LPFC issue with 2.6.12-rc3-git

Hi James,

I tried a latest git build and managed to wedge lpfc. This is a ppc64 box, and the cable was unplugged and replugged. Looks like we are stuck waiting for a scsi timeout and the other two threads are blocking on it. I waited 15 minutes to make sure it wouldnt clear itself.

Anton

scsi_wq_3 D 0000000000000000 11984 866 6 878 864 (L-TLB)

Call Trace:
[c0000000000120f0] .__switch_to+0xd0/0x160
[c00000000043a6a8] .schedule+0x488/0xe50
[c00000000043b638] .wait_for_completion+0xc8/0x150
[c0000000002efd60] .scsi_wait_req+0x90/0x100
[c0000000002f2ddc] .scsi_probe_and_add_lun+0x23c/0xa50
[c0000000002f3d04] .scsi_scan_target+0x164/0x540
[c0000000002fb5fc] .fc_scsi_scan_rport+0x2c/0x40
[c00000000006d354] .worker_thread+0x224/0x320
[c000000000074588] .kthread+0x178/0x190
[c000000000014988] .kernel_thread+0x4c/0x6c

lpfc_worker_1 D 0000000000000000 11552 864 6 866 705 (L-TLB)

Call Trace:
[c0000000000120f0] .__switch_to+0xd0/0x160
[c00000000043a6a8] .schedule+0x488/0xe50
[c00000000006db84] .flush_cpu_workqueue+0xc4/0x2a0
[c0000000002ea91c] .scsi_flush_work+0x2c/0x80
[c0000000002fd458] .fc_rport_tgt_remove+0xd8/0x110
[c0000000002fd4e0] .fc_rport_terminate+0x50/0xf0
[d00000000008673c] .lpfc_nlp_remove+0x40c/0x450 [lpfc]
[d000000000090050] .lpfc_disc_state_machine+0x1c0/0x1d0 [lpfc]
[d000000000089710] .lpfc_do_work+0x620/0xc00 [lpfc]
[c000000000074588] .kthread+0x178/0x190
[c000000000014988] .kernel_thread+0x4c/0x6c

rmmod D 000000000ff50640 10368 4127 4071 (NOTLB)

Call Trace:
[c0000000000120f0] .__switch_to+0xd0/0x160
[c00000000043a6a8] .schedule+0x488/0xe50
[c00000000006db84] .flush_cpu_workqueue+0xc4/0x2a0
[c0000000002ea91c] .scsi_flush_work+0x2c/0x80
[c0000000002fd458] .fc_rport_tgt_remove+0xd8/0x110
[c0000000002fd4e0] .fc_rport_terminate+0x50/0xf0
[c0000000002fd5c4] .fc_remove_host+0x44/0xd0
[d00000000008a650] .lpfc_pci_remove_one+0x90/0x2e0 [lpfc]
[c0000000001e87f0] .pci_device_remove+0x50/0x80
[c00000000025c194] .device_release_driver+0xa4/0xd0
[c00000000025c53c] .bus_remove_driver+0xcc/0x180
[c00000000025d038] .driver_unregister+0x18/0x40
[c0000000001e8620] .pci_unregister_driver+0x30/0xd0
[d0000000000968ec] .lpfc_exit+0x1c/0x2450 [lpfc]
[c00000000007e1c0] .sys_delete_module+0x240/0x330
[c00000000000d800] syscall_exit+0x0/0x18

Hi James,

Im also getting these errors:

lpfc 0004:21:01.0: 0:0327 Rsp ring 0 error - command completion for iotag x0 not found lpfc 0004:21:01.0: 0:0713 SCSI layer issued LUN reset (0, 0) Data: x2 x0 x0
lpfc 0004:21:01.0: 0:0317 Rsp ring 0 get: iotag x1700 greater then configured max x1000 wd0 x9032 lpfc 0004:21:01.0: 0:0327 Rsp ring 0 error - command completion for iotag x0 not found lpfc 0004:21:01.0: 0:0317 Rsp ring 0 get: iotag x1800 greater then configured max x1000 wd0 x9032

Which looks a lot like the problem last week which was solved with a firmware update. I'll look into that.

In the end the partition locked up with the following backtrace:

d000000000078198 .lpfc_sli_validate_iocb_cmd+0x58/0xe0 d000000000079e80 .lpfc_sli_abort_iocb+0xd0/0x230 d000000000092e40 .lpfc_reset_bus_handler+0x210/0x470 c0000000002ee50c .scsi_try_bus_reset+0x7c/0x160 c0000000002ef774 .scsi_error_handler+0x924/0xb20 c000000000014988 .kernel_thread+0x4c/0x6c

Anton