CVE-2025-38338

{"id": "CVE-2025-38338", "cveTags": [], "metrics": {"cvssMetricV31": [{"type": "Primary", "source": "nvd@nist.gov", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 7.8, "attackVector": "LOCAL", "baseSeverity": "HIGH", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H", "integrityImpact": "HIGH", "userInteraction": "NONE", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "LOW", "confidentialityImpact": "HIGH"}, "impactScore": 5.9, "exploitabilityScore": 1.8}]}, "published": "2025-07-10T09:15:28.510", "references": [{"url": "https://git.kernel.org/stable/c/14f5549ad163be2c018abc1bb38370fff617a243", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/1e93b61d3eaa14bfebcc2716ac09d43f3845d420", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/4c10fa44bc5f700e2ea21de2fbae520ba21f19d9", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/5bf0b9eeb0174686f22c2e5b8fb9f47ad25da6f5", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Primary", "source": "nvd@nist.gov", "description": [{"lang": "en", "value": "CWE-415"}]}], "descriptions": [{"lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nfs/nfs/read: fix double-unlock bug in nfs_return_empty_folio()\n\nSometimes, when a file was read while it was being truncated by\nanother NFS client, the kernel could deadlock because folio_unlock()\nwas called twice, and the second call would XOR back the `PG_locked`\nflag.\n\nMost of the time (depending on the timing of the truncation), nobody\nnotices the problem because folio_unlock() gets called three times,\nwhich flips `PG_locked` back off:\n\n 1. vfs_read, nfs_read_folio, ... nfs_read_add_folio,\n nfs_return_empty_folio\n 2. vfs_read, nfs_read_folio, ... netfs_read_collection,\n netfs_unlock_abandoned_read_pages\n 3. vfs_read, ... nfs_do_read_folio, nfs_read_add_folio,\n nfs_return_empty_folio\n\nThe problem is that nfs_read_add_folio() is not supposed to unlock the\nfolio if fscache is enabled, and a nfs_netfs_folio_unlock() check is\nmissing in nfs_return_empty_folio().\n\nRarely this leads to a warning in netfs_read_collection():\n\n ------------[ cut here ]------------\n R=0000031c: folio 10 is not locked\n WARNING: CPU: 0 PID: 29 at fs/netfs/read_collect.c:133 netfs_read_collection+0x7c0/0xf00\n [...]\n Workqueue: events_unbound netfs_read_collection_worker\n RIP: 0010:netfs_read_collection+0x7c0/0xf00\n [...]\n Call Trace:\n <TASK>\n netfs_read_collection_worker+0x67/0x80\n process_one_work+0x12e/0x2c0\n worker_thread+0x295/0x3a0\n\nMost of the time, however, processes just get stuck forever in\nfolio_wait_bit_common(), waiting for `PG_locked` to disappear, which\nnever happens because nobody is really holding the folio lock."}, {"lang": "es", "value": "En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: fs/nfs/read: corrige error de doble desbloqueo en nfs_return_empty_folio() En ocasiones, cuando se le\u00eda un archivo mientras otro cliente NFS lo estaba truncando, el kernel pod\u00eda bloquearse porque se llamaba a folio_unlock() dos veces y la segunda llamada XOR devolver\u00eda el indicador `PG_locked`. La mayor\u00eda de las veces (dependiendo del momento del truncamiento), nadie nota el problema porque folio_unlock() se llama tres veces, lo que desactiva `PG_locked`: 1. vfs_read, nfs_read_folio, ... nfs_read_add_folio, nfs_return_empty_folio 2. vfs_read, nfs_read_folio, ... netfs_read_collection, netfs_unlock_abandoned_read_pages 3. vfs_read, ... nfs_do_read_folio, nfs_read_add_folio, nfs_return_empty_folio El problema es que nfs_read_add_folio() no debe desbloquear el folio si fscache est\u00e1 habilitado, y falta una comprobaci\u00f3n nfs_netfs_folio_unlock() en nfs_return_empty_folio(). En raras ocasiones, esto genera una advertencia en netfs_read_collection(): ------------[ cortar aqu\u00ed ]------------ R=0000031c: el folio 10 no est\u00e1 bloqueado ADVERTENCIA: CPU: 0 PID: 29 en fs/netfs/read_collect.c:133 netfs_read_collection+0x7c0/0xf00 [...] Cola de trabajo: events_unbound netfs_read_collection_worker RIP: 0010:netfs_read_collection+0x7c0/0xf00 [...] Rastreo de llamadas: netfs_read_collection_worker+0x67/0x80 process_one_work+0x12e/0x2c0worker_thread+0x295/0x3a0 Sin embargo, la mayor\u00eda de las veces, los procesos simplemente se quedan atascados para siempre en folio_wait_bit_common(), esperando que `PG_locked` desaparezca, lo que nunca sucede porque nadie est\u00e1 realmente reteniendo el cerradura de folio. "}], "lastModified": "2025-11-18T12:52:58.413", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "F17A70D5-5B72-4411-9C16-98EF57D0C931", "versionEndExcluding": "6.6.95", "versionStartIncluding": "6.4"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "E569FD34-0076-4428-BE17-EECCF867611C", "versionEndExcluding": "6.12.35", "versionStartIncluding": "6.7"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "DFD174C5-1AA2-4671-BDDC-1A9FCC753655", "versionEndExcluding": "6.15.4", "versionStartIncluding": "6.13"}], "operator": "OR"}]}], "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}