{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-74591","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-08-15T05:44:03.918Z","datePublished":"2026-08-22T15:31:42.661Z","dateUpdated":"2026-08-25T05:40:18.794Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-25T05:40:18.794Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm/filemap: __filemap_add_folio() restore index before retrying\n\nIn __filemap_add_folio()'s split-a-conflict loop, xas_set_order() is\napplied repeatedly: each application modifies xas.xa_index, rounding it\ndown according to the split_order attempted at that stage: and if all goes\nas intended, it eventually (or immediately) converges on an\nxas_try_split() to the required folio_order, with xas.xa_index now the\nsame as index: then xas_store() puts the new folio into the xarray there.\n\nBut if a new node was needed, and GFP_NOWAIT allocation did not get one,\nthe lock is dropped, xas_nomem() used to allocate, and sequence retried. \nIf (that part of) the xarray is unchanged when the lock is reacquired, no\nproblem.  But what if the conflict was meanwhile resolved by another\nthread (perhaps even doing the same thing, inserting a folio at that same\nindex)?  Isn't there a danger of now putting our folio into the xarray at\nan intermediate rounded-down index?  With !folio_contains() bug to follow,\nwhen CONFIG_DEBUG_VM=y is checking for that.\n\nFix this with an xas_set_order() to restore the original xas.xa_index at\nthe bottom of the loop, so the retry does a full re-evaluation after\nreacquiring the lock, and cannot reach xas_store() with the wrong index.\n\nProduction was suffering from rare SIGILLs and SIGSEGVs, executable text\nfound a page away from where it belonged, !folio_contains() bug hit when\ndebug enabled: symptoms not seen since this patch went in."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H","baseScore":9.8,"baseSeverity":"CRITICAL"},"scenarios":[{"lang":"en","value":"AV:N - __filemap_add_folio() is reached from every page-cache insert path including nfsd vfs_iocb_iter_read() and ksmbd kernel_read() on internet-facing file servers, as well as local buffered reads, writes, and mmap faults via filemap_get_folio().\nAC:L - An attacker controls both sides of the race by issuing concurrent reads/writes or NFS/SMB requests to the same file offsets while applying memory pressure to force xas_nomem() xa_node allocation retries after lock drops.\nPR:N - Exploitation requires only the ability to trigger concurrent page-cache population on accessible files via network NFS/SMB reads or local syscalls; no real root, CAP_SYS_ADMIN, or init-namespace privileges are needed.\nUI:N - The corrupt insert is triggered entirely by attacker-driven concurrent I/O and memory pressure; no victim click, mount, or other interactive action beyond normal automated file access is required.\nS:U - Corruption is confined to the kernel page cache and process address spaces on the same host; it does not cross a VM, container runtime, or IOMMU security boundary to a separate authority.\nC:H - Storing a folio at a rounded-down xarray index maps the wrong physical page to a file offset, enabling cross-page disclosure of file or kernel data; production saw executable text one page away from its correct location.\nI:H - Wrong-index folio insertion corrupts page-cache mappings and served executable content, providing a memory-corruption primitive that can alter code and data seen by userspace mappings and enable further exploitation.\nA:H - Production systems experienced rare SIGILL and SIGSEGV crashes from corrupted mappings, and CONFIG_DEBUG_VM builds hit !folio_contains() BUG checks, demonstrating reliable kernel and process termination."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["mm/filemap.c"],"versions":[{"version":"200a89c159a7a416115e6e309183c82183bf98aa","lessThan":"4917e3ebcab50f0265e8ca01c8567de4c4a47511","status":"affected","versionType":"git"},{"version":"200a89c159a7a416115e6e309183c82183bf98aa","lessThan":"267ecd2eb7759c26f1a026eb0a5b231071534c9c","status":"affected","versionType":"git"},{"version":"200a89c159a7a416115e6e309183c82183bf98aa","lessThan":"86da3f7e1e609e1e8bfbab198af68467c5a015a5","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["mm/filemap.c"],"versions":[{"version":"6.15","status":"affected"},{"version":"0","lessThan":"6.15","status":"unaffected","versionType":"semver"},{"version":"6.18.45","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.1.9","lessThanOrEqual":"7.1.*","status":"unaffected","versionType":"semver"},{"version":"7.2","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.15","versionEndExcluding":"6.18.45"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.15","versionEndExcluding":"7.1.9"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.15","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/4917e3ebcab50f0265e8ca01c8567de4c4a47511"},{"url":"https://git.kernel.org/stable/c/267ecd2eb7759c26f1a026eb0a5b231071534c9c"},{"url":"https://git.kernel.org/stable/c/86da3f7e1e609e1e8bfbab198af68467c5a015a5"}],"title":"mm/filemap: __filemap_add_folio() restore index before retrying","x_generator":{"engine":"bippy-1.2.0"}}}}