{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-74374","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-08-15T05:44:03.889Z","datePublished":"2026-08-15T05:58:53.659Z","dateUpdated":"2026-08-17T05:46:27.554Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-17T05:46:27.554Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nmd/raid1,raid10: fix error-path detection with md_cloned_bio()\n\nDetect the error path using md_cloned_bio() instead of relying\non r1_bio in raid1 or r10_bio->read_slot in raid10, which may be\nNULL or -1 after splitting and resubmitting a failed bio.\n\nAs a result, the error path may not be recognized and memory\nallocations can incorrectly use GFP_NOIO instead of\n(GFP_NOIO | __GFP_HIGH), which can lead to a deadlock under\nmemory pressure."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H","baseScore":7.5,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:N - Linux file/NAS servers commonly export md RAID1/RAID10 volumes via nfsd, ksmbd, or iSCSI; remote client reads reach md_submit_bio()→raid1/raid10_read_request(), and mirror read failures drive handle_read_error() on the blocked raid1d/raid10d thread without local shell access.\nAC:L - An attacker can reliably hit the bug by reading sectors on a degraded mirror that return errors (common on production arrays), issuing large reads that bio_submit_split_bioset() splits on the error-retry path, and concurrently applying memory pressure so GFP_NOIO bio_alloc_clone() blocks reclaim while raid1d/raid10d holds the array frozen.\nPR:N - Triggering the fault path requires only permission to read exported storage or files on an md-backed filesystem; no CAP_SYS_ADMIN, md ioctl, or init-namespace root is needed because read I/O and mirror error recovery run with no extra capability checks beyond normal block-device access.\nUI:N - No victim interaction is required; exploitation uses the attacker's own read I/O (or scripted NFS/SMB/iSCSI reads) together with memory pressure, without anyone mounting disks, acknowledging bad blocks, or performing other cooperative actions.\nS:U - The circular GFP_NOIO allocation wait while raid1d/raid10d is blocked in handle_read_error() freezes I/O within the same host kernel and md security domain; it does not cross VM, container, or IOMMU boundaries.\nC:N - This is a GFP-flag misclassification causing reclaim/I/O circular wait (deadlock), not memory corruption, out-of-bounds access, use-after-free, or other primitives that could disclose kernel or user data.\nI:N - The flaw does not enable arbitrary writes, code execution, or silent data modification; impact is limited to blocking forward progress in the RAID error-recovery path under memory pressure.\nA:H - Without __GFP_HIGH on the error path, bio_alloc_clone() can stall reclaim while raid1d/raid10d blocks with the array frozen, deadlocking all I/O to the md device and potentially hanging root/data volumes on enterprise RAID1/RAID10 servers."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/md/raid1.c","drivers/md/raid10.c"],"versions":[{"version":"689389a06ce79fdced85b5115717f71c71e623e0","lessThan":"20fb582c92fd64e5c8bd83c6176264b2794603b7","status":"affected","versionType":"git"},{"version":"689389a06ce79fdced85b5115717f71c71e623e0","lessThan":"811545e0926d02a6a0b1a1258bb5544777c164d4","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["drivers/md/raid1.c","drivers/md/raid10.c"],"versions":[{"version":"4.12","status":"affected"},{"version":"0","lessThan":"4.12","status":"unaffected","versionType":"semver"},{"version":"7.1.5","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":"4.12","versionEndExcluding":"7.1.5"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"4.12","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/20fb582c92fd64e5c8bd83c6176264b2794603b7"},{"url":"https://git.kernel.org/stable/c/811545e0926d02a6a0b1a1258bb5544777c164d4"}],"title":"md/raid1,raid10: fix error-path detection with md_cloned_bio()","x_generator":{"engine":"bippy-1.2.0"}}}}