2025-06-18 14:04:05 +00:00

49 lines
3.6 KiB
JSON

{
"id": "CVE-2022-50202",
"sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
"published": "2025-06-18T11:15:50.923",
"lastModified": "2025-06-18T13:47:40.833",
"vulnStatus": "Awaiting Analysis",
"cveTags": [],
"descriptions": [
{
"lang": "en",
"value": "In the Linux kernel, the following vulnerability has been resolved:\n\nPM: hibernate: defer device probing when resuming from hibernation\n\nsyzbot is reporting hung task at misc_open() [1], for there is a race\nwindow of AB-BA deadlock which involves probe_count variable. Currently\nwait_for_device_probe() from snapshot_open() from misc_open() can sleep\nforever with misc_mtx held if probe_count cannot become 0.\n\nWhen a device is probed by hub_event() work function, probe_count is\nincremented before the probe function starts, and probe_count is\ndecremented after the probe function completed.\n\nThere are three cases that can prevent probe_count from dropping to 0.\n\n (a) A device being probed stopped responding (i.e. broken/malicious\n hardware).\n\n (b) A process emulating a USB device using /dev/raw-gadget interface\n stopped responding for some reason.\n\n (c) New device probe requests keeps coming in before existing device\n probe requests complete.\n\nThe phenomenon syzbot is reporting is (b). A process which is holding\nsystem_transition_mutex and misc_mtx is waiting for probe_count to become\n0 inside wait_for_device_probe(), but the probe function which is called\n from hub_event() work function is waiting for the processes which are\nblocked at mutex_lock(&misc_mtx) to respond via /dev/raw-gadget interface.\n\nThis patch mitigates (b) by deferring wait_for_device_probe() from\nsnapshot_open() to snapshot_write() and snapshot_ioctl(). Please note that\nthe possibility of (b) remains as long as any thread which is emulating a\nUSB device via /dev/raw-gadget interface can be blocked by uninterruptible\nblocking operations (e.g. mutex_lock()).\n\nPlease also note that (a) and (c) are not addressed. Regarding (c), we\nshould change the code to wait for only one device which contains the\nimage for resuming from hibernation. I don't know how to address (a), for\nuse of timeout for wait_for_device_probe() might result in loss of user\ndata in the image. Maybe we should require the userland to wait for the\nimage device before opening /dev/snapshot interface."
}
],
"metrics": {},
"references": [
{
"url": "https://git.kernel.org/stable/c/003a456ae6f70bb97e436e02fc5105be577c1570",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/2f0e18e0db42f4f8bc87d3d98333680065ceeff8",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/3c48d3067eaf878642276f053575a5c642600a50",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/5a283b59bce72c05c60e9f0fa92a28b5b850d8bb",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/8386c414e27caba8501119948e9551e52b527f59",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/8c90947e5f1801e6c7120021c6ea0f3ad6a4eb91",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/b8e1ae9433d7bd95f2dcc044a7a6f20a4c40d258",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
},
{
"url": "https://git.kernel.org/stable/c/f7042cf9dd40733f387b7cac021e626c74b8856f",
"source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
}
]
}