Kernel OOPS with aacraid driver

Kernel OOPS with aacraid driver

[Klaus Steinberger]

Hello,

since Kernel Version 2.6.10 I get kernel OOPS when I try to use the Adaptec 
Raidtools (ASMBE as well as the CLI tool: aacli). I reported this to Fedora 
Bugzilla as the following bug ( 148761 )

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=148761

They told me to report that to upstream.

Apr 21 10:53:33 www adaptec-smbe: arcpd startup succeeded
Apr 21 10:53:36 www adaptec-smbe: anotifyd startup succeeded
Apr 21 10:53:36 www kernel: Unable to handle kernel NULL pointer dereference 
at
virtual address 00000034
Apr 21 10:53:36 www kernel:  printing eip:
Apr 21 10:53:36 www kernel: c03227a0
Apr 21 10:53:36 www kernel: *pde = 359b2001
Apr 21 10:53:36 www kernel: Oops: 0000 [#2]
Apr 21 10:53:36 www kernel: SMP
Apr 21 10:53:36 www kernel: Modules linked in: md5 ipv6 iptable_filter 
ip_tables
tg3 floppy sg microcode ohci_hcd video button battery ac ext3 jbd dm_mod 
aacraid
aic79xx sd_modscsi_mod
Apr 21 10:53:36 www kernel: CPU:    3
Apr 21 10:53:36 www kernel: EIP:    0060:[<c03227a0>]    Not tainted VLI
Apr 21 10:53:36 www kernel: EFLAGS: 00010086   (2.6.11-1.14_FC2.rootsmp)
Apr 21 10:53:36 www kernel: EIP is at _spin_lock_irqsave+0x10/0x60
Apr 21 10:53:36 www kernel: eax: 00000030   ebx: 00000286   ecx: bffff080   
edx:
bffff080
Apr 21 10:53:36 www kernel: esi: 00000030   edi: 00000030   ebp: bffff080   
esp:
f5ecff14
Apr 21 10:53:36 www kernel: ds: 007b   es: 007b   ss: 0068
Apr 21 10:53:36 www kernel: Process aacsnmpd (pid: 3091, threadinfo=f5ecf000
task=f5a01020)
Apr 21 10:53:36 www kernel: Stack: 00000000 0000fe01 00042008 00000000 
f886ac4a
00042008 00000000 00000000
Apr 21 10:53:36 www kernel:        f8869b80 42676a06 00000000 42676a06 
00000000
42676a06 00042008 bffff080
Apr 21 10:53:36 www kernel:        00000000 bffff080 f886a560 f8871d80 
f54da880
ffffffe7 c017580a bffff080
Apr 21 10:53:36 www kernel: Call Trace:
Apr 21 10:53:36 www kernel:  [<f886ac4a>] fib_alloc+0x1a/0x70 [aacraid]
Apr 21 10:53:36 www kernel:  [<f8869b80>] ioctl_send_fib+0x10/0x110 [aacraid]
Apr 21 10:53:36 www kernel:  [<f886a560>] aac_do_ioctl+0x60/0xcf [aacraid]
Apr 21 10:53:36 www kernel:  [<c017580a>] do_ioctl+0x6a/0x90
Apr 21 10:53:36 www kernel:  [<c0175a7e>] vfs_ioctl+0x5e/0x1d0
Apr 21 10:53:36 www kernel:  [<c0175c4b>] sys_ioctl+0x5b/0x90
Apr 21 10:53:36 www kernel:  [<c0104203>] syscall_call+0x7/0xb
Apr 21 10:53:36 www kernel: Code: af de 75 0c f0 83 28 01 79 05 e8 9c e4 ff ff
c3 0f 0b d7 00 78 a7 33 c0 eb ea 90 83 ec 10 89 74 24 0c 89 c6 89 5c 24 08 9c 
5b
fa <81> 78 04 ad4e ad de 75 26 f0 fe 0e 79 13 f7 c3 00 02 00 00 74

The only workaround I currently have is to use the driver from Adaptec.

