Btw, totally unrelated to the report, but related to what the C reproducer does, killing it sometimes results in 12563 pts/1 Zl 0:00 [syzkaller] <defunct> [ 46.018014] mount.nfs (1163) used greatest stack depth: 23944 bytes left [ 9929.865478] syzkaller (12313) used greatest stack depth: 23216 bytes left [10159.658915] INFO: task syzkaller:12312 blocked for more than 120 seconds. [10159.663075] Not tainted 6.13.0-rc1+ #92 [10159.665618] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [10159.673650] task:syzkaller state:D stack:28944 pid:12312 tgid:12307 ppid:1 flags:0x00004006 [10159.681276] Call Trace: [10159.683004] <TASK> [10159.685636] __schedule+0x1b42/0x25b0 [10159.688521] schedule+0xb5/0x260 [10159.690415] __fuse_simple_request+0xc49/0x1350 [fuse] [10159.694677] ? wake_bit_function+0x210/0x210 [10159.697145] fuse_do_getattr+0x2cb/0x600 [fuse] Aborting the connection(s) 'fixes' that, but looks like it triggers another issue. Timeouts would certainly help, but it still should work automatically. Thanks, Bernd
