{"id":"AZL-115684","summary":"CVE-2026-99770 affecting package kernel 6.6.161.1-1","details":"In the Linux kernel, the following vulnerability has been resolved:\n\ndrm/vc4: Publish the output fence outside of job_lock\n\njob_lock is taken by the V3D interrupt handler, so nesting an IRQ-unsafe\nlock under it creates an inversion. drm_syncobj_replace_fence() takes\ndrm_syncobj::lock, which other paths acquire with interrupts enabled. For\nexample, a thread interrupted while holding it blocks the handler on\njob_lock, and the handler can only make progress once the submit path\nreleases job_lock, which it cannot do while waiting for drm_syncobj::lock.\n\nThis is reported by lockdep:\n\n[  188.139076] ========================================================\n[  188.139086] WARNING: possible irq lock inversion dependency detected\n[  188.139096] 7.3.0-rc3-00067-gc4f3feeaf60f #67 Tainted: G         C\n[  188.139108] --------------------------------------------------------\n[  188.139116] swapper/0/0 just changed the state of lock:\n[  188.139125] ffff0000054f8ff8 (&vc4-\u003ejob_lock){-...}-{3:3}, at: vc4_irq+0xa8/0x28c [vc4]\n[  188.139234] but this lock took another, HARDIRQ-unsafe lock in the past:\n[  188.139242]  (&syncobj-\u003elock){+.+.}-{3:3}\n[  188.139256] and interrupts could create inverse lock ordering between them.\n[  188.139267] other info that might help us debug this:\n[  188.139275]  Possible interrupt unsafe locking scenario:\n[  188.139284]        CPU0                    CPU1\n[  188.139291]        ----                    ----\n[  188.139299]   lock(&syncobj-\u003elock);\n[  188.139312]                                local_irq_disable();\n[  188.139320]                                lock(&vc4-\u003ejob_lock);\n[  188.139335]                                lock(&syncobj-\u003elock);\n[  188.139350]   \u003cInterrupt\u003e\n[  188.139358]     lock(&vc4-\u003ejob_lock);\n[  188.139372] *** DEADLOCK ***\n\nHand the fence over after job_lock is dropped. The job can complete and\ndrop exec-\u003efence in that window, so take a reference first and publish\nthrough the local pointer, as exec may already be gone.","modified":"2026-11-21T14:17:24.206817797Z","published":"2026-11-20T06:18:15Z","upstream":["CVE-2026-99770"],"references":[{"type":"WEB","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-99770"}],"affected":[{"package":{"name":"kernel","ecosystem":"Azure Linux:3","purl":"pkg:rpm/azure-linux/kernel"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"last_affected":"6.6.161.1-1"}]}],"database_specific":{"source":"https://github.com/microsoft/AzureLinuxVulnerabilityData/blob/main/osv/AZL-115684.json"}}],"schema_version":"1.9.2"}