Sincerly,
Klaus Steinberger

-- 
Klaus Steinberger         Maier-Leibnitz Labor
Phone: (+49 89)289 14287  Am Coulombwall 6, D-85748 Garching, Germany
FAX:   (+49 89)289 14280  EMail: Klaus.Steinberger@Physik.Uni-Muenchen.DE
URL: http://www.physik.uni-muenchen.de/~k2/

In a world without Walls and Fences, who needs Windows and Gates

Re: Kernel OOPS with aacraid driver

[Randy.Dunlap]

On Fri, 22 Apr 2005 07:49:56 +0200 Klaus Steinberger wrote:

| Hello,
| 
| since Kernel Version 2.6.10 I get kernel OOPS when I try to use the Adaptec 
| Raidtools (ASMBE as well as the CLI tool: aacli). I reported this to Fedora 
| Bugzilla as the following bug ( 148761 )
| 
| https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=148761
| 
| They told me to report that to upstream.
| 
| Apr 21 10:53:33 www adaptec-smbe: arcpd startup succeeded
| Apr 21 10:53:36 www adaptec-smbe: anotifyd startup succeeded
| Apr 21 10:53:36 www kernel: Unable to handle kernel NULL pointer dereference 
| at
| virtual address 00000034
| Apr 21 10:53:36 www kernel:  printing eip:
| Apr 21 10:53:36 www kernel: c03227a0
| Apr 21 10:53:36 www kernel: *pde = 359b2001
| Apr 21 10:53:36 www kernel: Oops: 0000 [#2]

BTW, this says that this is the second oops:  "[#2]".
What was the first one?

| Apr 21 10:53:36 www kernel: SMP
| Apr 21 10:53:36 www kernel: Modules linked in: md5 ipv6 iptable_filter 
| ip_tables
| tg3 floppy sg microcode ohci_hcd video button battery ac ext3 jbd dm_mod 
| aacraid
| aic79xx sd_modscsi_mod
| Apr 21 10:53:36 www kernel: CPU:    3
| Apr 21 10:53:36 www kernel: EIP:    0060:[<c03227a0>]    Not tainted VLI
| Apr 21 10:53:36 www kernel: EFLAGS: 00010086   (2.6.11-1.14_FC2.rootsmp)
| Apr 21 10:53:36 www kernel: EIP is at _spin_lock_irqsave+0x10/0x60
| Apr 21 10:53:36 www kernel: eax: 00000030   ebx: 00000286   ecx: bffff080   
| edx:
| bffff080
| Apr 21 10:53:36 www kernel: esi: 00000030   edi: 00000030   ebp: bffff080   
| esp:
| f5ecff14
| Apr 21 10:53:36 www kernel: ds: 007b   es: 007b   ss: 0068
| Apr 21 10:53:36 www kernel: Process aacsnmpd (pid: 3091, threadinfo=f5ecf000
| task=f5a01020)
| Apr 21 10:53:36 www kernel: Stack: 00000000 0000fe01 00042008 00000000 
| f886ac4a
| 00042008 00000000 00000000
| Apr 21 10:53:36 www kernel:        f8869b80 42676a06 00000000 42676a06 
| 00000000
| 42676a06 00042008 bffff080
| Apr 21 10:53:36 www kernel:        00000000 bffff080 f886a560 f8871d80 
| f54da880
| ffffffe7 c017580a bffff080
| Apr 21 10:53:36 www kernel: Call Trace:
| Apr 21 10:53:36 www kernel:  [<f886ac4a>] fib_alloc+0x1a/0x70 [aacraid]
| Apr 21 10:53:36 www kernel:  [<f8869b80>] ioctl_send_fib+0x10/0x110 [aacraid]
| Apr 21 10:53:36 www kernel:  [<f886a560>] aac_do_ioctl+0x60/0xcf [aacraid]
| Apr 21 10:53:36 www kernel:  [<c017580a>] do_ioctl+0x6a/0x90
| Apr 21 10:53:36 www kernel:  [<c0175a7e>] vfs_ioctl+0x5e/0x1d0
| Apr 21 10:53:36 www kernel:  [<c0175c4b>] sys_ioctl+0x5b/0x90
| Apr 21 10:53:36 www kernel:  [<c0104203>] syscall_call+0x7/0xb
| Apr 21 10:53:36 www kernel: Code: af de 75 0c f0 83 28 01 79 05 e8 9c e4 ff ff
| c3 0f 0b d7 00 78 a7 33 c0 eb ea 90 83 ec 10 89 74 24 0c 89 c6 89 5c 24 08 9c 
| 5b
| fa <81> 78 04 ad4e ad de 75 26 f0 fe 0e 79 13 f7 c3 00 02 00 00 74
| 
| The only workaround I currently have is to use the driver from Adaptec.

