Skip to contentSkip to conversation

6ae373394c4: kernel boot crash

NFSv4.1 Fengguang Wu 2 participants 4 messages Feb 7, 2015

lore
Share & tools
lore
#0 lore raw

Hi Trond,

We noticed the below kernel crash on

git://git.linux-nfs.org/projects/trondmy/linux-nfs.git testing
commit 6ae373394c4257bad562817aa60464ff7fe8f9c4 ("NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT")

testbox/testcase/testparams: vm-vp-quantal-x86_64/boot/1

0e3b137fbf0f4ab9  6ae373394c4257bad562817aa6  
----------------  --------------------------  
       fail:runs  %reproduction    fail:runs
           |             |             |    
           :100          4%           2:46    dmesg.BUG:kernel_test_crashed

The last dmesg lines before the kernel crash are

[    0.274266] NET: Registered protocol family 1
[    0.275236] RPC: Registered named UNIX socket transport module.
[    0.276123] RPC: Registered udp transport module.
[    0.276860] RPC: Registered tcp transport module.
[    0.277617] RPC: Registered tcp NFSv4.1 backchannel transport module

Thanks,
Fengguang

---
testcase: boot
default-monitors:
default_watchdogs:
watch-oom:
watchdog:
cpufreq_governor:
model: qemu-system-x86_64 -enable-kvm
nr_vm: 64
nr_cpu: 2
memory: 360M
rootfs: quantal-core-x86_64.cgz
boot-slabinfo:
boot-meminfo:
memmap:
boot-memory:
boot-time:
kernel-size:
sleep: 1
branch: nfs/linux-next
commit: 6ae373394c4257bad562817aa60464ff7fe8f9c4
repeat_to: 2
testbox: vm-vp-quantal-x86_64-3
tbox_group: vm-vp-quantal-x86_64
kconfig: x86_64-rhel
enqueue_time: 2015-02-05 18:54:42.323364417 +08:00
kernel: "/kernel/x86_64-rhel/6ae373394c4257bad562817aa60464ff7fe8f9c4/vmlinuz-3.19.0-rc5-g6ae3733"
user: lkp
queue: rand
result_root: "/result/vm-vp-quantal-x86_64/boot/1/quantal-core-x86_64.cgz/x86_64-rhel/6ae373394c4257bad562817aa60464ff7fe8f9c4/0"
job_file: "/lkp/scheduled/vm-vp-quantal-x86_64-3/rand_boot-1-quantal-core-x86_64.cgz-x86_64-rhel-6ae373394c4257bad562817aa60464ff7fe8f9c4-1.yaml"
dequeue_time: 2015-02-05 18:55:49.602324841 +08:00
job_state: finished

#1 lore raw
Quoted 19 lines · Fengguang Wu

Hi Trond,

We noticed the below kernel crash on

git://git.linux-nfs.org/projects/trondmy/linux-nfs.git testing
commit 6ae373394c4257bad562817aa60464ff7fe8f9c4 ("NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT")

testbox/testcase/testparams: vm-vp-quantal-x86_64/boot/1

0e3b137fbf0f4ab9  6ae373394c4257bad562817aa6
----------------  --------------------------
       fail:runs  %reproduction    fail:runs
           |             |             |
           :100          4%           2:46    dmesg.BUG:kernel_test_crashed

The last dmesg lines before the kernel crash are

Quoted 8 lines
[    0.274266] NET: Registered protocol family 1
[    0.275236] RPC: Registered named UNIX socket transport module.
[    0.276123] RPC: Registered udp transport module.
[    0.276860] RPC: Registered tcp transport module.
[    0.277617] RPC: Registered tcp NFSv4.1 backchannel transport module

Thanks,
Fengguang

Hi Fengguang,

I'm trying to reproduce this issue, but so far no luck. Do you have any details on the test itself that you could share with me?

Thanks
Trond

Signature · 3 lines
Trond Myklebust
Linux NFS client maintainer, PrimaryData
trond.myklebust(a)primarydata.com
#2 lore raw

Hi Trond,

Quoted 26 lines · Trond Myklebust
Quoted 25 lines · Fengguang Wu

Hi Trond,

We noticed the below kernel crash on

git://git.linux-nfs.org/projects/trondmy/linux-nfs.git testing
commit 6ae373394c4257bad562817aa60464ff7fe8f9c4 ("NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT")

testbox/testcase/testparams: vm-vp-quantal-x86_64/boot/1

0e3b137fbf0f4ab9  6ae373394c4257bad562817aa6
----------------  --------------------------
       fail:runs  %reproduction    fail:runs
           |             |             |
           :100          4%           2:46    dmesg.BUG:kernel_test_crashed

The last dmesg lines before the kernel crash are

[    0.274266] NET: Registered protocol family 1
[    0.275236] RPC: Registered named UNIX socket transport module.
[    0.276123] RPC: Registered udp transport module.
[    0.276860] RPC: Registered tcp transport module.
[    0.277617] RPC: Registered tcp NFSv4.1 backchannel transport module
Quoted 8 lines
Quoted 2 lines

Thanks,
Fengguang

Hi Fengguang,

I'm trying to reproduce this issue, but so far no luck. Do you have any details on the test itself that you could share with me?

It looks like a false report, sorry! The real crash dmesg is much longer and it ends here:

[   14.605487] i6300esb: Unexpected close, not stopping watchdog!

Which is not relevant to the NFS change at all.

The dmesg I looked at happen to be truncated (perhaps by email server) exactly at a NFS line, which treats me to believe it's a good bisect.

Cheers,
Fengguang

#3 lore raw
Quoted 43 lines · Fengguang Wu

Hi Trond,

Quoted 26 lines · Trond Myklebust
Quoted 25 lines · Fengguang Wu

Hi Trond,

We noticed the below kernel crash on

git://git.linux-nfs.org/projects/trondmy/linux-nfs.git testing
commit 6ae373394c4257bad562817aa60464ff7fe8f9c4 ("NFSv4.1: Ask for no delegation on OPEN if using O_DIRECT")

testbox/testcase/testparams: vm-vp-quantal-x86_64/boot/1

0e3b137fbf0f4ab9  6ae373394c4257bad562817aa6
----------------  --------------------------
       fail:runs  %reproduction    fail:runs
           |             |             |
           :100          4%           2:46    dmesg.BUG:kernel_test_crashed

The last dmesg lines before the kernel crash are

[    0.274266] NET: Registered protocol family 1
[    0.275236] RPC: Registered named UNIX socket transport module.
[    0.276123] RPC: Registered udp transport module.
[    0.276860] RPC: Registered tcp transport module.
[    0.277617] RPC: Registered tcp NFSv4.1 backchannel transport module
Quoted 8 lines
Quoted 2 lines

Thanks,
Fengguang

Hi Fengguang,

I'm trying to reproduce this issue, but so far no luck. Do you have any details on the test itself that you could share with me?

It looks like a false report, sorry! The real crash dmesg is much longer and it ends here:

[   14.605487] i6300esb: Unexpected close, not stopping watchdog!
Quoted 7 lines

Which is not relevant to the NFS change at all.

The dmesg I looked at happen to be truncated (perhaps by email server) exactly at a NFS line, which treats me to believe it's a good bisect.

Cheers,
Fengguang

Phew. So I'm off the hook then. :-)

Signature · 3 lines
Trond Myklebust
Linux NFS client maintainer, PrimaryData
trond.myklebust(a)primarydata.com