linux.git/drivers/gpu/drm/xe/xe_pt.c, branch master

drm/xe/pt: Reset current_op in xe_pt_update_ops_init()

2026-07-16T14:32:56Z

xe_pt_update_ops_init() fails to reset current_op to 0. On the vm_bind path, ops_execute() calls xe_pt_update_ops_prepare() inside the xe_validation_guard() / drm_exec_until_all_locked() loop. When that loop retries due to lock contention or OOM eviction (drm_exec_retry_on_contention() / xe_validation_retry_on_oom()), xe_pt_update_ops_prepare() runs again on the same vops, and each call to bind_op_prepare() increments current_op without resetting it. After N retries current_op exceeds the array size allocated by xe_vma_ops_alloc(), causing an out-of-bounds write into SLUB-poisoned memory and a subsequent UAF crash in xe_migrate_update_pgtables_cpu() when reading the corrupted pt_op->bind. Also reset needs_svm_lock and needs_invalidation which are derived in the same prepare pass and would otherwise cause wrong migrate ops selection and redundant TLB invalidation on retry. Fix this by resetting current_op, needs_svm_lock and needs_invalidation in xe_pt_update_ops_init(). v2 (Matt): - Add details in commit message. - Add Fixes tag and Cc to stable@vger.kernel.org Fixes: e8babb280b5e ("drm/xe: Convert multiple bind ops into single job") Suggested-by: Matthew Auld Cc: stable@vger.kernel.org Assisted-by: GitHub-Copilot:claude-sonnet-4.6 Signed-off-by: Zongyao Bai Reviewed-by: Matthew Brost Signed-off-by: Matthew Brost Link: https://patch.msgid.link/20260714232433.2737533-1-zongyao.bai@intel.com (cherry picked from commit 046045543e530605c441063535e7dca0075369a6) Signed-off-by: Thomas Hellström

drm/xe/userptr: Stub notifier_lock helpers when DRM_GPUSVM=n

2026-07-09T10:09:42Z

When CONFIG_DRM_GPUSVM=n (e.g. um-allyesconfig), the only caller of xe_pt_svm_userptr_notifier_lock() is compiled out, triggering: drivers/gpu/drm/xe/xe_pt.c:1418:13: warning: 'xe_pt_svm_userptr_notifier_lock' defined but not used [-Wunused-function] The helpers cannot simply be removed in this case: the matching xe_pt_svm_userptr_notifier_unlock() is also referenced from xe_pt_update_ops_run(), which lives outside any DRM_GPUSVM ifdef and is gated only at runtime by pt_update_ops->needs_svm_lock. The symbol must exist in all builds. Provide empty static inline stubs for !DRM_GPUSVM, matching the pattern used by xe_svm_notifier_lock()/_unlock() in xe_svm.h. Fixes: dca6e08c923a ("drm/xe/userptr: Hold notifier_lock for write on inject test path") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202606302210.QqcLbOEN-lkp@intel.com/ Reviewed-by: Matthew Brost Link: https://patch.msgid.link/20260630192221.2998168-1-shuicheng.lin@intel.com Signed-off-by: Shuicheng Lin (cherry picked from commit 3359422bf0a1140e96d783a19a397686e580a3ca) Signed-off-by: Thomas Hellström

drm/xe: Fix PTE index in xe_vm_populate_pgtable() for chunked binds

2026-07-09T10:09:18Z

xe_vm_populate_pgtable() indexed the source PTE array (update->pt_entries) by the per-call loop counter, assuming each call starts at the first entry of the update. That holds for the CPU bind path (xe_migrate_update_pgtables_cpu), which populates a whole update in a single call, but not for the GPU bind path: write_pgtable() splits an update into MAX_PTE_PER_SDI (510) sized MI_STORE_DATA_IMM chunks, invoking the populate callback once per chunk with an advancing qword_ofs but a fresh command- buffer destination pointer. As a result, every chunk after the first re-read pt_entries from index 0 instead of from its true offset, so PTEs beyond the first 510 entries of a single update were programmed with the wrong physical pages, shifting the mapping by exactly MAX_PTE_PER_SDI pages. This stayed latent because a single update only exceeds 510 qwords when a large (e.g. 2M) region is bound as individual 4K PTEs rather than a single huge-page entry, which happens when the backing store is sufficiently fragmented. It was surfaced by the BO defrag path, which deliberately rebinds such fragmented ranges via the GPU bind path, producing deterministic data corruption offset by 510 pages. Index pt_entries by the chunk's absolute offset relative to update->ofs so both the CPU and GPU paths pick the correct entries. Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Cc: stable@vger.kernel.org Assisted-by: GitHub_Copilot:claude-opus-4.8 Signed-off-by: Matthew Brost Reviewed-by: Matthew Auld Link: https://patch.msgid.link/20260702012434.3861171-1-matthew.brost@intel.com (cherry picked from commit e6f2d0b757c4fb577a513c577140109d1d292a9a) Signed-off-by: Thomas Hellström

