2.4.22-pre2

Loading and using the pl2303 serial module works fine the first time, but when you close the port the kernel goes oops. Unplugging the adapter makes life much worse - unfortunately, I can't find the message in my logs :( but it streams an error about usb interrupts. (I can reproduce if needed).

Its dead simple to reproduce - plug a pl2303 in, open /dev/ttyUSB0 with (eg) minicom or pppd, close it, check logs.

System seems stable after the crash. Kernel has swsusp but this oops is from a clean boot, before suspending/resuming.

This kernel has preempt and updated acpi; if that is a potential issue I can build a stock -pre3 and try to reproduce.

ksymoops 2.4.8 on i686 2.4.22-pre2-dis5. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.4.22-pre2-dis5/ (default) -m /boot/System.map-2.4.22-pre2-dis5 (default)

Warning: You did not tell me where to find symbol information. I will assume that the log matches the kernel and modules that are running right now and I'll use the default options above for symbol resolution. If the current kernel and/or modules do not match the log, you can get more accurate output by telling me the kernel version and where to find map, modules, ksyms etc. ksymoops -h explains the options.

Unable to handle kernel NULL pointer dereference at virtual address 00000014
e4d19794
*pde = 00000000
Oops: 0000
CPU:    0
EIP:    0010:[<e4d19794>]    Not tainted

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00210087
eax: d9ee8c00 ebx: ded0d880 ecx: 00000004 edx: 00000000
esi: d8b8cb80 edi: ded0d880 ebp: 00200292 esp: dc85be80
ds: 0018 es: 0018 ss: 0018
Process minicom (pid: 20629, stackpage=dc85b000)

Stack: ded0d880 d8b8cb80 00200282 ded0d884 e4d1a51b ded0d880 ded0d880 00000000
       c3d5d270 00000000 d7176c1c d7176c74 00000000 e28bd5fd ded0d880 e4d2846f
       ded0d880 e4d2933d 00000001 d7176c1c d7176c1c e29265a7 d7176c1c d6d33f00
Call Trace:    [<e4d1a51b>] [<e28bd5fd>] [<e4d2846f>] [<e4d2933d>] [<e29265a7>]
  [<e2926665>] [<c01c221e>] [<c0123283>] [<c014353a>] [<c0163130>] [<c01631cb>]
  [<c01c261f>] [<c0139f6c>] [<c01385bd>] [<c013863e>] [<c010915f>]
Code: 8b 52 14 83 ea 1c 8b 42 04 8b 5a 08 25 00 00 00 2f 0d 00 00

Trace; e4d1a51b <[uhci]uhci_unlink_urb+14b/190> Trace; e28bd5fd <[usbcore]usb_unlink_urb+3d/40> Trace; e4d2846f <[pl2303]pl2303_close+df/1c0> Trace; e4d2933d <[pl2303]__module_parm_desc_debug+25/e68> Trace; e29265a7 <[usbserial]__serial_close+87/b0> Trace; e2926665 <[usbserial]serial_close+95/c0> Trace; c01c221e <release_dev+55e/5a0>
Trace; c0123283 <in_group_p+23/30>
Trace; c014353a <vfs_permission+8a/130>
Trace; c0163130 <ext3_lookup+0/b0>
Trace; c01631cb <ext3_lookup+9b/b0>
Trace; c01c261f <tty_release+f/20>
Trace; c0139f6c <fput+fc/120>
Trace; c01385bd <filp_close+4d/80>
Trace; c013863e <sys_close+4e/60>
Trace; c010915f <system_call+33/38>

Code; e4d19794 <[uhci]uhci_reset_interrupt+24/a0> 00000000 <_EIP>:
Code; e4d19794 <[uhci]uhci_reset_interrupt+24/a0> <=====
0: 8b 52 14 mov 0x14(%edx),%edx <=====
Code; e4d19797 <[uhci]uhci_reset_interrupt+27/a0>
3: 83 ea 1c sub $0x1c,%edx
Code; e4d1979a <[uhci]uhci_reset_interrupt+2a/a0>
6: 8b 42 04 mov 0x4(%edx),%eax
Code; e4d1979d <[uhci]uhci_reset_interrupt+2d/a0>
9: 8b 5a 08 mov 0x8(%edx),%ebx
Code; e4d197a0 <[uhci]uhci_reset_interrupt+30/a0>
c: 25 00 00 00 2f and $0x2f000000,%eax
Code; e4d197a5 <[uhci]uhci_reset_interrupt+35/a0>
11: 0d 00 00 00 00 or $0x0,%eax

1 warning issued. Results may not be reliable.

Disconnect <[email protected]>

Known bug, search the archives :)

It's fixed in 2.5.

Someone needs to backport the 2.5 changes to 2.4 to fix this issue, as I'm getting tired of repeating myself...

thanks,

greg k-h

I've said it before on this list, and on the linux-usb-devel list, but here it is again for people to refer to in the future.

The locking and reference counting logic changes in the usb-serial drivers needs to be backported to the 2.4 kernel to fix this problem. The changes were spread out over a wide range of time during the 2.5 development process, so I can't point to a single changeset, sorry. It also requires changing all usb-serial drivers to do this properly.

If you do a diff of the two trees, you should see the major changes.

Good luck,

greg k-h