I hit the following warning after removing a disk
(/sys/class/scsi_device/.../device/delete) while doing IO via sg_dd:
Jun 29 09:37:47 wyvern kernel: ------------[ cut here ]------------
Jun 29 09:37:47 wyvern kernel: WARNING: at block/blk-core.c:1080
__blk_put_request+0xc6/0xe0()
Jun 29 09:37:47 wyvern kernel: Hardware name: ftServer 4500
Jun 29 09:37:47 wyvern kernel: Modules linked in: autofs4 sunrpc
be2iscsi bnx2i cnic uio cxgb3i iw_cxgb3 cxgb3 mdio ib_iser rdma_cm ib_cm
iw_cm ib_sa ib_mad ib_core ib_addr ipv6 iscsi_tcp libiscsi_tcp libiscsi
scsi_transport_iscsi dm_mirror dm_region_hash dm_log dm_multipath dm_mod
kvm_intel kvm uinput sg serio_raw shpchp pcspkr ext3 jbd mbcache raid1
sd_mod crc_t10dif lpfc mptsas scsi_transport_fc mptscsih mptbase
scsi_tgt scsi_transport_sas [last unloaded: scsi_wait_scan]
Jun 29 09:37:47 wyvern kernel: Pid: 7176, comm: sg_dd Not tainted 2.6.34
#2
Jun 29 09:37:47 wyvern kernel: Call Trace:
Jun 29 09:37:47 wyvern kernel: [<ffffffff8105d5fb>]
warn_slowpath_common+0x7b/0xc0
Jun 29 09:37:47 wyvern kernel: [<ffffffff8105d654>]
warn_slowpath_null+0x14/0x20
Jun 29 09:37:47 wyvern kernel: [<ffffffff811fdd86>]
__blk_put_request+0xc6/0xe0
Jun 29 09:37:47 wyvern kernel: [<ffffffff811fe53a>]
blk_put_request+0x3a/0x60
Jun 29 09:37:47 wyvern kernel: [<ffffffffa0160a97>]
sg_finish_rem_req+0x67/0x110 [sg]
Jun 29 09:37:47 wyvern kernel: [<ffffffffa0162e27>]
sg_common_write+0x537/0x580 [sg]
Jun 29 09:37:47 wyvern kernel: [<ffffffff811fdc2c>] ?
__freed_request+0x10c/0x140
Jun 29 09:37:47 wyvern kernel: [<ffffffffa0163092>]
sg_new_write+0x222/0x320 [sg]
Jun 29 09:37:47 wyvern kernel: [<ffffffffa0163cf7>] sg_ioctl+0xb67/0xd40
[sg]
Jun 29 09:37:47 wyvern kernel: [<ffffffff8107caa0>] ?
autoremove_wake_function+0x0/0x40
Jun 29 09:37:47 wyvern kernel: [<ffffffff8149f4c7>] ?
_lock_kernel+0x47/0xb3
Jun 29 09:37:47 wyvern kernel: [<ffffffff811512ff>] vfs_ioctl+0x9f/0xf0
Jun 29 09:37:47 wyvern kernel: [<ffffffff811518f4>]
do_vfs_ioctl+0x84/0x590
Jun 29 09:37:47 wyvern kernel: [<ffffffff81151e81>] sys_ioctl+0x81/0xa0
Jun 29 09:37:47 wyvern kernel: [<ffffffff8100b072>]
system_call_fastpath+0x16/0x1b
Jun 29 09:37:47 wyvern kernel: ---[ end trace c893c8e73cd7fd18 ]---
This is:
1070 void __blk_put_request(struct request_queue *q, struct request
*req)
1071 {
1072 if (unlikely(!q))
1073 return;
1074 if (unlikely(--req->ref_count))
1075 return;
1076
1077 elv_completed_request(q, req);
1078
1079 /* this is a bio leak */
1080 WARN_ON(req->bio != NULL);
Adding a 1 sec sleep just before this check on sdp->detached makes it
much easier to hit, but I did hit it originally without the sleep.
707 sg_common_write(Sg_fd * sfp, Sg_request * srp,
708 unsigned char *cmnd, int timeout, int blocking)
709 {
...
731 if (sdp->detached) {
732 sg_finish_rem_req(srp);
733 return -ENODEV;
734 }
I'm not sure what needs to be done to fix this (does sg_finish_rem_req
just need to null out req->bio, or is some cleanup actually missing?),
but I'm happy to test out a fix if anyone has any suggestions.
Thanks,
Nate Dailey,
Stratus Technologies
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
