Bug 10836 - ELO serial touchscreen driver unload issue
Summary: ELO serial touchscreen driver unload issue
Status: CLOSED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Serial (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Dmitry Torokhov
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-05-30 01:26 UTC by Roland Kletzing
Modified: 2008-07-22 12:46 UTC (History)
0 users

See Also:
Kernel Version: 2.6.26-rc4-git2
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
Don't leave kseriod events behind (1.31 KB, patch)
2008-05-30 13:51 UTC, Dmitry Torokhov
Details | Diff

Description Roland Kletzing 2008-05-30 01:26:22 UTC
Latest working kernel version: n/a
Earliest failing kernel version:  2.6.26-rc4-git2
Distribution: opensuse 10.1
Hardware Environment: x86 / p4 /gigabyte mobo
Software Environment:
Problem Description:

elo driver seems to have issues when unloading - at least on system without elo touchscreen attached

[  119.319962] BUG: unable to handle kernel paging request at f8cabfcc
[  119.319962] IP: [<c0269171>] driver_attach+0x3/0x16
[  119.319962] *pde = 37e08067 *pte = 00000000
[  119.319962] Oops: 0000 [#1] SMP
[  119.319962] Modules linked in: psmouse vboxdrv ipv6 button battery ac loop dm _mod md_mod mousedev evdev rtc_cmos rtc_core rtc_lib 8250_pnp e100 mii ehci_hcd uhci_hcd intel_agp usbcore agpgart i2c_i801 rng_core i2c_core iTCO_wdt iTCO_vend or_support parport_pc lp parport 8250 serial_core edd fan thermal processor ther mal_sys hwmon sg megaraid_mbox sr_mod cdrom megaraid_mm ata_piix libata dock sd_ mod scsi_mod [last unloaded: elo]
[  119.319962]
[  119.319962] Pid: 121, comm: kseriod Not tainted (2.6.26-rc4-git2 #1)
[  119.319962] EIP: 0060:[<c0269171>] EFLAGS: 00010246 CPU: 0
[  119.319962] EIP is at driver_attach+0x3/0x16
[  119.319962] EAX: f8cabfc8 EBX: f56adec0 ECX: f8cabfc8 EDX: 00000246
[  119.319962] ESI: f78ec560 EDI: f8cabfa0 EBP: f7877fd0 ESP: f7877f84
[  119.319962]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
[  119.319962] Process kseriod (pid: 121, ti=f7877000 task=f786be80 task.ti=f787 7000)
[  119.319962] Stack: f7877fd0 c026d651 00000246 00000000 00000000 00000000 0000 0000 00000000
[  119.319962]        00000001 f7819f28 f7819f2c 00000000 f786be80 c013262d f787 7fbc f7877fbc
[  119.319962]        00000000 c026d4f5 00000000 f7877fe0 c0132559 c013251e 0000 0000 00000000
[  119.319962] Call Trace:
[  119.319962]  [<c026d651>] ? serio_thread+0x15c/0x2f3
[  119.319962]  [<c013262d>] ? autoremove_wake_function+0x0/0x33
[  119.319962]  [<c026d4f5>] ? serio_thread+0x0/0x2f3
[  119.319962]  [<c0132559>] ? kthread+0x3b/0x62
[  119.319962]  [<c013251e>] ? kthread+0x0/0x62
[  119.319962]  [<c0104443>] ? kernel_thread_helper+0x7/0x10
[  119.319962]  =======================
[  119.319962] Code: 5d c3 55 89 e5 56 8d b0 b8 00 00 00 53 89 c3 89 f0 e8 b1 c9  ec ff 89 d8 e8 6d ff ff ff 89 f0 e8 39 c9 ec ff 5b 5e 5d c3 55 89 c1 <8b> 40 04  89 e5 31 d2 68 70 93 26 c0 e8 23 fa ff ff c9 c3 55 89
[  119.319962] EIP: [<c0269171>] driver_attach+0x3/0x16 SS:ESP 0068:f7877f84
[  119.319962] ---[ end trace 9f13a3e20519fbdd ]---


Steps to reproduce:

while true;do modprobe elo; echo -n load;modprobe -r elo;echo -n unload;done
Comment 1 Dmitry Torokhov 2008-05-30 13:51:13 UTC
Created attachment 16344 [details]
Don't leave kseriod events behind

Could you please try this patch?
Comment 2 Roland Kletzing 2008-05-30 15:59:06 UTC
thanks!
it looks good.
At least i can tell that the issue is fixed.
cannot test with elo hardware attached, since i don`t own that device.

since i trust you know what you`re doing, i´m giving an

Acked-by: Roland Kletzing <devzero@web.de>

:)
Comment 3 Dmitry Torokhov 2008-06-01 22:28:07 UTC
OK, great, thank you for testing. Since the oops is not spontaneous but you really have to work to get to it the fix can wait and I will stage it for .27.
Comment 4 Roland Kletzing 2008-06-05 14:32:11 UTC
thanks very much. 

i`d happy to leave this open until merge - just to make sure it won`t get lost.....
Comment 5 Dmitry Torokhov 2008-07-22 12:46:24 UTC
The fix is in mainline (for 2.6.27), closing.

Note You need to log in before you can comment on or make changes to this bug.