Bug 103711 - Null pointer dereference in md raid 5 on get_free_stripe
Summary: Null pointer dereference in md raid 5 on get_free_stripe
Status: NEW
Alias: None
Product: IO/Storage
Classification: Unclassified
Component: MD (show other bugs)
Hardware: x86-64 Linux
: P1 high
Assignee: io_md
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-08-29 14:40 UTC by e-kernel
Modified: 2016-02-16 22:05 UTC (History)
3 users (show)

See Also:
Kernel Version: 4.1.6-1-ARCH
Subsystem:
Regression: No
Bisected commit-id:

Attachments
kernel crash trace (11.62 KB, text/x-log)
2015-08-29 14:40 UTC, e-kernel 		Details
Kernel Log (13.16 KB, text/plain)
2015-09-03 01:23 UTC, Gene 		Details
ver_linux output (10.22 KB, text/plain)
2015-09-03 01:25 UTC, Gene 		Details
lspci -vvv output (22.14 KB, text/plain)
2015-09-03 01:25 UTC, Gene 		Details
/proc/iomem (2.75 KB, text/plain)
2015-09-03 01:26 UTC, Gene 		Details
/proc/ioreports (1.57 KB, text/plain)
2015-09-03 01:27 UTC, Gene 		Details
/proc/scsi/scsi (1.57 KB, text/plain)
2015-09-03 01:27 UTC, Gene 		Details
Add an attachment (proposed patch, testcase, etc.)

Description e-kernel 2015-08-29 14:40:20 UTC 
Created attachment 186171 [details]
kernel crash trace

My file server just crashed with the attached kernel trace.
Running arch linux with kernel 4.1.6.

Raid configuration:
Personalities : [raid6] [raid5] [raid4] 
md127 : active raid5 sdd1[2] sdc1[4] sda1[1] sdb1[0]
      5860147200 blocks super 1.2 level 5, 512k chunk, algorithm 2 [4/4] [UUUU]
      bitmap: 0/15 pages [0KB], 65536KB chunk

Hardware:
4x2TB hard drives
Intel(R) Atom(TM) CPU  C2550  @ 2.40GHz
8GB ECC UDIMM
Comment 1 Gene 2015-09-03 01:23:03 UTC 
I have the same problem. Takes quite a while to get this to happen. Typically 1 - 2 weeks of uptime. Best I can tell it crashes when the nightly backups are running and writing to the raid6 array.

Same kernel - 4.1.6 - details attached.

cat /proc/mdstat 
Personalities : [raid6] [raid5] [raid4] 
md0 : active raid6 sdf1[4] sdc1[1] sdh1[5] sde1[3] sdd1[2] sdb1[0]
      11720536064 blocks super 1.2 level 6, 512k chunk, algorithm 2 [6/6] [UUUUUU]
      bitmap: 0/22 pages [0KB], 65536KB chunk

unused devices: <none>
Comment 2 Gene 2015-09-03 01:23:49 UTC 
Created attachment 186531 [details]
Kernel Log
Comment 3 Gene 2015-09-03 01:25:08 UTC 
Created attachment 186541 [details]
ver_linux output
Comment 4 Gene 2015-09-03 01:25:46 UTC 
Created attachment 186551 [details]
lspci -vvv output
Comment 5 Gene 2015-09-03 01:26:50 UTC 
Created attachment 186561 [details]
/proc/iomem
Comment 6 Gene 2015-09-03 01:27:08 UTC 
Created attachment 186571 [details]
/proc/ioreports
Comment 7 Gene 2015-09-03 01:27:52 UTC 
Created attachment 186581 [details]
/proc/scsi/scsi
Comment 8 Neil Brown 2015-09-03 05:14:29 UTC 
Fix by commit 
49895bcc7e566ba455eb2996607d6fbd3447ce16

should get into 4.1-stable eventually.
Comment 9 Gene 2015-09-03 12:35:10 UTC 
Thanks Neil - assume this is in 4.2 already.
Note You need to log in before you can comment on or make changes to this bug.