CVE-2021-47226

{"id": "CVE-2021-47226", "cveTags": [], "metrics": {"cvssMetricV31": [{"type": "Primary", "source": "[email protected]", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 7.1, "attackVector": "LOCAL", "baseSeverity": "HIGH", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H", "integrityImpact": "HIGH", "userInteraction": "NONE", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "LOW", "confidentialityImpact": "NONE"}, "impactScore": 5.2, "exploitabilityScore": 1.8}]}, "published": "2024-05-21T15:15:11.823", "references": [{"url": "https://git.kernel.org/stable/c/002665dcba4bbec8c82f0aeb4bd3f44334ed2c14", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/a7748e021b9fb7739e3cb88449296539de0b6817", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/d8778e393afa421f1f117471144f8ce6deb6953a", "tags": ["Patch"], "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}, {"url": "https://git.kernel.org/stable/c/002665dcba4bbec8c82f0aeb4bd3f44334ed2c14", "tags": ["Patch"], "source": "af854a3a-2127-422b-91ae-364da2661108"}, {"url": "https://git.kernel.org/stable/c/a7748e021b9fb7739e3cb88449296539de0b6817", "tags": ["Patch"], "source": "af854a3a-2127-422b-91ae-364da2661108"}, {"url": "https://git.kernel.org/stable/c/d8778e393afa421f1f117471144f8ce6deb6953a", "tags": ["Patch"], "source": "af854a3a-2127-422b-91ae-364da2661108"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Primary", "source": "[email protected]", "description": [{"lang": "en", "value": "CWE-203"}]}], "descriptions": [{"lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nx86/fpu: Invalidate FPU state after a failed XRSTOR from a user buffer\n\nBoth Intel and AMD consider it to be architecturally valid for XRSTOR to\nfail with #PF but nonetheless change the register state. The actual\nconditions under which this might occur are unclear [1], but it seems\nplausible that this might be triggered if one sibling thread unmaps a page\nand invalidates the shared TLB while another sibling thread is executing\nXRSTOR on the page in question.\n\n__fpu__restore_sig() can execute XRSTOR while the hardware registers\nare preserved on behalf of a different victim task (using the\nfpu_fpregs_owner_ctx mechanism), and, in theory, XRSTOR could fail but\nmodify the registers.\n\nIf this happens, then there is a window in which __fpu__restore_sig()\ncould schedule out and the victim task could schedule back in without\nreloading its own FPU registers. This would result in part of the FPU\nstate that __fpu__restore_sig() was attempting to load leaking into the\nvictim task's user-visible state.\n\nInvalidate preserved FPU registers on XRSTOR failure to prevent this\nsituation from corrupting any state.\n\n[1] Frequent readers of the errata lists might imagine \"complex\n microarchitectural conditions\"."}, {"lang": "es", "value": "En el kernel de Linux, se ha resuelto la siguiente vulnerabilidad: x86/fpu: Invalida el estado de la FPU despu\u00e9s de un XRSTOR fallido desde un b\u00fafer de usuario. Tanto Intel como AMD consideran que es arquitect\u00f3nicamente v\u00e1lido que XRSTOR falle con #PF pero aun as\u00ed cambie el estado del registro. . Las condiciones reales bajo las cuales esto podr\u00eda ocurrir no est\u00e1n claras [1], pero parece plausible que esto pueda desencadenarse si un hilo hermano desasigna una p\u00e1gina e invalida el TLB compartido mientras otro hilo hermano est\u00e1 ejecutando XRSTOR en la p\u00e1gina en cuesti\u00f3n. __fpu__restore_sig() puede ejecutar XRSTOR mientras los registros de hardware se conservan en nombre de una tarea de v\u00edctima diferente (usando el mecanismo fpu_fpregs_owner_ctx) y, en teor\u00eda, XRSTOR podr\u00eda fallar pero modificar los registros. Si esto sucede, entonces hay una ventana en la que __fpu__restore_sig() podr\u00eda programar la salida y la tarea de la v\u00edctima podr\u00eda volver a programarse sin recargar sus propios registros FPU. Esto resultar\u00eda en parte del estado de la FPU en el que __fpu__restore_sig() intentaba cargar una filtraci\u00f3n en el estado visible para el usuario de la tarea de la v\u00edctima. Invalide los registros FPU preservados en caso de falla de XRSTOR para evitar que esta situaci\u00f3n corrompa cualquier estado. [1] Los lectores frecuentes de las listas de erratas podr\u00edan imaginar \"condiciones microarquitect\u00f3nicas complejas\"."}], "lastModified": "2025-04-29T19:26:36.690", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "71A9FF74-A7CD-4A97-B822-6302C3124C48", "versionEndExcluding": "5.10.46", "versionStartIncluding": "5.2"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "7806E7E5-6D4F-4E18-81C1-79B3C60EE855", "versionEndExcluding": "5.12.13", "versionStartIncluding": "5.11"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc1:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "0CBAD0FC-C281-4666-AB2F-F8E6E1165DF7"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc2:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "96AC23B2-D46A-49D9-8203-8E1BEDCA8532"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc3:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "DA610E30-717C-4700-9F77-A3C9244F3BFD"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc4:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "1ECD33F5-85BE-430B-8F86-8D7BD560311D"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc5:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "CF351855-2437-4CF5-AD7C-BDFA51F27683"}, {"criteria": "cpe:2.3:o:linux:linux_kernel:5.13:rc6:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "25A855BA-2118-44F2-90EF-EBBB12AF51EF"}], "operator": "OR"}]}], "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"}