* ide-scsi segfaults under 2.5.44
@ 2002-10-20 3:39 Thomas Molina
2002-10-20 18:18 ` Mike Anderson
0 siblings, 1 reply; 3+ messages in thread
From: Thomas Molina @ 2002-10-20 3:39 UTC (permalink / raw)
To: linux-kernel
I am getting an oops when attempting to remove ide-scsi modules:
Oct 19 22:10:30 dad kernel: Unable to handle kernel paging request at
virtual address d909b03c
Oct 19 22:10:30 dad kernel: printing eip:
Oct 19 22:10:30 dad kernel: c01d0665
Oct 19 22:10:30 dad kernel: *pde = 17d2d067
Oct 19 22:10:30 dad kernel: *pte = 00000000
Oct 19 22:10:30 dad kernel: Oops: 0000
Oct 19 22:10:30 dad kernel: sr_mod cdrom sg ymfpci ac97_codec soundcore
parport_pc lp parport scsi_mod rtc
Oct 19 22:10:30 dad kernel: CPU: 0
Oct 19 22:10:30 dad kernel: EIP: 0060:[<c01d0665>] Not tainted
Oct 19 22:10:30 dad kernel: EFLAGS: 00010202
Oct 19 22:10:30 dad kernel: EIP is at driver_detach+0x35/0x40
Oct 19 22:10:30 dad kernel: eax: d9097050 ebx: d905f784 ecx: d9099000
edx: d909b03c
Oct 19 22:10:30 dad kernel: esi: d905f780 edi: d9097028 ebp: d5cf0000
esp: d5d0bf5c
Oct 19 22:10:30 dad kernel: ds: 0068 es: 0068 ss: 0068
Oct 19 22:10:30 dad kernel: Process rmmod (pid: 953, threadinfo=d5d0a000
task=d5cee080)
Oct 19 22:10:30 dad kernel: Stack: d905f784 d905f780 d9097028 c01d0857
d9097028 d9097028 d905f780 00000000
Oct 19 22:10:30 dad kernel: c01d0b78 d9097028 d9090000 00000000
d90928ac d9097028 c0119e1e d9090000
Oct 19 22:10:30 dad kernel: 00000000 fffffff0 c01192f7 d9090000
00000000 d5d0a000 bffffcb5 00000001
Oct 19 22:10:30 dad kernel: Call Trace:
Oct 19 22:10:30 dad kernel: [<d905f784>]
scsi_driverfs_bus_type_Rfc6baff9+0x4/0x80 [scsi_mod]
Oct 19 22:10:30 dad kernel: [<d905f780>]
scsi_driverfs_bus_type_Rfc6baff9+0x0/0x80 [scsi_mod]
Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
Oct 19 22:10:30 dad kernel: [<c01d0857>] bus_remove_driver+0x37/0x70
Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
Oct 19 22:10:30 dad kernel: [<d905f780>]
scsi_driverfs_bus_type_Rfc6baff9+0x0/0x80 [scsi_mod]
Oct 19 22:10:30 dad kernel: [<c01d0b78>] put_driver+0x28/0x50
Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
Oct 19 22:10:30 dad kernel: [<d90928ac>] exit_sg+0x4c/0x50 [sg]
Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
Oct 19 22:10:30 dad kernel: [<c0119e1e>] free_module+0x1e/0xb0
Oct 19 22:10:30 dad kernel: [<c01192f7>] sys_delete_module+0xe7/0x1c0
Oct 19 22:10:30 dad kernel: [<c0108adf>] syscall_call+0x7/0xb
Oct 19 22:10:30 dad kernel:
Oct 19 22:10:30 dad kernel: Code: 8b 32 75 d7 5b 5e 5f c3 8d 76 00 57 56
53 8b 7c 24 10 ff b7
The problem does seem to be order dependent. I haven't had a chance to
fully explore this because I have to reboot each time to get the system
into a sane state. However, it appears that if I remove the sg module
first, I get the above oops. If I remove the ide-scsi module first I get
the following:
Oct 19 22:09:34 dad kernel: Badness in put_device at
drivers/base/core.c:139
Oct 19 22:09:34 dad kernel: Call Trace:
Oct 19 22:09:34 dad kernel: [<c01cfbfc>] put_device+0x9c/0xc0
Oct 19 22:09:34 dad kernel: [<d90927e8>] sg_detach+0x1a8/0x1d0 [sg]
Oct 19 22:09:34 dad kernel: [<d90970fc>] dev_attr_kdev+0x0/0x10 [sg]
Oct 19 22:09:34 dad kernel: [<d909710c>] dev_attr_type+0x0/0x10 [sg]
Oct 19 22:09:34 dad kernel: [<d90a8e55>] sr_detach+0x45/0x80 [sr_mod]
Oct 19 22:09:34 dad kernel: [<d9096fe0>] sg_template+0x0/0xa0 [sg]
Oct 19 22:09:34 dad kernel: [<d90541ba>]
scsi_host_chk_and_release+0x13a/0x200 [scsi_mod]
Oct 19 22:09:34 dad kernel: [<d905f810>] scsi_host_list+0x0/0x8
[scsi_mod]
Oct 19 22:09:34 dad kernel: [<d9054080>]
scsi_host_chk_and_release+0x0/0x200 [scsi_mod]
Oct 19 22:09:34 dad kernel: [<d90632c0>] idescsi_template+0x0/0x80
[ide-scsi]
Oct 19 22:09:34 dad kernel: [<d905401a>] scsi_tp_for_each_host+0x1a/0x30
[scsi_mod]
Oct 19 22:09:34 dad kernel: [<d90632c0>] idescsi_template+0x0/0x80
[ide-scsi]
Oct 19 22:09:34 dad kernel: [<d9054a07>]
scsi_unregister_host_R08850fd9+0x17/0x70 [scsi_mod]
Oct 19 22:09:34 dad kernel: [<d90632c0>] idescsi_template+0x0/0x80
[ide-scsi]
Oct 19 22:09:34 dad kernel: [<d9054080>]
scsi_host_chk_and_release+0x0/0x200 [scsi_mod]
