crossbow discuss - Jun 2009 - kernel panic: null pointer dereference in ip

I searched and did not find this anywhere.  I am running snv_111b on a Dell
2950.  I am using zones, vnics, and etherstubs.  The server is under fairly
light loads and is used for AMP services.

Thanks!

adb output:

# adb -k unix.0 vmcore.0
physmem 7ff2ef
*panicstr/s
0xfffffffffb92a180:             BAD TRAP: type=%x (#%s %s) rp=%p addr=%p
occurred in module "%s" due t
o %s
$c
tcp_do_getsockname+0xea(ffffff1a81d94980, ffffff091a3a2238,
ffffff003e581e6c)
tcp_getsockname+0x21(ffffff1a81d94780, ffffff091a3a2238, ffffff003e581e6c,
ffffff096223c388)
so_getsockname+0x67(ffffff1b28010148, ffffff091a3a2238, ffffff003e581e6c,
ffffff096223c388)
socket_getsockname+0x22(ffffff1b28010148, ffffff091a3a2238,
ffffff003e581e6c, ffffff096223c388)
getsockname+0x9c(28, 83b6f38, 83b6f24, 1)
_sys_sysenter_post_swapgs+0x14b()
$q

from /var/adm/messages:

Jun 14 15:24:41 waldorf unix: [ID 836849 kern.notice]
Jun 14 15:24:41 waldorf ^Mpanic[cpu7]/thread=ffffff09180dc200:
Jun 14 15:24:41 waldorf genunix: [ID 335743 kern.notice] BAD TRAP: type=e
(#pf Page fault) rp=ffffff003e581c80 addr=8 occurred in module "ip"
due to a
NULL pointer dereference
Jun 14 15:24:41 waldorf unix: [ID 100000 kern.notice]
Jun 14 15:24:41 waldorf unix: [ID 839527 kern.notice] httpd:
Jun 14 15:24:41 waldorf unix: [ID 753105 kern.notice] #pf Page fault
Jun 14 15:24:41 waldorf unix: [ID 532287 kern.notice] Bad kernel fault at
addr=0x8
Jun 14 15:24:41 waldorf unix: [ID 243837 kern.notice] pid=17007,
pc=0xfffffffff7b6daaa, sp=0xffffff003e581d78, eflags=0x10202
Jun 14 15:24:41 waldorf unix: [ID 211416 kern.notice] cr0:
80050033<pg,wp,ne,et,mp,pe> cr4: 6f8<xmme,fxsr,pge,mce,pae,pse,de>
Jun 14 15:24:41 waldorf unix: [ID 624947 kern.notice] cr2: 8
Jun 14 15:24:41 waldorf unix: [ID 625075 kern.notice] cr3: 86fe69000
Jun 14 15:24:41 waldorf unix: [ID 625715 kern.notice] cr8: c
Jun 14 15:24:41 waldorf unix: [ID 100000 kern.notice]
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   rdi:
ffffff1a81d94980 rsi: ffffff091a3a2238 rdx: ffffff003e581e6c
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   rcx:
ffffff096223c388  r8:                0  r9:                0
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   rax:
6 rbx: ffffff1b28010160 rbp: ffffff003e581d90
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   r10:
fffffffffb84b168 r11:                0 r12: ffffff096223c388
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   r13:
ffffff003e581e6c r14: ffffff091a3a2238 r15: ffffff1b28010148
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   fsb:
0 gsb: ffffff090f769080  ds:               4b
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]    es:
4b  fs:                0  gs:              1c3
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]   trp:
e err:                0 rip: fffffffff7b6daaa
Jun 14 15:24:41 waldorf unix: [ID 592667 kern.notice]    cs:
30 rfl:            10202 rsp: ffffff003e581d78
Jun 14 15:24:41 waldorf unix: [ID 266532 kern.notice]    ss:
38
Jun 14 15:24:41 waldorf unix: [ID 100000 kern.notice]
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581b60
unix:die+dd ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581c70
unix:trap+1752 ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581c80
unix:cmntrap+e9 ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581d90
ip:tcp_do_getsockname+ea ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581dc0
ip:tcp_getsockname+21 ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581e20
sockfs:so_getsockname+67 ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581e50
sockfs:socket_getsockname+22 ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581ec0
sockfs:getsockname+9c ()
Jun 14 15:24:41 waldorf genunix: [ID 655072 kern.notice] ffffff003e581f10
unix:brand_sys_sysenter+1e6 ()
Jun 14 15:24:41 waldorf unix: [ID 100000 kern.notice]
Jun 14 15:24:41 waldorf genunix: [ID 672855 kern.notice] syncing file
systems...
Jun 14 15:24:41 waldorf genunix: [ID 904073 kern.notice]  done

