Re: Null pointer dereference during iSCSI login

2009-03-03 Thread Mike Christie

Mike Christie wrote:
> Ulrich Windl wrote:
>> Hello,
>>
>> with SLES10 SP1 on x86_64 (open-iscsi-2.0.707-0.32) I'm seeing a problem 
>> during 
>> login using "iscsiadm -m node -L automatic". After a few logins, login 
>> suddenly 
>> fails:
> 
> I think you will have to make a suse bugzilla or try to recreate with 
> the current open-iscsi code. open-iscsi-2.0.707 is a little old and I 
> have no idea what suse adds to it.
> 
> I think suse might actually have this one fixed though. You can contact 
> Hannes Reinecke . I think he gets the suse bugzilla for 
> iscsi too, so either way you will reach him.
> 

Oh yeah, hannes, does the iscsi code in that kernel use sock_no_sendpage 
or kernel_sendmsg?

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---



Re: Null pointer dereference during iSCSI login

2009-03-02 Thread Mike Christie

Ulrich Windl wrote:
> Hello,
> 
> with SLES10 SP1 on x86_64 (open-iscsi-2.0.707-0.32) I'm seeing a problem 
> during 
> login using "iscsiadm -m node -L automatic". After a few logins, login 
> suddenly 
> fails:

I think you will have to make a suse bugzilla or try to recreate with 
the current open-iscsi code. open-iscsi-2.0.707 is a little old and I 
have no idea what suse adds to it.

I think suse might actually have this one fixed though. You can contact 
Hannes Reinecke . I think he gets the suse bugzilla for 
iscsi too, so either way you will reach him.

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---



Re: Null pointer dereference during iSCSI login

2009-03-02 Thread Ulrich Windl

On 2 Mar 2009 at 7:46, Konrad Rzeszutek wrote:

> 
> On Mon, Mar 02, 2009 at 10:53:19AM +0100, Ulrich Windl wrote:
> > 
> > Hello,
> > 
> > with SLES10 SP1 on x86_64 (open-iscsi-2.0.707-0.32) I'm seeing a problem 
> > during 
> > login using "iscsiadm -m node -L automatic". After a few logins, login 
> > suddenly 
> 
> What happens if you use the latest version of Open-iSCSI kernel modules and 
> Open-iSCSI utils?
> (You will need to patch the kernel directory with the 
> kernel/2.6.16-suse.patch for them to compile).

Unfortunately I have no development environment ready at the moment, and the 
machine is serving some production users. Plans are to upgrade to a newer 
service 
pack, but I thought this problem could be interesting to examine.

I can apply ksymoops to dmesg output:
Unable to handle kernel NULL pointer dereference at 0232 RIP:
{inet_sendmsg+23}
Oops:  [1] SMP
CPU 3
Pid: 25485, comm: scsi_wq_25 Not tainted 2.6.16.54-0.2.11-xen #1
RIP: e030:[] {inet_sendmsg+23}
Using defaults from ksymoops -t elf64-x86-64 -a i386:x86-64
RSP: e02b:880011e0db78  EFLAGS: 00010296
RAX: 802f1c40 RBX:  RCX: 0200
RDX: 880011e0dd58 RSI: 8800080988c0 RDI: 880011e0dba8
RBP: 0200 R08: 0200 R09: 8000
R10: dbb545c6 R11: 0001 R12: 880011e0dd58
R13: 880011e0dba8 R14: 88000bdc52c0 R15: 0200
FS:  2b77ef71e6d0() GS:803a2180() knlGS:
CS:  e033 DS:  ES: 
Stack: 0030 8800080988c0 0200 880011e0dd58
    8026e1da 0018 880011e0e000
    0001
Call Trace: {sock_sendmsg+249} 
{__kprobes_text_start+845}
   {autoremove_wake_function+0} 
{__alloc_pages+101}
   {kernel_sendmsg+53} 
{sock_no_sendpage+130}
   {monotonic_clock+53} 
{:iscsi_tcp:iscsi_tcp_mtask_xmit+502}
   {:libiscsi:iscsi_xmitworker+0} 
{:libiscsi:iscsi_xmit_mtask+84}
   {:libiscsi:iscsi_xmitworker+310} 
{run_workqueue+148}
   {worker_thread+0} 
{keventd_create_kthread+0}
   {worker_thread+240} 
{default_wake_function+0}
   {keventd_create_kthread+0} 
{keventd_create_kthread+0}
   {kthread+212} {child_rip+8}
   {keventd_create_kthread+0} 
{kthread+0}
   {child_rip+0}
Code: 66 83 bb 32 02 00 00 00 75 0c 48 89 df e8 ad f5 ff ff 85 c0


