{"id":"CVE-2026-97960","published":"2026-09-25T11:17:23.720","lastModified":"2026-09-25T11:17:23.720","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nperf/x86/intel: Prevent drain_pebs() reentry\n\nThe PEBS buffer is shared by all events on a CPU, so drain_pebs() must\nnot be reentered. If so, one instance may observe stale buffer state and\npotentially access out-of-bound memory.\n\nMost invocations happen in NMI context, which naturally prevents reentry.\nHowever, drain_pebs() is also reachable from process context via\nintel_pmu_drain_pebs_buffer().\n\nIn those paths, the PMU is often already disabled, but not guaranteed.\nFor example, __intel_pmu_pebs_disable() only disables the target counter,\nso other active counters can still raise a PMI and interrupt an in-flight\ndrain_pebs(). Here is an example,\n\n__perf_addr_filters_adjust()\n  perf_event_stop()\n    __perf_event_stop()\n      x86_pmu_stop() (event->pmu->stop)\n        intel_pmu_disable_event()\n          intel_pmu_pebs_disable()\n            __intel_pmu_pebs_disable()\n              intel_pmu_drain_large_pebs()\n                intel_pmu_drain_pebs_buffer()\n\nIntroduce __intel_pmu_quiesce() and __intel_pmu_resume() helpers and\nuse them in intel_pmu_drain_large_pebs() to disable the full PMU\naround the intel_pmu_drain_pebs_buffer() call, preventing reentry.\n\nAlso add a warning in intel_pmu_drain_pebs_buffer() when the full PMU is\nnot disabled.","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/a56c03a397e2cd0c4cf8da96dcd6214f7d0e7d8c","tags":[]},{"url":"https://git.kernel.org/stable/c/a5fe19dd8b3ed5fad6e5e0f0c58c7245043ee4af","tags":[]},{"url":"https://git.kernel.org/stable/c/c55599c0ec2aa020e41a0599c3044c56d8a2e7d9","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}