drm/xe/pt: prevent invalid cursor access for purged BOs

2026-07-02T10:29:43Z

During a page table walk for binding, xe_pt_stage_bind() explicitly skips initializing the xe_res_cursor for purged BOs, treating them similarly to NULL VMAs by only setting the cursor size. However, xe_pt_hugepte_possible() and xe_pt_scan_64K() did not check if the BO was purged before attempting to walk the cursor using xe_res_dma() and xe_res_next(). Because the cursor was left uninitialized for purged BOs, this falls through and triggers warnings like: WARNING: drivers/gpu/drm/xe/xe_res_cursor.h:274 at xe_res_next Fix this by explicitly checking if the BO is purged in both xe_pt_hugepte_possible() and xe_pt_scan_64K(), returning early just as we do for NULL VMAs, avoiding the invalid cursor accesses entirely. As a precaution, also zero-initialize the cursor in xe_pt_stage_bind() to ensure we don't pass garbage data into the page table walkers if we ever hit a similar edge case in the future. Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/work_items/8418 Fixes: ad9843aac91a ("drm/xe/madvise: Implement purgeable buffer object support") Assisted-by: Copilot:gemini-3.1-pro-preview Reported-by: Matthew Schwartz Signed-off-by: Matthew Auld Cc: Thomas Hellström Cc: Matthew Brost Cc: Arvind Yadav Reviewed-by: Matthew Brost Tested-by: Matthew Schwartz Link: https://patch.msgid.link/20260625152054.450125-8-matthew.auld@intel.com (cherry picked from commit 4c7b9c6ece32440e5a435a92076d049450cd2d2e) Signed-off-by: Thomas Hellström

drm/xe/userptr: Hold notifier_lock for write on inject test path

2026-07-02T10:29:43Z

When CONFIG_DRM_XE_USERPTR_INVAL_INJECT=y, xe_pt_svm_userptr_pre_commit() runs vma_check_userptr() with the svm notifier_lock taken for read. The test injection causes vma_check_userptr() to call xe_vma_userptr_force_invalidate(), which feeds into xe_vma_userptr_do_inval() with drm_gpusvm_ctx.in_notifier=true. That flag tells drm_gpusvm_unmap_pages() the caller already holds notifier_lock for write and only asserts the mode. Because the caller actually holds it for read, the assertion fires: WARNING: drivers/gpu/drm/drm_gpusvm.c:1669 at \ drm_gpusvm_unmap_pages+0xd4/0x130 [drm_gpusvm_helper] Call Trace: xe_vma_userptr_do_inval+0x40d/0xfd0 [xe] xe_vma_userptr_invalidate_pass1+0x3e6/0x8d0 [xe] xe_vma_userptr_force_invalidate+0xde/0x290 [xe] vma_check_userptr.constprop.0+0x1c6/0x220 [xe] xe_pt_svm_userptr_pre_commit+0x6a3/0xc60 [xe] ... xe_vm_bind_ioctl+0x3a0a/0x4480 [xe] Acquire notifier_lock for write in pre-commit when the inject Kconfig is enabled, via new helpers xe_pt_svm_userptr_notifier_lock()/_unlock(). Rename xe_svm_assert_held_read() to xe_svm_assert_held_read_or_inject_write() so it asserts the correct mode under each build configuration. Production builds (CONFIG_DRM_XE_USERPTR_INVAL_INJECT=n) keep the existing read-mode behavior bit-for-bit. Fixes: 9e9787414882 ("drm/xe/userptr: replace xe_hmm with gpusvm") Assisted-by: Claude:claude-opus-4.7 Cc: Matthew Auld Cc: Zongyao Bai Reviewed-by: Matthew Brost Link: https://patch.msgid.link/20260625215615.3016892-1-shuicheng.lin@intel.com Signed-off-by: Shuicheng Lin (cherry picked from commit 80ccbd97ffee8ad2e73167d826fe7be548364365) Signed-off-by: Thomas Hellström

drm/xe/pt: Fix NULL pointer dereference in xe_pt_zap_ptes_entry()