---
~Randy

Re: Kernel OOPS with aacraid driver

[Klaus Steinberger]

Hello Randy,

> BTW, this says that this is the second oops:  "[#2]".
> What was the first one?
Apr 21 10:53:27 www adaptec-smbe: aiomgrd startup succeeded
Apr 21 10:53:27 www kernel: Unable to handle kernel NULL pointer dereference 
at virtual address 00000034
Apr 21 10:53:27 www kernel:  printing eip:
Apr 21 10:53:27 www kernel: c03227a0
Apr 21 10:53:27 www kernel: *pde = 36c2f001
Apr 21 10:53:27 www kernel: Oops: 0000 [#1]
Apr 21 10:53:27 www kernel: SMP
Apr 21 10:53:27 www kernel: Modules linked in: md5 ipv6 iptable_filter 
ip_tables tg3 floppy sg microcode ohci_hcd video button battery ac ext
3 jbd dm_mod aacraid aic79xx sd_mod scsi_mod
Apr 21 10:53:27 www kernel: CPU:    3
Apr 21 10:53:27 www kernel: EIP:    0060:[<c03227a0>]    Not tainted VLI
Apr 21 10:53:27 www kernel: EFLAGS: 00010086   (2.6.11-1.14_FC2.rootsmp)
Apr 21 10:53:27 www kernel: EIP is at _spin_lock_irqsave+0x10/0x60
Apr 21 10:53:27 www kernel: eax: 00000030   ebx: 00000286   ecx: bfff9b4c   
edx: bfff9b4c
Apr 21 10:53:27 www kernel: esi: 00000030   edi: 00000030   ebp: bfff9b4c   
esp: f5f52f14
Apr 21 10:53:27 www kernel: ds: 007b   es: 007b   ss: 0068
Apr 21 10:53:27 www kernel: Process aiomgrd (pid: 3003, threadinfo=f5f52000 
task=f7381020)
Apr 21 10:53:27 www kernel: Stack: 00000000 0000fe01 00042008 00000000 
f886ac4a 00042008 00000000 00000000
Apr 21 10:53:27 www kernel:        f8869b80 42676a06 00000000 42676a06 
00000000 42676a06 00042008 bfff9b4c
Apr 21 10:53:27 www kernel:        00000000 bfff9b4c f886a560 f8871d80 
f6c96880 ffffffe7 c017580a bfff9b4c
Apr 21 10:53:27 www kernel: Call Trace:
Apr 21 10:53:27 www kernel:  [<f886ac4a>] fib_alloc+0x1a/0x70 [aacraid]
Apr 21 10:53:27 www kernel:  [<f8869b80>] ioctl_send_fib+0x10/0x110 [aacraid]
Apr 21 10:53:27 www kernel:  [<f886a560>] aac_do_ioctl+0x60/0xcf [aacraid]
Apr 21 10:53:27 www kernel:  [<c017580a>] do_ioctl+0x6a/0x90
Apr 21 10:53:27 www kernel:  [<c0175a7e>] vfs_ioctl+0x5e/0x1d0
Apr 21 10:53:27 www kernel:  [<c0175c4b>] sys_ioctl+0x5b/0x90
Apr 21 10:53:27 www kernel:  [<c0104203>] syscall_call+0x7/0xb
Apr 21 10:53:27 www kernel: Code: af de 75 0c f0 83 28 01 79 05 e8 9c e4 ff ff 
c3 0f 0b d7 00 78 a7 33 c0 eb ea 90 83 ec 10 89 74 24 0c 89 c6
 89 5c 24 08 9c 5b fa <81> 78 04 ad 4e ad de 75 26 f0 fe 0e 79 13 f7 c3 00 02 
00 00 74


The same can be triggered by starting Adaptecs aaccli and giving the following 
command:

aaccli
CLI > controller list

Sincerly,
Klaus

-- 
Klaus Steinberger         Maier-Leibnitz Labor
Phone: (+49 89)289 14287  Am Coulombwall 6, D-85748 Garching, Germany
FAX:   (+49 89)289 14280  EMail: Klaus.Steinberger@Physik.Uni-Muenchen.DE
URL: http://www.physik.uni-muenchen.de/~k2/

In a world without Walls and Fences, who needs Windows and Gates

RE: Kernel OOPS with aacraid driver

[Yum Rayan]

> > BTW, this says that this is the second oops:  "[#2]".
> > What was the first one?
> Apr 21 10:53:27 www adaptec-smbe: aiomgrd startup succeeded
> Apr 21 10:53:27 www kernel: Unable to handle kernel NULL 
> pointer dereference 
> at virtual address 00000034

Seems to me like someone is issuing IOCTL's against devices that do
not exist. Does this patch against linux-2.6.12-rc3 make sense and
applicable to you?

Seems like aac_cfg_open() will return success even if it fails? This
patch fixes it to actually return error.

Regards,
Rayan

diff -Nupr -X dontdiff linux-2.6.12-rc3.a/drivers/scsi/aacraid/linit.c
linux-2.6.12-rc3.b/drivers/scsi/aacraid/linit.c
--- linux-2.6.12-rc3.a/drivers/scsi/aacraid/linit.c	2005-04-21
23:56:31.000000000 -0700
+++ linux-2.6.12-rc3.b/drivers/scsi/aacraid/linit.c	2005-04-21
23:57:26.000000000 -0700
@@ -450,7 +450,7 @@ static int aac_cfg_open(struct inode *in
 		}
 	}
 
-	return 0;
+	return err;
 }
 
 /**

Re: Kernel OOPS with aacraid driver

[Klaus Steinberger]

Hello Rayan,

> Seems to me like someone is issuing IOCTL's against devices that do
> not exist. Does this patch against linux-2.6.12-rc3 make sense and
> applicable to you?

Sounds good. I have an kernel build on the way, and will test it on Monday.

Sincerly,
Klaus

-- 
Klaus Steinberger         Maier-Leibnitz Labor
Phone: (+49 89)289 14287  Am Coulombwall 6, D-85748 Garching, Germany
FAX:   (+49 89)289 14280  EMail: Klaus.Steinberger@Physik.Uni-Muenchen.DE
URL: http://www.physik.uni-muenchen.de/~k2/

In a world without Walls and Fences, who needs Windows and Gates

RE: Kernel OOPS with aacraid driver

[Mark Haverkamp]

On Fri, 2005-04-22 at 00:08 -0700, Yum Rayan wrote:
> > > BTW, this says that this is the second oops:  "[#2]".
> > > What was the first one?
> > Apr 21 10:53:27 www adaptec-smbe: aiomgrd startup succeeded
> > Apr 21 10:53:27 www kernel: Unable to handle kernel NULL 
> > pointer dereference 
> > at virtual address 00000034
> 
> Seems to me like someone is issuing IOCTL's against devices that do
> not exist. Does this patch against linux-2.6.12-rc3 make sense and
> applicable to you?
> 
A fix for this was submitted to the list last week.

Mark.


-- 
Mark Haverkamp <markh@osdl.org>

Re: Kernel OOPS with aacraid driver

[Klaus Steinberger]

Am Freitag 22 April 2005 13:57 schrieb Klaus Steinberger:
> Hello Rayan,
>
> > Seems to me like someone is issuing IOCTL's against devices that do
> > not exist. Does this patch against linux-2.6.12-rc3 make sense and
> > applicable to you?
>
> Sounds good. I have an kernel build on the way, and will test it on Monday.

Yep I got by occasion a test window today, the fix works fine.

Sincerly,
Klaus

-- 
Klaus Steinberger         Maier-Leibnitz Labor
Phone: (+49 89)289 14287  Am Coulombwall 6, D-85748 Garching, Germany
FAX:   (+49 89)289 14280  EMail: Klaus.Steinberger@Physik.Uni-Muenchen.DE
URL: http://www.physik.uni-muenchen.de/~k2/

In a world without Walls and Fences, who needs Windows and Gates