Public bug reported:

Problem Description:
=============
- PCI passthru Qlogic Fibre Channel adapter from Ubuntu 18.04 KVM host to 
Ubuntu 18.04 KVM guest.

- Stress test on Qlogic Fibre Channel on Ubuntu KVM guest caused KVM
host crashed in qlt_free_session_done call.

- Below stack traces from KVM host:


91:mon> t
[c000200e4e81fb60] c00800001162f044 qlt_free_session_done+0x4ec/0x680 [qla2xxx] 
(unreliable)
[c000200e4e81fc90] c00000000012fbb8 process_one_work+0x298/0x5a0
[c000200e4e81fd20] c00000000012ff58 worker_thread+0x98/0x630
[c000200e4e81fdc0] c000000000138ae8 kthread+0x1a8/0x1b0
[c000200e4e81fe30] c00000000000b528 ret_from_kernel_thread+0x5c/0xb4

91:mon> e
cpu 0x91: Vector: 300 (Data Access) at [c000200e4e81f8e0]
    pc: c00800001162ed58: qlt_free_session_done+0x200/0x680 [qla2xxx]
    lr: c00800001162eca8: qlt_free_session_done+0x150/0x680 [qla2xxx]
    sp: c000200e4e81fb60
   msr: 900000000280b033
   dar: 20
 dsisr: 40000000
  current = 0xc000200e4e7b0e00
  paca    = 0xc00000000fae3b00   softe: 0        irq_happened: 0x01
    pid   = 1119, comm = kworker/145:1
Linux version 4.15.0-041500rc9-generic (kernel@tangerine) (gcc version 7.2.0 
(Ubuntu 7.2.0-6ubuntu1)) #201801212130 SMP Mon Jan 22 03:36:42 UTC 2018


91:mon> r
R00 = c00800001162eca8   R16 = 0000000000000000
R01 = c000200e4e81fb60   R17 = 0000000000000000
R02 = c00800001166ad60   R18 = 0000000000000000
R03 = 0000000000000001   R19 = 0000000000000000
R04 = c000200e44f8c7f8   R20 = c000200e618e7d80
R05 = 000000000000f087   R21 = 0000000000000000
R06 = c00800001165e6c8   R22 = 0000000000000001
R07 = c00800001164adb0   R23 = c000200e44f99d24
R08 = 0000000000000000   R24 = 0000000000000402
R09 = 0000000000000000   R25 = 0000000000000000
R10 = 0000000000000000   R26 = c000000fe1270c20
R11 = c00800001163e170   R27 = c000200e44f99000
R12 = c000000000cfccf0   R28 = c00800001164adb0
R13 = c00000000fae3b00   R29 = c000000fe1270c00
R14 = c000000000138948   R30 = c000200e44f8c7f8
R15 = c000200e4f019440   R31 = c000000fe1270cc0
pc  = c00800001162ed58 qlt_free_session_done+0x200/0x680 [qla2xxx]
cfar= c00800001162ed1c qlt_free_session_done+0x1c4/0x680 [qla2xxx]
lr  = c00800001162eca8 qlt_free_session_done+0x150/0x680 [qla2xxx]
msr = 900000000280b033   cr  = 28002284
ctr = c000000000cfccf0   xer = 0000000000000000   trap =  300
dar = 0000000000000020   dsisr = 40000000
91:mon>

The crash location seems close to this one fixed about two weeks ago:

https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-
next.git/commit/drivers/scsi/qla2xxx/qla_os.c?h=next-20180212&id=2ce87cc5b269510de9ca1185ca8a6e10ec78c069

scsi: qla2xxx: Fix memory corruption during hba reset test
This patch fixes memory corrpution while performing HBA Reset test.

Following stack trace is seen:

[  466.397219] BUG: unable to handle kernel NULL pointer dereference at 
0000000000000020
[  466.433669] IP: [<ffffffffc06f5dd0>] qlt_free_session_done+0x260/0x5f0 
[qla2xxx]
[  466.467731] PGD 0
[  466.476718] Oops: 0000 [#1] SMP

- Luciano built and provided the patch with new Qlogic change on Friday
last week.

root@bos1u1p1:~/chavez# ls linux-image*
linux-image-4.15.0-041500rc9-generic_4.15.0-041500rc9.201801212130_ppc64el.deb
linux-image-extra-4.15.0-041500rc9-generic_4.15.0-041500rc9.201801212130_ppc64el.deb

- I configured and ran same test over weekend and test ran good. KVM
host did not crash in qlt_free_session_done call like before.

- So the patch fixed the problem.

Hi Canonical,

Please review and consider this a request to pull in commit
2ce87cc5b269510de9ca1185ca8a6e10ec78c069 please. Thanks!

** Affects: linux (Ubuntu)
     Importance: Undecided
     Assignee: Ubuntu on IBM Power Systems Bug Triage (ubuntu-power-triage)
         Status: New


** Tags: architecture-ppc64le bugnameltc-164551 severity-high 
targetmilestone-inin1804

** Tags added: architecture-ppc64le bugnameltc-164551 severity-high
targetmilestone-inin1804

** Changed in: ubuntu
     Assignee: (unassigned) => Ubuntu on IBM Power Systems Bug Triage 
(ubuntu-power-triage)

** Package changed: ubuntu => linux (Ubuntu)

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1750441

Title:
  Boston-LC:bos1u1: Stress test on Qlogic Fibre Channel on Ubuntu KVM
  guest that caused KVM host crashed in qlt_free_session_done call

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1750441/+subscriptions

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to