Bug 4995

Summary: Kernel came out with OOPs when tried to kill vmmstress [mmap] programs
Product: Memory Management Reporter: Sharyathi (sharyathi)
Component: OtherAssignee: Andrew Morton (akpm)
Status: REJECTED INSUFFICIENT_DATA    
Severity: high CC: bnpoorni, bunk, sglass, ssant
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.13-rc4-mm1 Subsystem:
Regression: --- Bisected commit-id:
Attachments: This is the dump analysis taken after observing OOps

Description Sharyathi 2005-08-04 05:56:09 UTC
Distribution:
RHEL 4 Update 1
-----------------------------
Hardware Environment:
SMP, 2 way Intel(R) Xeon(TM) CPU 2.40GHz, 1.5 GB RAM
Network Interface (Tigon3)
Disk I/O
SCSI storage controller: usion-MPT Dual Ultra320
-----------------------------
Software Environment:
Linux x225b.in.ibm.com 2.6.13-rc4-mm1 #1 Thu Jul 28 17:43:05 IST 2005 i686 i686
i386 GNU/Linux
-----------------------------
Problem Description:
Kernel came out with this OOPs message when tried to kill VM stress Test suite..

------------[ cut here ]------------
kernel BUG at <bad filename>:63113!
invalid operand: 0000 [#1]
SMP
last sysfs file: /class/net/eth0/ifindex
Modules linked in: ipv6 ipt_REJECT ipt_state ip_conntrack iptable_filter
ip_tables button battery ac i2e
CPU:    2
EIP:    0060:[<c015ee6c>]    Not tainted VLI
EFLAGS: 00010286   (2.6.13-rc4-mm1-I)
EIP is at exit_mmap+0x15c/0x170
eax: 00000068   ebx: 00000000   ecx: f7fa4b80   edx: ffffffff
esi: 00000000   edi: c5fc7680   ebp: c5fc76cc   esp: d11a4e50
ds: 007b   es: 007b   ss: 0068
Process mmap3 (pid: 8080, threadinfo=d11a4000 task=e89fe030)
Stack: 00000000 ffffffff d11a4e64 00000000 c5c1f900 00014071 c5fc7680 c5fc76dc
       e89fe030 e89fe4d8 c01235f4 f707ee00 f707ee00 f707ee00 c01284e5 d11a4000
       d11a4f20 c0130315 00000008 00000001 0000000f 00000009 f707ee00 d11a4000
Call Trace:
 [<c01235f4>] mmput+0x54/0xf0
 [<c01284e5>] do_exit+0xe5/0x430
 [<c0130315>] __dequeue_signal+0x105/0x1b0
 [<c01288a7>] do_group_exit+0x37/0xa0
 [<c0132074>] get_signal_to_deliver+0x234/0x340
 [<c0103e50>] do_signal+0x60/0x140
 [<c0515e65>] _read_unlock_irq+0x5/0x10
 [<c014aebc>] find_get_pages_tag+0x6c/0x80
 [<c014a51d>] wait_on_page_writeback_range+0x6d/0x120
 [<c016c271>] do_fsync+0x41/0xd0
 [<c0103f67>] do_notify_resume+0x37/0x3c
 [<c01040fa>] work_notifysig+0x13/0x15
Code: c3 c7 43 08 00 00 00 00 8b 03 e8 c0 8b fb ff 8b 53 04 83 fa ff 74 8e 8d 43
14 e8 80 34 00 00 c7 4
 <0>Fatal exception: panic in 5 seconds
Kernel panic - not syncing: Fatal exception

-----------------------------
Comment 1 Andrew Morton 2005-08-04 10:26:12 UTC
bugme-daemon@kernel-bugs.osdl.org wrote:
>

This:

> kernel BUG at <bad filename>:63113!

Is very irritating.  Why is it happening?  Can you find some way to fix it?

Comment 2 Sharyathi 2005-08-05 04:44:11 UTC
Created attachment 5516 [details]
This is the dump analysis taken after observing OOps

Andrew 
  I was able to replicate it following same process[killing vmmstress]
  I am attaching the Preliminary analysis of the dump.
It seems to be the problem with mm/mmap.c line 1962. I couldn't exactly figure
out the cause for the OOPs
Comment 3 Andrew Morton 2006-01-16 00:53:44 UTC
Is this fixed?
Comment 4 Sharyathi 2006-01-16 21:17:45 UTC
I looked into 2.6.15-rc7 but the problem was not observed. I am looking into it
will update later.
Comment 5 Adrian Bunk 2006-04-22 10:17:30 UTC
Sharyathi, what is the status of this issue in kernel 2.6.16?
Comment 6 Adrian Bunk 2006-08-06 03:47:05 UTC
Please reopen this bug if it's still present in kernel 2.6.17.