{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-64588","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-07-19T15:36:31.798Z","datePublished":"2026-08-06T07:13:46.688Z","dateUpdated":"2026-08-17T04:58:09.400Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-08-17T04:58:09.400Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nfuse-uring: fix data races on ring->ready\n\nOn weakly-ordered architectures, the store to fiq->ops can be\nreordered past the store to ring->ready, allowing a CPU that sees\nring->ready == true via fuse_uring_ready() to dispatch requests\nthrough a stale fiq->ops pointer. Upgrade the store to\nsmp_store_release() and the load in fuse_uring_ready() to\nsmp_load_acquire() so that the preceding WRITE_ONCE(fiq->ops, ...)\nis visible to any CPU that observes ring->ready == true.\n\nAdditionally, fuse_uring_do_register() publishes ring->ready with\nWRITE_ONCE() but the fast-path check reads it with a plain load.\nThis is a marked-vs-unmarked access that KCSAN will flag. Wrap it in\nREAD_ONCE() to mark it without adding unnecessary ordering.\n\nAlso wrap the fc->ring load in fuse_uring_ready() in READ_ONCE() to\nprevent the compiler from reloading it between the NULL check and the\ndereference."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H","baseScore":7.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - The bug is reached locally by mounting FUSE with io-uring and submitting IORING_OP_URING_CMD registration while other local threads generate FUSE requests; no network protocol path invokes fuse_uring_ready() or fuse_uring_do_register().\nAC:L - The attacker controls both sides of the race by registering ring entries on /dev/fuse while concurrently issuing filesystem operations on the same mount, so the publish window on weakly ordered CPUs can be driven and retried without relying on victim state.\nPR:L - FUSE is FS_USERNS_MOUNT, so an ordinary user can create a user-namespace mount, open /dev/fuse, and enable FUSE_OVER_IO_URING without init-namespace root privileges.\nUI:N - The attacker can run both the FUSE daemon that registers io-uring queues and the client threads that generate requests on that mount, so no separate victim action is required.\nS:U - Impact is confined to the host kernel authority as a local FUSE/io-uring memory-safety failure, not a VM, IOMMU, or guest-to-host boundary escape.\nC:H - Missing acquire on ring->ready lets CPUs observe readiness while fiq->ops and related ring publication are stale, and the unfixed fc->ring reload in fuse_uring_ready() is a check-then-use UAF window against ring teardown, enabling disclosure of reclaimed kernel memory.\nI:H - Dispatching through inconsistent ops/ring state and the concurrent ring reload UAF allow writes and list/queue operations on stale or reallocated fuse_ring objects, which is exploitable for heap corruption and control-flow hijacking.\nA:H - The same races can NULL-dereference or use a freed fuse_ring in fuse_uring_task_to_queue()/fuse_uring_ready(), hang requests on the wrong queue path, and trigger WARN/oops or panic under panic_on_warn."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["fs/fuse/dev_uring.c","fs/fuse/dev_uring_i.h"],"versions":[{"version":"c2c9af9a0b13261c36909036057a116f2edb5e1a","lessThan":"b156bb9966972122b148acab8bdf415cdb8176a3","status":"affected","versionType":"git"},{"version":"c2c9af9a0b13261c36909036057a116f2edb5e1a","lessThan":"d01a09b442cb786cd44ccc7c84d57e2856d6737c","status":"affected","versionType":"git"},{"version":"c2c9af9a0b13261c36909036057a116f2edb5e1a","lessThan":"46725a0056c884cf58a6897f222892807327d82d","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["fs/fuse/dev_uring.c","fs/fuse/dev_uring_i.h"],"versions":[{"version":"6.14","status":"affected"},{"version":"0","lessThan":"6.14","status":"unaffected","versionType":"semver"},{"version":"6.18.39","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.1.4","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":"6.14","versionEndExcluding":"6.18.39"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.14","versionEndExcluding":"7.1.4"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.14","versionEndExcluding":"7.2"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/b156bb9966972122b148acab8bdf415cdb8176a3"},{"url":"https://git.kernel.org/stable/c/d01a09b442cb786cd44ccc7c84d57e2856d6737c"},{"url":"https://git.kernel.org/stable/c/46725a0056c884cf58a6897f222892807327d82d"}],"title":"fuse-uring: fix data races on ring->ready","x_generator":{"engine":"bippy-1.2.0"}}}}