-Kristin

-- 
Mae: Come see! (points to ketchup in the refrigerator)  See, dat''s
he''s best
friend (points to the hershey''s syrup)

http://tomorrowisobsolete.blogspot.com & http://kamundse.blogspot.com
http://flickr.com/photos/kamundse/
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/crossbow-discuss/attachments/20090614/e472456d/attachment.html>

Kristin Amundsen writes:
> tcp_do_getsockname+0xea(ffffff1a81d94980, ffffff091a3a2238,
> ffffff003e581e6c)
> tcp_getsockname+0x21(ffffff1a81d94780, ffffff091a3a2238, ffffff003e581e6c,
> ffffff096223c388)
File a bug.  Your system seems to have died here:

static int
tcp_do_getsockname(tcp_t *tcp, struct sockaddr *sa, uint_t *salenp)
{
[...]
			} else {
-->				sin6->sin6_addr = tcp->tcp_ip6h->ip6_src;

... because tcp->tcp_ip6h was NULL.  It''s some sort of IPv6 problem,
possibly related to the fix for CR 6795817.

-- 
James Carlson, Solaris Networking              <james.d.carlson at
sun.com>
Sun Microsystems / 35 Network Drive        71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677

The bug #6855646 is closed because of duplicating #6877954.
What is a current state of #6877954 ?
-- 
This message posted from opensolaris.org

> The bug #6855646 is closed because of duplicating #6877954.
 > What is a current state of #6877954 ?

According to the bug database, Anders has a fix targeting build 129.

--
meem

Any workaround for that until that time?

On Thu, Sep 10, 2009 at 7:30 PM, Peter Memishian
<peter.memishian at sun.com> wrote:
>
> ?> The bug #6855646 is closed because of duplicating #6877954.
> ?> What is a current state of #6877954 ?
>
> According to the bug database, Anders has a fix targeting build 129.
>
> --
> meem
>


-- 
Piotr Jasiukajtis | estibi | SCA OS0072
http://estseg.blogspot.com

> > ??> The bug #6855646 is closed because of duplicating #6877954.
 > > ??> What is a current state of #6877954 ?
 > >
 > > According to the bug database, Anders has a fix targeting build 129.
 >
 > Any workaround for that until that time?

There''s none listed, but Anders (CC''d may know of one).

--
meem

On Thu, Sep 10, 2009 at 10:30:57AM -0700, Peter Memishian
wrote:
> 
>  > The bug #6855646 is closed because of duplicating #6877954.
>  > What is a current state of #6877954 ?
> 
> According to the bug database, Anders has a fix targeting build 129.
And a fix currently underway for OpenSolaris 2009.06.

Anders

On Thu, Sep 10, 2009 at 07:49:20PM +0200, Piotr Jasiukajtis
wrote:
> Any workaround for that until that time?
Yes, you can revert to STREAMS based sockets.

Edit /etc/sock2path and change:

	2	2	0	tcp
	2	2	6	tcp

	26	2	0	tcp
	26	2	6	tcp

To:

	2	2	0	/dev/tcp
	2	2	6	/dev/tcp

	26	2	0	/dev/tcp6
	26	2	6	/dev/tcp6

Anders

Yay, thanks for the bug fix!!  We will be glad to have this fixed.

-Kristin

-- 
Mae: Come see! (points to ketchup in the refrigerator)  See, dat''s
he''s best
friend (points to the hershey''s syrup)

http://tomorrowisobsolete.blogspot.com & http://kamundse.blogspot.com
http://flickr.com/photos/kamundse/
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/crossbow-discuss/attachments/20090910/a0263dbc/attachment.html>

Thanks!

On Thu, Sep 10, 2009 at 8:11 PM, Anders Persson <anders.persson at
sun.com> wrote:
> On Thu, Sep 10, 2009 at 07:49:20PM +0200, Piotr Jasiukajtis wrote:
>> Any workaround for that until that time?
>
> Yes, you can revert to STREAMS based sockets.
>
> Edit /etc/sock2path and change:
>
> ? ? ? ?2 ? ? ? 2 ? ? ? 0 ? ? ? tcp
> ? ? ? ?2 ? ? ? 2 ? ? ? 6 ? ? ? tcp
>
> ? ? ? ?26 ? ? ?2 ? ? ? 0 ? ? ? tcp
> ? ? ? ?26 ? ? ?2 ? ? ? 6 ? ? ? tcp
>
> To:
>
> ? ? ? ?2 ? ? ? 2 ? ? ? 0 ? ? ? /dev/tcp
> ? ? ? ?2 ? ? ? 2 ? ? ? 6 ? ? ? /dev/tcp
>
> ? ? ? ?26 ? ? ?2 ? ? ? 0 ? ? ? /dev/tcp6
> ? ? ? ?26 ? ? ?2 ? ? ? 6 ? ? ? /dev/tcp6
>
> Anders
>


