fixup: mm/migrate_device: handle partially mapped folios during
Subsystems (2)
Matthew Brost
reviewerPATCH v2 1/2 fixup: mm/migrate_device: handle partially mapped folios during
Splitting a partially mapped folio caused a regression in the Intel Xe SVM test suite in the mremap section, resulting in the following stack trace:
NFO: task kworker/u65:2:1642 blocked for more than 30 seconds.
[ 212.624286] Tainted: G S W 6.18.0-rc6-xe+ #1719 [ 212.638288] Workqueue: xe_page_fault_work_queue xe_pagefault_queue_work [xe] [ 212.638323] Call Trace: [ 212.638324] <TASK> [ 212.638325] __schedule+0x4b0/0x990 [ 212.638330] schedule+0x22/0xd0 [ 212.638331] io_schedule+0x41/0x60 [ 212.638333] migration_entry_wait_on_locked+0x1d8/0x2d0 [ 212.638336] ? __pfx_wake_page_function+0x10/0x10 [ 212.638339] migration_entry_wait+0xd2/0xe0 [ 212.638341] hmm_vma_walk_pmd+0x7c9/0x8d0 [ 212.638343] walk_pgd_range+0x51d/0xa40 [ 212.638345] __walk_page_range+0x75/0x1e0 [ 212.638347] walk_page_range_mm+0x138/0x1f0 [ 212.638349] hmm_range_fault+0x59/0xa0 [ 212.638351] drm_gpusvm_get_pages+0x194/0x7b0 [drm_gpusvm_helper] [ 212.638354] drm_gpusvm_range_get_pages+0x2d/0x40 [drm_gpusvm_helper] [ 212.638355] __xe_svm_handle_pagefault+0x259/0x900 [xe] [ 212.638375] ? update_load_avg+0x7f/0x6c0 [ 212.638377] ? update_curr+0x13d/0x170 [ 212.638379] xe_svm_handle_pagefault+0x37/0x90 [xe] [ 212.638396] xe_pagefault_queue_work+0x2da/0x3c0 [xe] [ 212.638420] process_one_work+0x16e/0x2e0 [ 212.638422] worker_thread+0x284/0x410 [ 212.638423] ? __pfx_worker_thread+0x10/0x10 [ 212.638425] kthread+0xec/0x210 [ 212.638427] ? __pfx_kthread+0x10/0x10 [ 212.638428] ? __pfx_kthread+0x10/0x10 [ 212.638430] ret_from_fork+0xbd/0x100 [ 212.638433] ? __pfx_kthread+0x10/0x10 [ 212.638434] ret_from_fork_asm+0x1a/0x30 [ 212.638436] </TASK>
The issue appears to be that migration PTEs are not properly removed after a split due to incorrect retry handling after a split failure or success. Upon failure, collect a skip, and upon success, continue the collection from the current position in the sequence.
Also, while here, fix migrate_vma_split_folio to only lock the new fault folio if it is different from the original fault folio (i.e., it’s possible the original fault folio is not the same as the one being split).
---
This fixup should be squashed into the patch "mm/migrate_device: handle partially mapped folios during" in mm/mm-unstable
--- mm/migrate_device.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-)
mm/migrate_device.c +11 -6
@@ -110,8 +110,10 @@ static int migrate_vma_split_folio(struct folio *folio,folio_unlock (folio);folio_put (folio); }else if (folio != new_fault_folio) { -folio_get (new_fault_folio); -folio_lock (new_fault_folio); +if (new_fault_folio != fault_folio) { +folio_get (new_fault_folio); +folio_lock (new_fault_folio); + }folio_unlock (folio);folio_put (folio); }
@@ -266,10 +268,11 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,return 0 ; } - ptep =pte_offset_map_lock (mm, pmdp, addr, &ptl); + ptep =pte_offset_map_lock (mm, pmdp, start, &ptl);if (!ptep) goto again;arch_enter_lazy_mmu_mode (); + ptep += (addr - start) /PAGE_SIZE ;for (; addr < end; addr +=PAGE_SIZE , ptep++) {struct dev_pagemap *pgmap;
@@ -351,16 +354,18 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,if (folio &&folio_test_large (folio)) {int ret; +arch_leave_lazy_mmu_mode ();pte_unmap_unlock (ptep, ptl); ret =migrate_vma_split_folio (folio, migrate->fault_page);if (ret) { - ptep =pte_offset_map_lock (mm, pmdp, addr, &ptl); - goto next; +if (unmapped) +flush_tlb_range (walk->vma, start, end); + +return migrate_vma_collect_skip (addr, end, walk); } - addr = start; goto again; } mpfn =migrate_pfn (pfn) |MIGRATE_PFN_MIGRATE ;
Signature · 1 line
2.34.1
Matthew Brost
reviewerPATCH v2 2/2 fixup: mm/migrate_device: add THP splitting during migration
Fix splitting a folio in PMD collection to collect a skip on folio split failure and continue the loop at the current position on success. This fixes an issue where migration entries that had already been collected could accidentally be left behind.
---
This fixup should be squashed into the patch "mm/migrate_device: add THP splitting during migration" in mm/mm-unstable
--- mm/migrate_device.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
mm/migrate_device.c +5 -3
@@ -313,16 +313,18 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,if (folio_test_large (folio)) {int ret; +arch_leave_lazy_mmu_mode ();pte_unmap_unlock (ptep, ptl); ret =migrate_vma_split_folio (folio, migrate->fault_page);if (ret) { - ptep =pte_offset_map_lock (mm, pmdp, addr, &ptl); - goto next; +if (unmapped) +flush_tlb_range (walk->vma, start, end); + +return migrate_vma_collect_skip (addr, end, walk); } - addr = start; goto again; }
Signature · 1 line
2.34.1