kernel/linux.git/drivers/gpu/drm, branch v6.13.6

amdgpu/pm/legacy: fix suspend/resume issues

2025-03-07T17:27:11+00:00

commit 91dcc66b34beb72dde8412421bdc1b4cd40e4fb8 upstream. resume and irq handler happily races in set_power_state() * amdgpu_legacy_dpm_compute_clocks() needs lock * protect irq work handler * fix dpm_enabled usage v2: fix clang build, integrate Lijo's comments (Alex) Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/2524 Fixes: 3712e7a49459 ("drm/amd/pm: unified lock protections in amdgpu_dpm.c") Reviewed-by: Lijo Lazar Tested-by: Maciej S. Szmigiero # on Oland PRO Signed-off-by: chr[] Signed-off-by: Alex Deucher (cherry picked from commit ee3dc9e204d271c9c7a8d4d38a0bce4745d33e71) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amd/display: Fix HPD after gpu reset

2025-03-07T17:27:07+00:00

commit 4de141b8b1b7991b607f77e5f4580e1c67c24717 upstream. [Why] DC is not using amdgpu_irq_get/put to manage the HPD interrupt refcounts. So when amdgpu_irq_gpu_reset_resume_helper() reprograms all of the IRQs, HPD gets disabled. [How] Use amdgpu_irq_get/put() for HPD init/fini in DM in order to sync refcounts Cc: Mario Limonciello Cc: Alex Deucher Reviewed-by: Mario Limonciello Reviewed-by: Aurabindo Pillai Signed-off-by: Roman Li Signed-off-by: Zaeem Mohamed Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher (cherry picked from commit f3dde2ff7fcaacd77884502e8f572f2328e9c745) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amd/display: add a quirk to enable eDP0 on DP1

2025-03-07T17:27:07+00:00

commit b5f7242e49b927cfe488b369fa552f2eff579ef1 upstream. [why] some board designs have eDP0 connected to DP1, need a way to enable support_edp0_on_dp1 flag, otherwise edp related features cannot work [how] do a dmi check during dm initialization to identify systems that require support_edp0_on_dp1. Optimize quirk table with callback functions to set quirk entries, retrieve_dmi_info can set quirks according to quirk entries Cc: Mario Limonciello Reviewed-by: Mario Limonciello Reviewed-by: Nicholas Kazlauskas Signed-off-by: Yilin Chen Signed-off-by: Zaeem Mohamed Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher (cherry picked from commit f6d17270d18a6a6753fff046330483d43f8405e4) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amd/display: Disable PSR-SU on eDP panels

2025-03-07T17:27:07+00:00

commit e8863f8b0316d8ee1e7e5291e8f2f72c91ac967d upstream. [Why] PSR-SU may cause some glitching randomly on several panels. [How] Temporarily disable the PSR-SU and fallback to PSR1 for all eDP panels. Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3388 Cc: Mario Limonciello Cc: Alex Deucher Reviewed-by: Sun peng Li Signed-off-by: Tom Chung Signed-off-by: Roman Li Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher (cherry picked from commit 6deeefb820d0efb0b36753622fb982d03b37b3ad) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: init return value in amdgpu_ttm_clear_buffer

2025-03-07T17:27:07+00:00

commit d3c7059b6a8600fc62cd863f1ea203b8675e63e1 upstream. Otherwise an uninitialized value can be returned if amdgpu_res_cleared returns true for all regions. Possibly closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3812 Fixes: a68c7eaa7a8f ("drm/amdgpu: Enable clear page functionality") Signed-off-by: Pierre-Eric Pelloux-Prayer Acked-by: Alex Deucher Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit 7c62aacc3b452f73a1284198c81551035fac6d71) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: disable BAR resize on Dell G5 SE

2025-03-07T17:27:07+00:00

commit 099bffc7cadff40bfab1517c3461c53a7a38a0d7 upstream. There was a quirk added to add a workaround for a Sapphire RX 5600 XT Pulse that didn't allow BAR resizing. However, the quirk caused a regression with runtime pm on Dell laptops using those chips, rather than narrowing the scope of the resizing quirk, add a quirk to prevent amdgpu from resizing the BAR on those Dell platforms unless runtime pm is disabled. v2: update commit message, add runpm check Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1707 Fixes: 907830b0fc9e ("PCI: Add a REBAR size quirk for Sapphire RX 5600 XT Pulse") Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher (cherry picked from commit 5235053f443cef4210606e5fb71f99b915a9723d) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdkfd: Preserve cp_hqd_pq_control on update_mqd

2025-03-07T17:27:06+00:00

commit 3502ab5022bb5ef1edd063bdb6465a8bf3b46e66 upstream. When userspace applications call AMDKFD_IOC_UPDATE_QUEUE. Preserve bitfields that do not need to be modified as they contain flags to track queue states that are used by CP FW. Signed-off-by: David Yat Sin Reviewed-by: Jay Cornwall Signed-off-by: Alex Deucher (cherry picked from commit 8150827990b709ab5a40c46c30d21b7f7b9e9440) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/fbdev-dma: Add shadow buffering for deferred I/O

2025-03-07T17:27:06+00:00

commit 3603996432997f7c88da37a97062a46cda01ac9d upstream. DMA areas are not necessarily backed by struct page, so we cannot rely on it for deferred I/O. Allocate a shadow buffer for drivers that require deferred I/O and use it as framebuffer memory. Fixes driver errors about being "Unable to handle kernel NULL pointer dereference at virtual address" or "Unable to handle kernel paging request at virtual address". The patch splits drm_fbdev_dma_driver_fbdev_probe() in an initial allocation, which creates the DMA-backed buffer object, and a tail that sets up the fbdev data structures. There is a tail function for direct memory mappings and a tail function for deferred I/O with the shadow buffer. It is no longer possible to use deferred I/O without shadow buffer. It can be re-added if there exists a reliably test for usable struct page in the allocated DMA-backed buffer object. Signed-off-by: Thomas Zimmermann Reported-by: Nuno Gonçalves CLoses: https://lore.kernel.org/dri-devel/CAEXMXLR55DziAMbv_+2hmLeH-jP96pmit6nhs6siB22cpQFr9w@mail.gmail.com/ Tested-by: Nuno Gonçalves Fixes: 5ab91447aa13 ("drm/tiny/ili9225: Use fbdev-dma") Cc: Thomas Zimmermann Cc: # v6.11+ Reviewed-by: Simona Vetter Reviewed-by: Javier Martinez Canillas Link: https://patchwork.freedesktop.org/patch/msgid/20241211090643.74250-1-tzimmermann@suse.de Signed-off-by: Greg Kroah-Hartman

drm/xe/userptr: fix EFAULT handling

2025-03-07T17:27:06+00:00

commit a9f4fa3a7efa65615ff7db13023ac84516e99e21 upstream. Currently we treat EFAULT from hmm_range_fault() as a non-fatal error when called from xe_vm_userptr_pin() with the idea that we want to avoid killing the entire vm and chucking an error, under the assumption that the user just did an unmap or something, and has no intention of actually touching that memory from the GPU. At this point we have already zapped the PTEs so any access should generate a page fault, and if the pin fails there also it will then become fatal. However it looks like it's possible for the userptr vma to still be on the rebind list in preempt_rebind_work_func(), if we had to retry the pin again due to something happening in the caller before we did the rebind step, but in the meantime needing to re-validate the userptr and this time hitting the EFAULT. This explains an internal user report of hitting: [ 191.738349] WARNING: CPU: 1 PID: 157 at drivers/gpu/drm/xe/xe_res_cursor.h:158 xe_pt_stage_bind.constprop.0+0x60a/0x6b0 [xe] [ 191.738551] Workqueue: xe-ordered-wq preempt_rebind_work_func [xe] [ 191.738616] RIP: 0010:xe_pt_stage_bind.constprop.0+0x60a/0x6b0 [xe] [ 191.738690] Call Trace: [ 191.738692] [ 191.738694] ? show_regs+0x69/0x80 [ 191.738698] ? __warn+0x93/0x1a0 [ 191.738703] ? xe_pt_stage_bind.constprop.0+0x60a/0x6b0 [xe] [ 191.738759] ? report_bug+0x18f/0x1a0 [ 191.738764] ? handle_bug+0x63/0xa0 [ 191.738767] ? exc_invalid_op+0x19/0x70 [ 191.738770] ? asm_exc_invalid_op+0x1b/0x20 [ 191.738777] ? xe_pt_stage_bind.constprop.0+0x60a/0x6b0 [xe] [ 191.738834] ? ret_from_fork_asm+0x1a/0x30 [ 191.738849] bind_op_prepare+0x105/0x7b0 [xe] [ 191.738906] ? dma_resv_reserve_fences+0x301/0x380 [ 191.738912] xe_pt_update_ops_prepare+0x28c/0x4b0 [xe] [ 191.738966] ? kmemleak_alloc+0x4b/0x80 [ 191.738973] ops_execute+0x188/0x9d0 [xe] [ 191.739036] xe_vm_rebind+0x4ce/0x5a0 [xe] [ 191.739098] ? trace_hardirqs_on+0x4d/0x60 [ 191.739112] preempt_rebind_work_func+0x76f/0xd00 [xe] Followed by NPD, when running some workload, since the sg was never actually populated but the vma is still marked for rebind when it should be skipped for this special EFAULT case. This is confirmed to fix the user report. v2 (MattB): - Move earlier. v3 (MattB): - Update the commit message to make it clear that this indeed fixes the issue. Fixes: 521db22a1d70 ("drm/xe: Invalidate userptr VMA on page pin fault") Signed-off-by: Matthew Auld Cc: Matthew Brost Cc: Thomas Hellström Cc: # v6.10+ Reviewed-by: Matthew Brost Reviewed-by: Thomas Hellström Link: https://patchwork.freedesktop.org/patch/msgid/20250221143840.167150-5-matthew.auld@intel.com Signed-off-by: Lucas De Marchi (cherry picked from commit 6b93cb98910c826c2e2004942f8b060311e43618) Signed-off-by: Rodrigo Vivi Signed-off-by: Greg Kroah-Hartman

drm/xe/userptr: restore invalidation list on error

2025-03-07T17:27:06+00:00

commit e043dc16c28c8446e66c55adfe7c6e862a6a7bb7 upstream. On error restore anything still on the pin_list back to the invalidation list on error. For the actual pin, so long as the vma is tracked on either list it should get picked up on the next pin, however it looks possible for the vma to get nuked but still be present on this per vm pin_list leading to corruption. An alternative might be then to instead just remove the link when destroying the vma. v2: - Also add some asserts. - Keep the overzealous locking so that we are consistent with the docs; updating the docs and related bits will be done as a follow up. Fixes: ed2bdf3b264d ("drm/xe/vm: Subclass userptr vmas") Suggested-by: Matthew Brost Signed-off-by: Matthew Auld Cc: Thomas Hellström Cc: # v6.8+ Reviewed-by: Matthew Brost Link: https://patchwork.freedesktop.org/patch/msgid/20250221143840.167150-4-matthew.auld@intel.com Signed-off-by: Lucas De Marchi (cherry picked from commit 4e37e928928b730de9aa9a2f5dc853feeebc1742) Signed-off-by: Rodrigo Vivi Signed-off-by: Greg Kroah-Hartman