-- 
Piotr Jasiukajtis | estibi | SCA OS0072
http://estseg.blogspot.com

Hi Anders and all,

i''ve been hit by this bug as well ...
> And a fix currently underway for OpenSolaris 2009.06.
... and would very much appreciate a fix for 2009.06. As the last posting is now
one month old I thought it might have already got released. Am I missing
something, pkg doesn''t tell me about a kernel update on the release
repository?

root at u0:/var/crash/u0# pkg publisher
PUBLISHER                             TYPE     STATUS   URI
opensolaris.org          (preferred)  origin   online  
http://pkg.opensolaris.org/release/
root at u0:/var/crash/u0# pkg list -uv
FMRI                                                             STATE      UFIX
pkg:/SUNWipkg at 0.5.11,5.11-0.111:20090508T161015Z                 installed 
u---
pkg:/SUNWipkg-brand at 0.5.11,5.11-0.111:20090508T161021Z           installed 
u---
pkg:/SUNWipkg-gui at 0.5.11,5.11-0.111:20090508T161022Z             installed 
u---
pkg:/SUNWipkg-um at 0.5.11,5.11-0.111:20090508T161027Z              installed 
u---

Also, thank you very much for looking after this. I was very happy about finding
this issue already fixed when I had a look at the hg commit log.
-- 
This message posted from opensolaris.org

Hi Nils,

The fix has been included in the OpenSolaris update, which I believe
is scheduled for release today. You can track what updates have been
published at:

  http://sunsolve.sun.com/show.do?target=opensolaris

There are instructions there on how to access the support repo.

Cheers,
Anders

On Tue, Oct 13, 2009 at 04:08:33AM -0700, Nils Goroll
wrote:
> Hi Anders and all,
> 
> i''ve been hit by this bug as well ...
> 
> > And a fix currently underway for OpenSolaris 2009.06.
> 
> ... and would very much appreciate a fix for 2009.06. As the last posting
is now one month old I thought it might have already got released. Am I missing
something, pkg doesn''t tell me about a kernel update on the release
repository?
> 
> root at u0:/var/crash/u0# pkg publisher
> PUBLISHER                             TYPE     STATUS   URI
> opensolaris.org          (preferred)  origin   online  
http://pkg.opensolaris.org/release/
> root at u0:/var/crash/u0# pkg list -uv
> FMRI                                                             STATE     
UFIX
> pkg:/SUNWipkg at 0.5.11,5.11-0.111:20090508T161015Z                
installed  u---
> pkg:/SUNWipkg-brand at 0.5.11,5.11-0.111:20090508T161021Z          
installed  u---
> pkg:/SUNWipkg-gui at 0.5.11,5.11-0.111:20090508T161022Z            
installed  u---
> pkg:/SUNWipkg-um at 0.5.11,5.11-0.111:20090508T161027Z             
installed  u---
> 
> Also, thank you very much for looking after this. I was very happy about
finding this issue already fixed when I had a look at the hg commit log.
> -- 
> This message posted from opensolaris.org
> _______________________________________________
> crossbow-discuss mailing list
> crossbow-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/crossbow-discuss

> The fix has been included in the OpenSolaris update, which I believe
> is scheduled for release today. You can track what updates have been
> published at:
Ah, I''d thought you actually wanted to include this as a publicly
accessible update.

Thanks for the info anyway.

Nils

On Tue, Oct 13, 2009 at 08:16:07PM +0200, Nils Goroll
wrote:
>
>> The fix has been included in the OpenSolaris update, which I believe
>> is scheduled for release today. You can track what updates have been
>> published at:
>
> Ah, I''d thought you actually wanted to include this as a publicly
accessible update.
This bug has been fixed in Nevada build 125 as well, which will be
publicly available in the dev repository.

Anders

>>> The fix has been included in the OpenSolaris update, which I
believe
>>> is scheduled for release today. You can track what updates have
been
>>> published at:
>> Ah, I''d thought you actually wanted to include this as a
publicly accessible update.
> 
> This bug has been fixed in Nevada build 125 as well, which will be
> publicly available in the dev repository.
Yes, I am aware of the fact that (all) fixes (should also) go to the dev repo. 
Thank you again, I simply misunderstood your first posting.

Nils

BTW, According to

http://sunsolve.sun.com/search/document.do?assetkey=1-68-nv_osol0906u5-1

the fix has not made it into update 5 :-/ ?!?

On Sun, Nov 08, 2009 at 07:17:38PM +0100, Nils Goroll
wrote:
> BTW, According to
>
> http://sunsolve.sun.com/search/document.do?assetkey=1-68-nv_osol0906u5-1
>
> the fix has not made it into update 5 :-/ ?!?
>
Hi Nils,

It missed update 5, but it''s in update 6.

Anders