Hi,
I am experiencing a repeatable crash with 2.6.9-rc2-mm1 on Fedora devel, when unplugging a uba device. I can quite reliably trigger this crash by simply pulling the hardware out of the USB port. Testing so far with both a USB digital camera, and also a USB pen stick, and I'm seeing the same trace:
------------[ cut here ]------------
kernel BUG at kernel/timer.c:414!
invalid operand: 0000 [#1]
PREEMPT SMP
Modules linked in: nls_iso8859_1 nls_cp437 vfat fat nls_base ub it87 eeprom i2cd
CPU: 0
EIP: 0060:[<c01206b2>] Not tainted VLI
EFLAGS: 00010007 (2.6.9-rc2-mm1)
EIP is at cascade+0x2f/0x39
eax: c140ab60 ebx: dccd5578 ecx: 0000000f edx: dccd5578
esi: c140b3e8 edi: c140ab60 ebp: 0000000f esp: c03b1f34
ds: 007b es: 007b ss: 0068
Process swapper (pid: 0, threadinfo=c03b0000 task=c032bb00)
Stack: 00000000 c03af188 c140ab60 c03b1f4c c0120c4b 00000247 c03b1f4c c03b1f4c
c032bb00 00000001 c03af188 c03decc0 00000000 c011d2b2 0000000a 00000046
00000000 c03dc488 00429007 c011d34a c03dc48c c010fc2e c03b0008 c03b0000
Call Trace:
[<c0120c4b>] run_timer_softirq+0x10f/0x180
[<c011d2b2>] __do_softirq+0x62/0xcd
[<c011d34a>] do_softirq+0x2d/0x35
[<c010fc2e>] smp_apic_timer_interrupt+0xf1/0xf5
[<c0104692>] apic_timer_interrupt+0x1a/0x20
[<c0101ed7>] mwait_idle+0x25/0x4a
[<c0101e8e>] cpu_idle+0x35/0x59
[<c03b27df>] start_kernel+0x163/0x17e
[<c03b2314>] unknown_bootoption+0x0/0x136
Code: 89 c7 56 8d 34 ca 53 8b 1e eb 10 39 7b 20 89 da 75 19 89 f8 8b 1b e8 71 f
I had this same crash with 2.6.9-rc1-mm4 and well, -mm5 was completely unstable on my box so I couldn't test that one at all.
I am also getting a lot of messages logged: "uba: was not changed"
..and in the case of my USB pen drive was getting some of these messages before the crash:
Buffer I/O error on device uba, logical block 3 end_request: I/O error, dev uba, sector 4 Buffer I/O error on device uba, logical block 2 end_request: I/O error, dev uba, sector 2 Buffer I/O error on device uba, logical block 1 end_request: I/O error, dev uba, sector 0 Buffer I/O error on device uba, logical block 0 end_request: I/O error, dev uba, sector 2
However with my USB digital camera, there weren't any messages before the crash.
I note that this patch is included in the source for rc2-mm1:
From: Pete Zaitcev <[EMAIL PROTECTED]> Patch for 3 ub bugs in 2.6.9-rc1-mm4 2004-09-11 20:25
Also noted this once in /var/log/messages:
uba: was not changed
usb_unlink_urb() is deprecated for synchronous unlinks. Use usb_kill_urb() instead
Badness in usb_unlink_urb at drivers/usb/core/urb.c:457
[<c025a07c>] usb_unlink_urb+0x6c/0x89
[<e0d5e8e3>] ub_urb_timeout+0x16/0x21 [ub]
[<e0d5e8cd>] ub_urb_timeout+0x0/0x21 [ub]
[<c0120bff>] run_timer_softirq+0xc3/0x180
[<c011d2b2>] __do_softirq+0x62/0xcd
[<c011d34a>] do_softirq+0x2d/0x35
[<c010fc2e>] smp_apic_timer_interrupt+0xf1/0xf5
[<c0104692>] apic_timer_interrupt+0x1a/0x20
[<c0101ed7>] mwait_idle+0x25/0x4a
[<c0101e8e>] cpu_idle+0x35/0x59
[<c03b27df>] start_kernel+0x163/0x17e
[<c03b2314>] unknown_bootoption+0x0/0x136
Not sure if this is the "usb_unlink_urb(&sc->work_urb);" statement on line 825 of ub.c.
System is a 2.8 Intel P4, SMP, HT, PREEMPT, compiled with gcc-3.4. (yep I know, probably all the bad things but still, I don't think the kernel should crash like this)
Reuben
------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement on who ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