2026-07-01T16:01:10Z

The page-table walk framework may pass a NULL *child pointer for unpopulated entries. xe_pt_zap_ptes_entry() called container_of(*child) before checking for NULL, then dereferenced the result, causing a crash. Move the container_of() call after a NULL guard, so the function returns early instead of proceeding with an invalid pointer. XE_WARN_ON is kept to help root cause the issue, but we now bail instead of crashing the driver. v2: Comment that triggering XE_WARN_ON is unexpected behavior (Matt Brost) Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Cc: Matthew Brost Cc: Thomas Hellström Reviewed-by: Matthew Brost Link: https://lore.kernel.org/r/20260616081756.286918-1-francois.dugast@intel.com Signed-off-by: Francois Dugast (cherry picked from commit b9297d19d9df5d4b6c994648570c5dcd1cac68ff) Signed-off-by: Thomas Hellström

drm/xe: Add compact-PT and addr mask handling for page reclaim

2026-06-16T17:18:52Z

Current implementation of generate_reclaim_entry() overlooks some differences between the different page implementations: address masking and compact 64K page handling. Address masking of each leaf varies depending on the leaf entry size. generate_reclaim_entry() is using XE_PTE_ADDR_MASK [51:12] for all leaf entries. For 2MB PTEs, bit 12 (PAT) is part of the flags so the old mask corrupts the physical address extraction. 64K pages can be represented as PS64 and a compact PT, which the latter was not handled. Compact pages aren't walked by the unbind walker, so we separately walk through the compact PT to ensure none of the leaf 64K PTEs are dropped. Previously, compact PT were causing an abort since it was considered covered and not descended into. v2: - Update 64K entry/unbind walker for 64K compact PT handling. (Matthew) - Rework calculations of reclamation and address mask size. - Add new func abstracting the error handling before generating the reclaim entry. v3: - Report finer addr granularity in abort debug print for compact. (Zongyao) - Add comments for ADDR_MASK usage. (Zongyao) - Drop existing phys_addr asserts, the new XE_PAGE_ADDR_MASK clears bits checked, so redundant asserts. (Sashiko) - WARN_ON to verify compact pt and edge pt won't be possible. Fixes: b912138df299 ("drm/xe: Create page reclaim list on unbind") Assisted-by: Sashiko-Review:gemini-3.1-pro-preview Cc: stable@vger.kernel.org Cc: Matthew Auld Suggested-by: Zongyao Bai Signed-off-by: Brian Nguyen Reviewed-by: Matthew Auld Reviewed-by: Zongyao Bai Link: https://patch.msgid.link/20260605224257.2194194-2-brian3.nguyen@intel.com Signed-off-by: Matt Roper (cherry picked from commit 669252801a4aa4098fbc5dd9dd0bd93f0625abd7) Signed-off-by: Matthew Brost

drm/xe/pat: Introduce xe_cache_pat_idx() macro helper

2026-04-21T17:20:17Z

Wrap pat.idx[] reads with xe_cache_pat_idx() so invalid PAT index use is caught by xe_assert() in debug builds. Suggested-by: Matthew Auld Reviewed-by: Matthew Auld Signed-off-by: Xin Wang Link: https://patch.msgid.link/20260416045526.536497-4-x.wang@intel.com Signed-off-by: Matt Roper

drm/xe/madvise: Implement purgeable buffer object support

2026-03-27T02:57:10Z

This allows userspace applications to provide memory usage hints to the kernel for better memory management under pressure: Add the core implementation for purgeable buffer objects, enabling memory reclamation of user-designated DONTNEED buffers during eviction. This patch implements the purge operation and state machine transitions: Purgeable States (from xe_madv_purgeable_state): - WILLNEED (0): BO should be retained, actively used - DONTNEED (1): BO eligible for purging, not currently needed - PURGED (2): BO backing store reclaimed, permanently invalid Design Rationale: - Async TLB invalidation via trigger_rebind (no blocking xe_vm_invalidate_vma) - i915 compatibility: retained field, "once purged always purged" semantics - Shared BO protection prevents multi-process memory corruption - Scratch PTE reuse avoids new infrastructure, safe for fault mode Note: The madvise_purgeable() function is implemented but not hooked into the IOCTL handler (madvise_funcs[] entry is NULL) to maintain bisectability. The feature will be enabled in the final patch when all supporting infrastructure (shrinker, per-VMA tracking) is complete. Cc: Matthew Brost Cc: Himal Prasad Ghimiray Reviewed-by: Thomas Hellström Signed-off-by: Arvind Yadav Signed-off-by: Matthew Brost Link: https://patch.msgid.link/20260326130843.3545241-4-arvind.yadav@intel.com

