Am Thu, 19 Apr 2018 11:50:33 +0200
schrieb Kevin Wolf <[email protected]>:

> The real question is why somebody else still holds a lock.

I think the bug is that something locked 0xc9, but did not unlock it:

2018-05-07T08:34:00.466283Z qemu-system-i386: qemu_open: 39 = 
/nfs_vmimages/fv_sles12sp1/vdisk.fv_sles12sp1.disk0.qcow2 0 Success
2018-05-07T08:34:00.466968Z qemu-system-i386: qemu_lock_fcntl: 39 64 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467017Z qemu-system-i386: qemu_lock_fcntl: 39 65 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467048Z qemu-system-i386: qemu_lock_fcntl: 39 66 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467076Z qemu-system-i386: qemu_lock_fcntl: 39 67 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467104Z qemu-system-i386: qemu_lock_fcntl: 39 68 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467369Z qemu-system-i386: qemu_lock_fcntl: 39 c8 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467669Z qemu-system-i386: qemu_lock_fcntl: 39 c9 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467701Z qemu-system-i386: qemu_lock_fcntl: 39 ca 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467729Z qemu-system-i386: qemu_lock_fcntl: 39 cb 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.467757Z qemu-system-i386: qemu_lock_fcntl: 39 cc 1 
F_UNLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.468168Z qemu-system-i386: qemu_lock_fcntl: 39 64 1 
F_RDLCK>F_RDLCK 0 Success
2018-05-07T08:34:00.468905Z qemu-system-i386: qemu_lock_fcntl: 39 65 1 
F_RDLCK>F_RDLCK 0 Success
2018-05-07T08:34:00.469618Z qemu-system-i386: qemu_lock_fcntl: 39 67 1 
F_RDLCK>F_RDLCK 0 Success
2018-05-07T08:34:00.470193Z qemu-system-i386: qemu_lock_fcntl: 39 c9 1 
F_RDLCK>F_RDLCK 0 Success
2018-05-07T08:34:00.470701Z qemu-system-i386: qemu_lock_fcntl: 39 cb 1 
F_RDLCK>F_RDLCK 0 Success
2018-05-07T08:34:00.471183Z qemu-system-i386: qemu_lock_fd_test: 39 c8 1 
F_WRLCK>F_UNLCK 0 Success
2018-05-07T08:34:00.471229Z qemu-system-i386: qemu_lock_fd_test: 39 c9 1 
F_WRLCK>F_RDLCK 0 Success
raw_check_lock_bytes: 123 fcntl on 39 returned -11/0
2018-05-07T08:34:00.721403Z qemu-system-i386: qemu_lock_fd_test: 39 c9 1 
F_WRLCK>F_RDLCK 0 Success
raw_check_lock_bytes: 122 fcntl on 39 returned -11/0
...

It looks like raw_apply_lock_bytes cycles through RAW_LOCK_PERM_BASE and 
RAW_LOCK_SHARED_BASE and does unlock. Then something locks 0xc9 and tries to 
lock it twice.

Olaf

Attachment: pgpMV62yWv57M.pgp
Description: Digitale Signatur von OpenPGP

Reply via email to