>>RIP; 802ba089<=

>>RAX; 802f1c40 
>>RDX; 880011e0dd58 <__start___xen_guest+880011a15a78/7fd07d20>
>>RSI; 8800080988c0 <__start___xen_guest+880007ca05e0/7fd07d20>
>>RDI; 880011e0dba8 <__start___xen_guest+880011a158c8/7fd07d20>
>>R10; dbb545c6 <__start___xen_guest+db75c2e6/7fd07d20>
>>R12; 880011e0dd58 <__start___xen_guest+880011a15a78/7fd07d20>
>>R13; 880011e0dba8 <__start___xen_guest+880011a158c8/7fd07d20>
>>R14; 88000bdc52c0 <__start___xen_guest+88000b9ccfe0/7fd07d20>

Trace; 8026e1da 
Trace; 8014195d 
Trace; 8026fac1 
Trace; 8010df38 
Trace; 8839ba18 <_end+7f1151c/7f175b04>
Trace; 8839bb4e <_end+7f11652/7f175b04>
Trace; 8013e34e 
Trace; 8013e43e 
Trace; 80141582 
Trace; 80141826 
Trace; 80141582 
Trace; 8010baae 

Code;  802ba089 
 <_RIP>:
Code;  802ba089<=
   0:   66 83 bb 32 02 00 00  cmpw   $0x0,0x232(%rbx)   <=
Code;  802ba090 
   7:   00
Code;  802ba091 
   8:   75 0c jne16 <_RIP+0x16>
Code;  802ba093 
   a:   48 89 df  mov%rbx,%rdi
Code;  802ba096 
   d:   e8 ad f5 ff ffcallq  f5bf 
<_RIP+0xf5bf>
Code;  802ba09b 
  12:   85 c0 test   %eax,%eax

CR2: 0232


> 
> I see the 'kprobes' symbol in there - do you have any of them loaded?

Not that I knew; I only had heard the word before, but never did anything with 
that.

Regards,
Ulrich


--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---



Re: Null pointer dereference during iSCSI login

2009-03-02 Thread Konrad Rzeszutek

On Mon, Mar 02, 2009 at 10:53:19AM +0100, Ulrich Windl wrote:
> 
> Hello,
> 
> with SLES10 SP1 on x86_64 (open-iscsi-2.0.707-0.32) I'm seeing a problem 
> during 
> login using "iscsiadm -m node -L automatic". After a few logins, login 
> suddenly 

What happens if you use the latest version of Open-iSCSI kernel modules and 
Open-iSCSI utils?
(You will need to patch the kernel directory with the kernel/2.6.16-suse.patch 
for them to compile).

I see the 'kprobes' symbol in there - do you have any of them loaded?

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---



Re: Null pointer dereference during iSCSI login

2009-03-02 Thread Ulrich Windl

Hi!

As an add-on to this problem: Any subsequent attempt to run "iscsiadm -m node 
-L 
automatic" just silently hangs. Trace follows:

# strace iscsiadm -m node -L automatic
execve("/sbin/iscsiadm", ["iscsiadm", "-m", "node", "-L", "automatic"], [/* 56 
vars */]) = 0
brk(0)  = 0x522000
[...]
umask(0177) = 022
socket(PF_FILE, SOCK_STREAM, 0) = 3
connect(3, {sa_family=AF_FILE, pa...@iscsiadm_abstract_namespace}, 110) = 0
write(3, "\10\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 2504) = 
2504
recvfrom(3,  

Regards,
Ulrich


--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---