"value":"In the Linux kernel, the following vulnerability has been resolved:\n\nio_uring: fix shared sqpoll cancellation hangs\n\n[ 736.982891] INFO: task iou-sqp-4294:4295 blocked for more than 122 seconds.\n[ 736.982897] Call Trace:\n[ 736.982901] schedule+0x68/0xe0\n[ 736.982903] io_uring_cancel_sqpoll+0xdb/0x110\n[ 736.982908] io_sqpoll_cancel_cb+0x24/0x30\n[ 736.982911] io_run_task_work_head+0x28/0x50\n[ 736.982913] io_sq_thread+0x4e3/0x720\n\nWe call io_uring_cancel_sqpoll() one by one for each ctx either in\nsq_thread() itself or via task works, and it's intended to cancel all\nrequests of a specified context. However the function uses per-task\ncounters to track the number of inflight requests, so it counts more\nrequests than available via currect io_uring ctx and goes to sleep for\nthem to appear (e.g. from IRQ), that will never happen.\n\nCancel a bit more than before, i.e. all ctxs that share sqpoll\nand continue to use shared counters. Don't forget that we should not\nremove ctx from the list before running that task_work sqpoll-cancel,\notherwise the function wouldn't be able to find the context and will\nhang."