Oct 19 22:09:34 dad kernel: [<d906251a>] exit_idescsi_module+0xa/0x20
[ide-scsi]
Oct 19 22:09:34 dad kernel: [<d90632c0>] idescsi_template+0x0/0x80
[ide-scsi]
Oct 19 22:09:34 dad kernel: [<c0119e1e>] free_module+0x1e/0xb0
Oct 19 22:09:34 dad kernel: [<c01192f7>] sys_delete_module+0xe7/0x1c0
Oct 19 22:09:34 dad kernel: [<c0108adf>] syscall_call+0x7/0xb
Oct 19 22:09:34 dad kernel:
In either case I get a segfault on shutdown, and a lockup. I've not seen
any filesystem corruption and the journal restores things in an orderly
fashion on reboot.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: ide-scsi segfaults under 2.5.44
2002-10-20 3:39 ide-scsi segfaults under 2.5.44 Thomas Molina
@ 2002-10-20 18:18 ` Mike Anderson
2002-10-20 20:53 ` Thomas Molina
0 siblings, 1 reply; 3+ messages in thread
From: Mike Anderson @ 2002-10-20 18:18 UTC (permalink / raw)
To: Thomas Molina; +Cc: linux-kernel
Thomas,
It looks like sg.c was missed in the update from put_device to
device_unregister.
I have only compile tested the patch below, but it should fix the
problem. I will have a chance to run later on today.
Thomas Molina [tmolina@cox.net] wrote:
> Oct 19 22:10:30 dad kernel: [<d905f780>]
> scsi_driverfs_bus_type_Rfc6baff9+0x0/0x80 [scsi_mod]
> Oct 19 22:10:30 dad kernel: [<c01d0b78>] put_driver+0x28/0x50
> Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
> Oct 19 22:10:30 dad kernel: [<d90928ac>] exit_sg+0x4c/0x50 [sg]
> Oct 19 22:10:30 dad kernel: [<d9097028>] sg_template+0x48/0xa0 [sg]
> Oct 19 22:10:30 dad kernel: [<c0119e1e>] free_module+0x1e/0xb0
> Oct 19 22:10:30 dad kernel: [<c01192f7>] sys_delete_module+0xe7/0x1c0
> Oct 19 22:10:30 dad kernel: [<c0108adf>] syscall_call+0x7/0xb
> Oct 19 22:10:30 dad kernel:
-andmike
--
Michael Anderson
andmike@us.ibm.com
sg.c | 2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
------
===== drivers/scsi/sg.c 1.30 vs edited =====
--- 1.30/drivers/scsi/sg.c Fri Oct 18 11:27:30 2002
+++ edited/drivers/scsi/sg.c Sun Oct 20 10:59:33 2002
@@ -1607,7 +1607,7 @@
sdp->de = NULL;
device_remove_file(&sdp->sg_driverfs_dev, &dev_attr_type);
device_remove_file(&sdp->sg_driverfs_dev, &dev_attr_kdev);
- put_device(&sdp->sg_driverfs_dev);
+ device_unregister(&sdp->sg_driverfs_dev);
if (NULL == sdp->headfp)
vfree((char *) sdp);
}
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: ide-scsi segfaults under 2.5.44
2002-10-20 18:18 ` Mike Anderson
@ 2002-10-20 20:53 ` Thomas Molina
0 siblings, 0 replies; 3+ messages in thread
From: Thomas Molina @ 2002-10-20 20:53 UTC (permalink / raw)
To: Mike Anderson; +Cc: linux-kernel
On Sun, 20 Oct 2002, Mike Anderson wrote:
> Thomas,
> It looks like sg.c was missed in the update from put_device to
> device_unregister.
>
> I have only compile tested the patch below, but it should fix the
> problem. I will have a chance to run later on today.
OK, well that was interesting. Your patch cleared up my problem with sg.
Now I have a new ide-scsi failure mode. I'm pounding on this one because
whatever magic was supposed to happen with ide-cd didn't. I've documented
it in a separate message.
I can now insert and remove ide-scsi modules (sr_mod, sg, scsi_mod, cdrom,
ide-scsi). However, when I then insert the ide-scsi modules a second
time, it sees my disks at a different device! I have a CDROM at ide1
primary, and a CDRW at ide1 secondary. ide-scsi normally sees them as
/dev/scd0 and /dev/scd1, respectively. After the modules are reinserted I
get them reversed. ide1 primary is now seen at /dev/scd1 and ide1
secondary is seen at /dev/scd0. Furthermore, this happens each time I
repeat the cycle. On the next cycle they are back to ide1 primary at
/dev/scd0. In other words, they alternate, depending on how many
remove/insert cycles I go through.
If I simply insert the modules and leave them, thing act normally.
However, once I remove them I get a segfault and system hang at reboot.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2002-10-20 20:47 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-10-20 3:39 ide-scsi segfaults under 2.5.44 Thomas Molina
2002-10-20 18:18 ` Mike Anderson
2002-10-20 20:53 ` Thomas Molina
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome