{"api_version":"1","generated_at":"2026-11-20T07:29:57+00:00","cve":"CVE-2026-99705","urls":{"html":"https://cve.report/CVE-2026-99705","api":"https://cve.report/api/cve/CVE-2026-99705.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-99705","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-99705"},"summary":{"title":"mm: memory: fix truncation removing mapped folios","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm: memory: fix truncation removing mapped folios\n\nAyush reported an issue that shmem/tmpfs folios are still mapped when\ntruncating them on production:\n\"\nBUG: Bad page cache in process ...  pfn:...\npage dumped because: still mapped when deleted\n\"\n\nI reproduced the issue with Ayush's reproducer [1] with the latest\nmm-unstable branch, and I got the following crash info.  From the dump\nmessage, we can see that truncate_inode_folio() is really trying to remove\nmapped folios, which is incorrect.\n\n\"\n[  189.171977] page: refcount:3 mapcount:1 mapping:00000000fe168ee1\nindex:0x3880 pfn:0x191647\n[  189.171995] memcg:ffff0000cc073d40\n[  189.171997] aops:shmem_aops ino:3c01 dentry name(?):\"memfd:runsc-memory\"\n[  189.172006] flags: 0x17fffef0002022d(locked|referenced|uptodate|lru|\nworkingset|swapbacked|node=0|zone=2|lastcpupid=0x3ffff)\n.......\n[  189.172017] page dumped because: VM_BUG_ON_FOLIO(folio_mapped(folio))\n[  189.172026] ------------[ cut here ]------------\n[  189.172027] kernel BUG at mm/filemap.c:155!\n[  189.172057] Internal error: Oops - BUG: 00000000f2000800 [#1]  SMP\n.......\n[  189.177721] CPU: 3 UID: 0 PID: 5521 Comm: repro Kdump: loaded Tainted:\n7.3.0-rc4+ #277 PREEMPT(lazy)\n.......\n[  189.183986] Call trace:\n[  189.184124]  filemap_unaccount_folio+0xf0/0x1e8 (P)\n[  189.184391]  __filemap_remove_folio+0x34/0x160\n[  189.184633]  filemap_remove_folio+0x4c/0xb0\n[  189.184859]  truncate_inode_folio+0x34/0x58\n[  189.185087]  shmem_undo_range+0x220/0x658\n[  189.185308]  shmem_fallocate+0x2f0/0x470\n[  189.185527]  vfs_fallocate+0x128/0x328\n......\n[  189.187078]  el0t_64_sync+0x184/0x188\n\"\n\nAfter analysis, I believe the race exists between truncation and\nMADV_DONTNEED, and shmem's ->map_pages() merely makes the issue easier to\nreproduce.  Since MADV_DONTNEED synchronously releases the pagetable page\nbefore calling tlb_flush_rmaps(), this could cause another thread's\ntruncation to skip zap_pte_range() (due to pmd is none) but still observe\nthe folio's mapcount as non-zero.  A possible race scenario is as follows:\n\nCPU 0                           CPU 1\nshmem_fallocate                 madvise_dontneed_single_vma\n  unmap_mapping_range            .......\n  ......\n  (filemap_map_pages() remap folios)\n\t\t\t\t  zap_pte_range\n  truncate_inode_folio\t\t    zap_empty_pte_table(pmd clear)\n    unmap_mapping_folio\n    ......\n      zap_pmd_range (saw pmd none, skip)\n\n  filemap_remove_folio\n    BUG_ON(folio_mapped)\n                                    tlb_flush_rmaps(remove rmaps)\n\nTo fix this issue, we should move the pmd clear operation (via\nzap_empty_pte_table()) to after tlb_flush_rmaps() and use\nsync_with_folio_pmd_zap() to sync with concurrent zapping, so that when we\nfind a pmd_none() while unmapping a folio without holding the PTL, the\nfolio's mappings are guaranteed to have been removed via\ntlb_flush_rmaps().\n\nMeanwhile move the second pmd_none() check to where it belongs: if\nsplitting zapped the PMD.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-11-20 06:18:06","updated_at":"2026-11-20 06:18:06"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/8011e987026b65f59ea6b6082c777807a008b227","name":"https://git.kernel.org/stable/c/8011e987026b65f59ea6b6082c777807a008b227","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/b57cfd4da476a4636bf1ed69dee176cc266e5a06","name":"https://git.kernel.org/stable/c/b57cfd4da476a4636bf1ed69dee176cc266e5a06","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/243cf1f67b365668389832463fc31cbb9fec47af","name":"https://git.kernel.org/stable/c/243cf1f67b365668389832463fc31cbb9fec47af","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-99705","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-99705","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6375e95f381e3dc85065b6f74263a61522736203 8011e987026b65f59ea6b6082c777807a008b227 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6375e95f381e3dc85065b6f74263a61522736203 b57cfd4da476a4636bf1ed69dee176cc266e5a06 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6375e95f381e3dc85065b6f74263a61522736203 243cf1f67b365668389832463fc31cbb9fec47af git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 6.14","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.14 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.18.64 6.18.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.3.5 7.3.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.4-rc2 * original_commit_for_fix","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["mm/internal.h","mm/memory.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"8011e987026b65f59ea6b6082c777807a008b227","status":"affected","version":"6375e95f381e3dc85065b6f74263a61522736203","versionType":"git"},{"lessThan":"b57cfd4da476a4636bf1ed69dee176cc266e5a06","status":"affected","version":"6375e95f381e3dc85065b6f74263a61522736203","versionType":"git"},{"lessThan":"243cf1f67b365668389832463fc31cbb9fec47af","status":"affected","version":"6375e95f381e3dc85065b6f74263a61522736203","versionType":"git"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["mm/internal.h","mm/memory.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"6.14"},{"lessThan":"6.14","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"6.18.*","status":"unaffected","version":"6.18.64","versionType":"semver"},{"lessThanOrEqual":"7.3.*","status":"unaffected","version":"7.3.5","versionType":"semver"},{"lessThanOrEqual":"*","status":"unaffected","version":"7.4-rc2","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.18.64","versionStartIncluding":"6.14","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3.5","versionStartIncluding":"6.14","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.4-rc2","versionStartIncluding":"6.14","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm: memory: fix truncation removing mapped folios\n\nAyush reported an issue that shmem/tmpfs folios are still mapped when\ntruncating them on production:\n\"\nBUG: Bad page cache in process ...  pfn:...\npage dumped because: still mapped when deleted\n\"\n\nI reproduced the issue with Ayush's reproducer [1] with the latest\nmm-unstable branch, and I got the following crash info.  From the dump\nmessage, we can see that truncate_inode_folio() is really trying to remove\nmapped folios, which is incorrect.\n\n\"\n[  189.171977] page: refcount:3 mapcount:1 mapping:00000000fe168ee1\nindex:0x3880 pfn:0x191647\n[  189.171995] memcg:ffff0000cc073d40\n[  189.171997] aops:shmem_aops ino:3c01 dentry name(?):\"memfd:runsc-memory\"\n[  189.172006] flags: 0x17fffef0002022d(locked|referenced|uptodate|lru|\nworkingset|swapbacked|node=0|zone=2|lastcpupid=0x3ffff)\n.......\n[  189.172017] page dumped because: VM_BUG_ON_FOLIO(folio_mapped(folio))\n[  189.172026] ------------[ cut here ]------------\n[  189.172027] kernel BUG at mm/filemap.c:155!\n[  189.172057] Internal error: Oops - BUG: 00000000f2000800 [#1]  SMP\n.......\n[  189.177721] CPU: 3 UID: 0 PID: 5521 Comm: repro Kdump: loaded Tainted:\n7.3.0-rc4+ #277 PREEMPT(lazy)\n.......\n[  189.183986] Call trace:\n[  189.184124]  filemap_unaccount_folio+0xf0/0x1e8 (P)\n[  189.184391]  __filemap_remove_folio+0x34/0x160\n[  189.184633]  filemap_remove_folio+0x4c/0xb0\n[  189.184859]  truncate_inode_folio+0x34/0x58\n[  189.185087]  shmem_undo_range+0x220/0x658\n[  189.185308]  shmem_fallocate+0x2f0/0x470\n[  189.185527]  vfs_fallocate+0x128/0x328\n......\n[  189.187078]  el0t_64_sync+0x184/0x188\n\"\n\nAfter analysis, I believe the race exists between truncation and\nMADV_DONTNEED, and shmem's ->map_pages() merely makes the issue easier to\nreproduce.  Since MADV_DONTNEED synchronously releases the pagetable page\nbefore calling tlb_flush_rmaps(), this could cause another thread's\ntruncation to skip zap_pte_range() (due to pmd is none) but still observe\nthe folio's mapcount as non-zero.  A possible race scenario is as follows:\n\nCPU 0                           CPU 1\nshmem_fallocate                 madvise_dontneed_single_vma\n  unmap_mapping_range            .......\n  ......\n  (filemap_map_pages() remap folios)\n\t\t\t\t  zap_pte_range\n  truncate_inode_folio\t\t    zap_empty_pte_table(pmd clear)\n    unmap_mapping_folio\n    ......\n      zap_pmd_range (saw pmd none, skip)\n\n  filemap_remove_folio\n    BUG_ON(folio_mapped)\n                                    tlb_flush_rmaps(remove rmaps)\n\nTo fix this issue, we should move the pmd clear operation (via\nzap_empty_pte_table()) to after tlb_flush_rmaps() and use\nsync_with_folio_pmd_zap() to sync with concurrent zapping, so that when we\nfind a pmd_none() while unmapping a folio without holding the PTL, the\nfolio's mappings are guaranteed to have been removed via\ntlb_flush_rmaps().\n\nMeanwhile move the second pmd_none() check to where it belongs: if\nsplitting zapped the PMD."}],"providerMetadata":{"dateUpdated":"2026-11-20T05:50:46.263Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/8011e987026b65f59ea6b6082c777807a008b227"},{"url":"https://git.kernel.org/stable/c/b57cfd4da476a4636bf1ed69dee176cc266e5a06"},{"url":"https://git.kernel.org/stable/c/243cf1f67b365668389832463fc31cbb9fec47af"}],"title":"mm: memory: fix truncation removing mapped folios","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-99705","datePublished":"2026-11-20T05:50:46.263Z","dateReserved":"2026-09-25T10:25:14.430Z","dateUpdated":"2026-11-20T05:50:46.263Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-11-20 06:18:06","lastModifiedDate":"2026-11-20 06:18:06","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"99705","Ordinal":"1","Title":"mm: memory: fix truncation removing mapped folios","CVE":"CVE-2026-99705","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"99705","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm: memory: fix truncation removing mapped folios\n\nAyush reported an issue that shmem/tmpfs folios are still mapped when\ntruncating them on production:\n\"\nBUG: Bad page cache in process ...  pfn:...\npage dumped because: still mapped when deleted\n\"\n\nI reproduced the issue with Ayush's reproducer [1] with the latest\nmm-unstable branch, and I got the following crash info.  From the dump\nmessage, we can see that truncate_inode_folio() is really trying to remove\nmapped folios, which is incorrect.\n\n\"\n[  189.171977] page: refcount:3 mapcount:1 mapping:00000000fe168ee1\nindex:0x3880 pfn:0x191647\n[  189.171995] memcg:ffff0000cc073d40\n[  189.171997] aops:shmem_aops ino:3c01 dentry name(?):\"memfd:runsc-memory\"\n[  189.172006] flags: 0x17fffef0002022d(locked|referenced|uptodate|lru|\nworkingset|swapbacked|node=0|zone=2|lastcpupid=0x3ffff)\n.......\n[  189.172017] page dumped because: VM_BUG_ON_FOLIO(folio_mapped(folio))\n[  189.172026] ------------[ cut here ]------------\n[  189.172027] kernel BUG at mm/filemap.c:155!\n[  189.172057] Internal error: Oops - BUG: 00000000f2000800 [#1]  SMP\n.......\n[  189.177721] CPU: 3 UID: 0 PID: 5521 Comm: repro Kdump: loaded Tainted:\n7.3.0-rc4+ #277 PREEMPT(lazy)\n.......\n[  189.183986] Call trace:\n[  189.184124]  filemap_unaccount_folio+0xf0/0x1e8 (P)\n[  189.184391]  __filemap_remove_folio+0x34/0x160\n[  189.184633]  filemap_remove_folio+0x4c/0xb0\n[  189.184859]  truncate_inode_folio+0x34/0x58\n[  189.185087]  shmem_undo_range+0x220/0x658\n[  189.185308]  shmem_fallocate+0x2f0/0x470\n[  189.185527]  vfs_fallocate+0x128/0x328\n......\n[  189.187078]  el0t_64_sync+0x184/0x188\n\"\n\nAfter analysis, I believe the race exists between truncation and\nMADV_DONTNEED, and shmem's ->map_pages() merely makes the issue easier to\nreproduce.  Since MADV_DONTNEED synchronously releases the pagetable page\nbefore calling tlb_flush_rmaps(), this could cause another thread's\ntruncation to skip zap_pte_range() (due to pmd is none) but still observe\nthe folio's mapcount as non-zero.  A possible race scenario is as follows:\n\nCPU 0                           CPU 1\nshmem_fallocate                 madvise_dontneed_single_vma\n  unmap_mapping_range            .......\n  ......\n  (filemap_map_pages() remap folios)\n\t\t\t\t  zap_pte_range\n  truncate_inode_folio\t\t    zap_empty_pte_table(pmd clear)\n    unmap_mapping_folio\n    ......\n      zap_pmd_range (saw pmd none, skip)\n\n  filemap_remove_folio\n    BUG_ON(folio_mapped)\n                                    tlb_flush_rmaps(remove rmaps)\n\nTo fix this issue, we should move the pmd clear operation (via\nzap_empty_pte_table()) to after tlb_flush_rmaps() and use\nsync_with_folio_pmd_zap() to sync with concurrent zapping, so that when we\nfind a pmd_none() while unmapping a folio without holding the PTL, the\nfolio's mappings are guaranteed to have been removed via\ntlb_flush_rmaps().\n\nMeanwhile move the second pmd_none() check to where it belongs: if\nsplitting zapped the PMD.","Type":"Description","Title":"mm: memory: fix truncation removing mapped folios"}]}}}