary of fixes pull: amdgpu: - Disable AMD_CTX_PRIORITY_UNSET bridge: - ti-sn65dsi86: Fix device lifetime edid: - Add quirk for BenQ GW2765 ivpu: - Extend address range for MMU mmap nouveau: - DP-connector fixes - Documentation fixes panel: - Move AUX B116XW03 into panel-simple scheduler: - Eliminate DRM_SCHED_PRIORITY_UNSET ttm: - Fix possible NULL-ptr deref in cleanup Signed-off-by: Dave Airlie <airlied@redhat.com> From: Thomas Zimmermann <tzimmermann@suse.de> Link: https://patchwork.freedesktop.org/patch/msgid/20231019114605.GA22540@linux-uq9g 
Short summary of fixes pull:

amdgpu:
- Disable AMD_CTX_PRIORITY_UNSET

bridge:
- ti-sn65dsi86: Fix device lifetime

edid:
- Add quirk for BenQ GW2765

ivpu:
- Extend address range for MMU mmap

nouveau:
- DP-connector fixes
- Documentation fixes

panel:
- Move AUX B116XW03 into panel-simple

scheduler:
- Eliminate DRM_SCHED_PRIORITY_UNSET

ttm:
- Fix possible NULL-ptr deref in cleanup

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20231019114605.GA22540@linux-uq9g
drm/amdgpu: Reserve fences for VM update 2023-10-19T22:56:57+00:00 Felix Kuehling Felix.Kuehling@amd.com 2023-07-17T19:28:52+00:00 316baf09d355aec1179981b6dfe28eba50c5ee5b In amdgpu_dma_buf_move_notify reserve fences for the page table updates in amdgpu_vm_clear_freed and amdgpu_vm_handle_moved. This fixes a BUG_ON in dma_resv_add_fence when using SDMA for page table updates. 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> 
In amdgpu_dma_buf_move_notify reserve fences for the page table updates
in amdgpu_vm_clear_freed and amdgpu_vm_handle_moved. This fixes a BUG_ON
in dma_resv_add_fence when using SDMA for page table updates.

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>
drm/amdgpu: Fix possible null pointer dereference 2023-10-19T22:56:50+00:00 Felix Kuehling Felix.Kuehling@amd.com 2023-10-17T20:51:03+00:00 51b79f33817544e3b4df838d86e8e8e4388ff684 abo->tbo.resource may be NULL in amdgpu_vm_bo_update. Fixes: 180253782038 ("drm/ttm: stop allocating dummy resources during BO creation") 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> 
abo->tbo.resource may be NULL in amdgpu_vm_bo_update.

Fixes: 180253782038 ("drm/ttm: stop allocating dummy resources during BO creation")
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>
drm/i915: Retry gtt fault when out of fence registers 2023-10-18T02:08:54+00:00 Ville Syrjälä ville.syrjala@linux.intel.com 2023-10-12T13:28:01+00:00 e339c6d628fe66c9b64bf31040a55770952aec57 If we can't find a free fence register to handle a fault in the GMADR range just return VM_FAULT_NOPAGE without populating the PTE so that userspace will retry the access and trigger another fault. Eventually we should find a free fence and the fault will get properly handled. A further improvement idea might be to reserve a fence (or one per CPU?) for the express purpose of handling faults without having to retry. But that would require some additional work. Looks like this may have gotten broken originally by commit 39965b376601 ("drm/i915: don't trash the gtt when running out of fences") as that changed the errno to -EDEADLK which wasn't handle by the gtt fault code either. But later in commit 2feeb52859fc ("drm/i915/gt: Fix -EDEADLK handling regression") I changed it again to -ENOBUFS as -EDEADLK was now getting used for the ww mutex dance. So this fix only makes sense after that last commit. Cc: stable@vger.kernel.org Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/9479 Fixes: 2feeb52859fc ("drm/i915/gt: Fix -EDEADLK handling regression") Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231012132801.16292-1-ville.syrjala@linux.intel.com Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com> (cherry picked from commit 7f403caabe811b88ab0de3811ff3f4782c415761) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> 
If we can't find a free fence register to handle a fault in the GMADR
range just return VM_FAULT_NOPAGE without populating the PTE so that
userspace will retry the access and trigger another fault. Eventually
we should find a free fence and the fault will get properly handled.

