{"id":"CVE-2026-97948","published":"2026-09-25T11:17:22.317","lastModified":"2026-09-25T11:17:22.317","description":"In the Linux kernel, the following vulnerability has been resolved:\n\npowerpc/eeh: Fix recursive locking on devices without EEH sensitive driver\n\nThe commit 1010b4c012b0 (\"powerpc/eeh: Make EEH driver device hotplug\nsafe\") refactored the EEH code such that the pci_rescan_remove_lock is\nheld at the beginning of eeh_handle_normal_event() and the\neeh_reset_device() is called with that lock being held. Looks like the\ncommit missed to remove the existing lock/unlock inside eeh_rmv_device()\nwhich is no longer necessary. This is causing the eehd to hang on the\nlock which it actually holds when that code path is taken.\n\n[<0>] 0xc00000011c78f870\n[<0>] __switch_to+0xfc/0x1a0\n[<0>] pci_lock_rescan_remove+0x30/0x44\n[<0>] eeh_rmv_device+0x290/0x2e0\n[<0>] eeh_pe_dev_traverse+0x80/0x130\n[<0>] eeh_reset_device+0xcc/0x23c\n[<0>] eeh_handle_normal_event+0x830/0xa80\n[<0>] eeh_event_handler+0xf8/0x190\n[<0>] kthread+0x194/0x1b0\n[<0>] start_kernel_thread+0x14/0x18\n\nThe issue is seen for cases where the errors are detected on the PHB\ndirectly AND|OR for devices where the driver error_detected() returns\nPCI_ERS_RESULT_NEED_RESET, and driver being not EEH sensitive(i.e no\nerror handlers like slot_reset(), resume() etc defined).","cvssScore":null,"cvssSeverity":null,"cvssVector":null,"cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/102e3dc5ab5ba052e294819e83384166b242c1ec","tags":[]},{"url":"https://git.kernel.org/stable/c/2920af33d097ca335e492b346af70b72986ad6dc","tags":[]},{"url":"https://git.kernel.org/stable/c/85d8eaefc052cf3e5ae2c7bafeda2db68b8898b4","tags":[]},{"url":"https://git.kernel.org/stable/c/c5e68706527968282e49de205cc2b935823cb88a","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}