drm/xe: always keep track of remap prev/next

2026-03-25T11:25:18Z

During 3D workload, user is reporting hitting: [ 413.361679] WARNING: drivers/gpu/drm/xe/xe_vm.c:1217 at vm_bind_ioctl_ops_unwind+0x1e2/0x2e0 [xe], CPU#7: vkd3d_queue/9925 [ 413.361944] CPU: 7 UID: 1000 PID: 9925 Comm: vkd3d_queue Kdump: loaded Not tainted 7.0.0-070000rc3-generic #202603090038 PREEMPT(lazy) [ 413.361949] RIP: 0010:vm_bind_ioctl_ops_unwind+0x1e2/0x2e0 [xe] [ 413.362074] RSP: 0018:ffffd4c25c3df930 EFLAGS: 00010282 [ 413.362077] RAX: 0000000000000000 RBX: ffff8f3ee817ed10 RCX: 0000000000000000 [ 413.362078] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [ 413.362079] RBP: ffffd4c25c3df980 R08: 0000000000000000 R09: 0000000000000000 [ 413.362081] R10: 0000000000000000 R11: 0000000000000000 R12: ffff8f41fbf99380 [ 413.362082] R13: ffff8f3ee817e968 R14: 00000000ffffffef R15: ffff8f43d00bd380 [ 413.362083] FS: 00000001040ff6c0(0000) GS:ffff8f4696d89000(0000) knlGS:00000000330b0000 [ 413.362085] CS: 0010 DS: 002b ES: 002b CR0: 0000000080050033 [ 413.362086] CR2: 00007ddfc4747000 CR3: 00000002e6262005 CR4: 0000000000f72ef0 [ 413.362088] PKRU: 55555554 [ 413.362089] Call Trace: [ 413.362092] [ 413.362096] xe_vm_bind_ioctl+0xa9a/0xc60 [xe] Which seems to hint that the vma we are re-inserting for the ops unwind is either invalid or overlapping with something already inserted in the vm. It shouldn't be invalid since this is a re-insertion, so must have worked before. Leaving the likely culprit as something already placed where we want to insert the vma. Following from that, for the case where we do something like a rebind in the middle of a vma, and one or both mapped ends are already compatible, we skip doing the rebind of those vma and set next/prev to NULL. As well as then adjust the original unmap va range, to avoid unmapping the ends. However, if we trigger the unwind path, we end up with three va, with the two ends never being removed and the original va range in the middle still being the shrunken size. If this occurs, one failure mode is when another unwind op needs to interact with that range, which can happen with a vector of binds. For example, if we need to re-insert something in place of the original va. In this case the va is still the shrunken version, so when removing it and then doing a re-insert it can overlap with the ends, which were never removed, triggering a warning like above, plus leaving the vm in a bad state. With that, we need two things here: 1) Stop nuking the prev/next tracking for the skip cases. Instead relying on checking for skip prev/next, where needed. That way on the unwind path, we now correctly remove both ends. 2) Undo the unmap va shrinkage, on the unwind path. With the two ends now removed the unmap va should expand back to the original size again, before re-insertion. v2: - Update the explanation in the commit message, based on an actual IGT of triggering this issue, rather than conjecture. - Also undo the unmap shrinkage, for the skip case. With the two ends now removed, the original unmap va range should expand back to the original range. v3: - Track the old start/range separately. vma_size/start() uses the va info directly. Link: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/7602 Fixes: 8f33b4f054fc ("drm/xe: Avoid doing rebinds") Signed-off-by: Matthew Auld Cc: Matthew Brost Cc: # v6.8+ Reviewed-by: Matthew Brost Link: https://patch.msgid.link/20260318100208.78097-2-matthew.auld@intel.com