kernel-packages team mailing list archive

[Chris J Arges <1468416@xxxxxxxxxxxxxxxxxx>]

Sent to kteam ML.

** Description changed:

- Hi Canonical,
+ SRU Justification:
  
- Please apply this kernel commit for Trusty SRU / 14.04.x (already in 14.10, 15.04, and later):
- - 27f344eb15dd0da80ebec80c7245e8c85043f841 lpfc: Add iotag memory barrier
+ [Impact]
  
- Tested on top of linux 3.13.0-55.94 (applies cleanly; cherrypick)
+ The following issue occurs with the lpfc driver, seen under I/O stress.
  
- It fixes the this issue w/ the lpfc driver, seen under I/O stress.
- 
- (In reply to comment #12)
  > [  177.972210] lpfc 0000:00:05.0: 0:0369 No entry from fast-path completion
  > queue fcpcqid=1
  > [  355.950403] lpfc 0000:00:04.0: 1:0372 iotag x85d is out off range: max
  > iotag (xa10)
  > [  421.007268] lpfc 0000:00:04.0: 1:(0):0748 abort handler timed out waiting
  > for abortng I/O (xri:xdf7) to complete: ret 0x2003, ID 0, LUN 2
  > [  421.008177] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_LUN_RESET to TGT 0 LUN
  > 2 failed (1, 4096) iocb_flag x6
  > [  421.008309] lpfc 0000:00:04.0: 1:(0):0713 SCSI layer issued Device Reset
  > (0, 2) return x2002
  > [  481.017214] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > LUN : cnt x1
  > [  481.019292] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_TARGET_RESET to TGT 0
  > LUN 2 failed (1, 4096) iocb_flag x6
  > [  481.019398] lpfc 0000:00:04.0: 1:(0):0723 SCSI layer issued Target Reset
  > (0, 2) return x2002
  > [  541.027198] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > TGT : cnt x1
  > [  541.029168] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_TARGET_RESET to TGT 0
  > LUN 0 failed (1, 4096) iocb_flag x6
  ...
  > [  601.037193] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > HOST : cnt x1
  > [  601.037291] lpfc 0000:00:04.0: 1:(0):0714 SCSI layer issued Bus Reset
  > Data: x2003
  > [  601.037388] lpfc 0000:00:04.0: 1:(0):3172 SCSI layer issued Host Reset
  > Data:
  > [  603.560526] lpfc 0000:00:04.0: 1:1303 Link Up Event x1 received Data: x1
  > x0 x80 x0 x0 x0 0
  > [  612.687916] sd 2:0:0:0: timing out command, waited 30s
  > [  612.689217] sd 2:0:0:1: timing out command, waited 30s
  > [  612.690407] sd 2:0:0:3: timing out command, waited 30s
  > [  612.694220] sd 2:0:0:4: timing out command, waited 30s
  > [  634.667144]  rport-2:0-4: blocked FC remote port time out: removing rport
+ 
+ [Fix]
+ 
+ 27f344eb15dd0da80ebec80c7245e8c85043f841 lpfc: Add iotag memory barrier
+ 
+ Tested on top of linux 3.13.0-55.94 (applies cleanly; cherrypick)
+ 
+ [Test Case]
+ Stress I/O using the lpfc driver.

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1468416

Title:
  [Trusty SRU] patch for "lpfc: iotag is out of range: max iotag"

Status in linux package in Ubuntu:
  Fix Released
Status in linux source package in Trusty:
  In Progress

Bug description:
  SRU Justification:

  [Impact]

  The following issue occurs with the lpfc driver, seen under I/O
  stress.

  > [  177.972210] lpfc 0000:00:05.0: 0:0369 No entry from fast-path completion
  > queue fcpcqid=1
  > [  355.950403] lpfc 0000:00:04.0: 1:0372 iotag x85d is out off range: max
  > iotag (xa10)
  > [  421.007268] lpfc 0000:00:04.0: 1:(0):0748 abort handler timed out waiting
  > for abortng I/O (xri:xdf7) to complete: ret 0x2003, ID 0, LUN 2
  > [  421.008177] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_LUN_RESET to TGT 0 LUN
  > 2 failed (1, 4096) iocb_flag x6
  > [  421.008309] lpfc 0000:00:04.0: 1:(0):0713 SCSI layer issued Device Reset
  > (0, 2) return x2002
  > [  481.017214] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > LUN : cnt x1
  > [  481.019292] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_TARGET_RESET to TGT 0
  > LUN 2 failed (1, 4096) iocb_flag x6
  > [  481.019398] lpfc 0000:00:04.0: 1:(0):0723 SCSI layer issued Target Reset
  > (0, 2) return x2002
  > [  541.027198] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > TGT : cnt x1
  > [  541.029168] lpfc 0000:00:04.0: 1:(0):0727 TMF FCP_TARGET_RESET to TGT 0
  > LUN 0 failed (1, 4096) iocb_flag x6
  ...
  > [  601.037193] lpfc 0000:00:04.0: 1:(0):0724 I/O flush failure for context
  > HOST : cnt x1
  > [  601.037291] lpfc 0000:00:04.0: 1:(0):0714 SCSI layer issued Bus Reset
  > Data: x2003
  > [  601.037388] lpfc 0000:00:04.0: 1:(0):3172 SCSI layer issued Host Reset
  > Data:
  > [  603.560526] lpfc 0000:00:04.0: 1:1303 Link Up Event x1 received Data: x1
  > x0 x80 x0 x0 x0 0
  > [  612.687916] sd 2:0:0:0: timing out command, waited 30s
  > [  612.689217] sd 2:0:0:1: timing out command, waited 30s
  > [  612.690407] sd 2:0:0:3: timing out command, waited 30s
  > [  612.694220] sd 2:0:0:4: timing out command, waited 30s
  > [  634.667144]  rport-2:0-4: blocked FC remote port time out: removing rport

  [Fix]

  27f344eb15dd0da80ebec80c7245e8c85043f841 lpfc: Add iotag memory
  barrier

  Tested on top of linux 3.13.0-55.94 (applies cleanly; cherrypick)

  [Test Case]
  Stress I/O using the lpfc driver.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1468416/+subscriptions