A further improvement idea might be to reserve a fence (or one per CPU?)
for the express purpose of handling faults without having to retry. But
that would require some additional work.

Looks like this may have gotten broken originally by
commit 39965b376601 ("drm/i915: don't trash the gtt when running out of fences")
as that changed the errno to -EDEADLK which wasn't handle by the gtt
fault code either. But later in commit 2feeb52859fc ("drm/i915/gt: Fix
-EDEADLK handling regression") I changed it again to -ENOBUFS as -EDEADLK
was now getting used for the ww mutex dance. So this fix only makes
sense after that last commit.

Cc: stable@vger.kernel.org
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/9479
Fixes: 2feeb52859fc ("drm/i915/gt: Fix -EDEADLK handling regression")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231012132801.16292-1-ville.syrjala@linux.intel.com
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
(cherry picked from commit 7f403caabe811b88ab0de3811ff3f4782c415761)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
drm/i915/cx0: Only clear/set the Pipe Reset bit of the PHY Lanes Owned 2023-10-18T02:08:49+00:00 Khaled Almahallawy khaled.almahallawy@intel.com 2023-10-05T00:13:10+00:00 5e4c16fe08c8b894b258f4110349dc9b642669f9 Currently, with MFD/pin assignment D, the driver clears the pipe reset bit of lane 1 which is not owned by display. This causes the display to block S0iX. By not clearing this bit for lane 1 and keeping whatever default, S0ix started to work. This is already what the driver does at the end of the phy lane reset sequence (Step#8) Bspec: 65451 Fixes: 619a06dba6fa ("drm/i915/mtl: Reset only one lane in case of MFD") Cc: Mika Kahola <mika.kahola@intel.com> Cc: Gustavo Sousa <gustavo.sousa@intel.com> Signed-off-by: Khaled Almahallawy <khaled.almahallawy@intel.com> Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com> Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20231005001310.154396-1-khaled.almahallawy@intel.com (cherry picked from commit 4a07f063d20c46524f00976f4537de72d9f31c4e) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> 
Currently, with MFD/pin assignment D, the driver clears the pipe reset bit
of lane 1 which is not owned by display. This causes the display
to block S0iX.

By not clearing this bit for lane 1 and keeping whatever default, S0ix
started to work. This is already what the driver does at the end
of the phy lane reset sequence (Step#8)

Bspec: 65451
Fixes: 619a06dba6fa ("drm/i915/mtl: Reset only one lane in case of MFD")
Cc: Mika Kahola <mika.kahola@intel.com>
Cc: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Khaled Almahallawy <khaled.almahallawy@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231005001310.154396-1-khaled.almahallawy@intel.com
(cherry picked from commit 4a07f063d20c46524f00976f4537de72d9f31c4e)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
gpu/drm: Eliminate DRM_SCHED_PRIORITY_UNSET 2023-10-18T00:35:38+00:00 Luben Tuikov luben.tuikov@amd.com 2023-10-17T02:48:56+00:00 fa8391ad68c16716e2c06ada397e99ceed2fb647 Eliminate DRM_SCHED_PRIORITY_UNSET, value of -2, whose only user was amdgpu. Furthermore, eliminate an index bug, in that when amdgpu boots, it calls drm_sched_entity_init() with DRM_SCHED_PRIORITY_UNSET, which uses it to index sched->sched_rq[]. Cc: Alex Deucher <Alexander.Deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Acked-by: Alex Deucher <Alexander.Deucher@amd.com> Link: https://lore.kernel.org/r/20231017035656.8211-2-luben.tuikov@amd.com 
Eliminate DRM_SCHED_PRIORITY_UNSET, value of -2, whose only user was
amdgpu. Furthermore, eliminate an index bug, in that when amdgpu boots, it
calls drm_sched_entity_init() with DRM_SCHED_PRIORITY_UNSET, which uses it to
index sched->sched_rq[].

Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Alex Deucher <Alexander.Deucher@amd.com>
Link: https://lore.kernel.org/r/20231017035656.8211-2-luben.tuikov@amd.com
drm/amdgpu: Unset context priority is now invalid 2023-10-18T00:35:38+00:00 Luben Tuikov luben.tuikov@amd.com 2023-10-17T02:24:39+00:00 eab0261967aeab528db4d0a51806df8209aec179 A context priority value of AMD_CTX_PRIORITY_UNSET is now invalid--instead of carrying it around and passing it to the Direct Rendering Manager--and it becomes AMD_CTX_PRIORITY_NORMAL in amdgpu_ctx_ioctl(), the gateway to context creation. Cc: Alex Deucher <Alexander.Deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Acked-by: Alex Deucher <Alexander.Deucher@amd.com> Link: https://lore.kernel.org/r/20231017035656.8211-1-luben.tuikov@amd.com 
A context priority value of AMD_CTX_PRIORITY_UNSET is now invalid--instead of
carrying it around and passing it to the Direct Rendering Manager--and it
becomes AMD_CTX_PRIORITY_NORMAL in amdgpu_ctx_ioctl(), the gateway to context
creation.

Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Alex Deucher <Alexander.Deucher@amd.com>
Link: https://lore.kernel.org/r/20231017035656.8211-1-luben.tuikov@amd.com
drm/mediatek: Correctly free sg_table in gem prime vmap 2023-10-16T15:44:03+00:00 Chen-Yu Tsai wenst@chromium.org 2023-10-04T08:32:24+00:00 dcc583c225e659d5da34b4ad83914fd6b51e3dbf The MediaTek DRM driver implements GEM PRIME vmap by fetching the sg_table for the object, iterating through the pages, and then vmapping them. In essence, unlike the GEM DMA helpers which vmap when the object is first created or imported, the MediaTek version does it on request. Unfortunately, the code never correctly frees the sg_table contents. This results in a kernel memory leak. On a Hayato device with a text console on the internal display, this results in the system running out of memory in a few days from all the console screen cursor updates. Add sg_free_table() to correctly free the contents of the sg_table. This was missing despite explicitly required by mtk_gem_prime_get_sg_table(). Also move the "out" shortcut label to after the kfree() call for the sg_table. Having sg_free_table() together with kfree() makes more sense. The shortcut is only used when the object already has a kernel address, in which case the pointer is NULL and kfree() does nothing. Hence this change causes no functional change. Fixes: 3df64d7b0a4f ("drm/mediatek: Implement gem prime vmap/vunmap function") Cc: <stable@vger.kernel.org> Signed-off-by: Chen-Yu Tsai <wenst@chromium.org> Reviewed-by: CK Hu <ck.hu@mediatek.com> Link: https://patchwork.kernel.org/project/dri-devel/patch/20231004083226.1940055-1-wenst@chromium.org/ Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org> 
The MediaTek DRM driver implements GEM PRIME vmap by fetching the
sg_table for the object, iterating through the pages, and then
vmapping them. In essence, unlike the GEM DMA helpers which vmap
when the object is first created or imported, the MediaTek version
does it on request.

Unfortunately, the code never correctly frees the sg_table contents.
This results in a kernel memory leak. On a Hayato device with a text
console on the internal display, this results in the system running
out of memory in a few days from all the console screen cursor updates.

Add sg_free_table() to correctly free the contents of the sg_table. This
was missing despite explicitly required by mtk_gem_prime_get_sg_table().

Also move the "out" shortcut label to after the kfree() call for the
sg_table. Having sg_free_table() together with kfree() makes more sense.
The shortcut is only used when the object already has a kernel address,
in which case the pointer is NULL and kfree() does nothing. Hence this
change causes no functional change.

Fixes: 3df64d7b0a4f ("drm/mediatek: Implement gem prime vmap/vunmap function")
Cc: <stable@vger.kernel.org>
Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
Link: https://patchwork.kernel.org/project/dri-devel/patch/20231004083226.1940055-1-wenst@chromium.org/
Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>