[Bug 378509] New: Machine hangs when using pppd with Kyocera CDMA card
https://bugzilla.novell.com/show_bug.cgi?id=378509 Summary: Machine hangs when using pppd with Kyocera CDMA card Product: openSUSE 11.0 Version: Factory Platform: x86-64 OS/Version: openSUSE 11.0 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: lchiquitto@novell.com QAContact: qa@suse.de Found By: Development I recently installed OpenSUSE on a x86-64 notebook and started to follow Factory. When using a Kyocera CDMA card to connect to the Internet, the machine will eventually hang. I couldn't find a way to reproduce it at will, but the issue seems to appear at least a couple of times a day. Traffic amount do not seem related to the hang. I will attach the call trace that was logged. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c1 --- Comment #1 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-04-09 13:19:44 MST --- Created an attachment (id=207105) --> (https://bugzilla.novell.com/attachment.cgi?id=207105) Logged kernel call trace -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User gregkh@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c2 Greg Kroah-Hartman <gregkh@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |INVALID --- Comment #2 from Greg Kroah-Hartman <gregkh@novell.com> 2008-04-09 14:17:05 MST --- This is a warning caused by your closed source kernel modules that you have loaded. If you can reproduce this without them loaded, please reopen this bug with the new kernel log messages. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c3 --- Comment #3 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-11 14:52:47 MST --- Finally was able to reproduce the issue and get a - hopefully now correct - call trace. BUG: spinlock recursion on CPU#1, swapper/0 lock: ffffffff8813a990, .magic: dead4ead, .owner: swapper/0, .owner_cpu: 1 Pid: 0, comm: swapper Tainted: G N 2.6.25-26-debug #1 Call Trace: [<ffffffff8020d686>] dump_trace+0xc4/0x576 [<ffffffff8020db78>] show_trace+0x40/0x57 [<ffffffff8046c564>] _etext+0x72/0x7b [<ffffffff80337a92>] spin_bug+0xa2/0xaa [<ffffffff80337b67>] _raw_spin_lock+0x42/0x133 [<ffffffff80465f90>] _spin_lock+0x9/0xb [<ffffffff88116eb7>] :usbcore:usb_hcd_unlink_urb_from_ep+0x18/0x43 [<ffffffff88154971>] :uhci_hcd:uhci_giveback_urb+0xf2/0x19d [<ffffffff881550d6>] :uhci_hcd:uhci_scan_schedule+0x5a4/0x85f [<ffffffff88156fa8>] :uhci_hcd:uhci_irq+0x13f/0x159 [<ffffffff88116d80>] :usbcore:usb_hcd_irq+0x2b/0x63 [<ffffffff80271810>] handle_IRQ_event+0x2e/0x65 [<ffffffff80272a29>] handle_fasteoi_irq+0x9d/0xdd [<ffffffff8020ebae>] do_IRQ+0x6e/0xd8 [<ffffffff8020c59d>] ret_from_intr+0x0/0x19 [<ffffffff88116ece>] :usbcore:usb_hcd_unlink_urb_from_ep+0x2f/0x43 [<ffffffff88448ab7>] :ohci_hcd:finish_urb+0x5e/0xb6 [<ffffffff88448b51>] :ohci_hcd:takeback_td+0x42/0xb5 [<ffffffff88448be4>] :ohci_hcd:dl_done_list+0x20/0x2d [<ffffffff8844aff2>] :ohci_hcd:ohci_irq+0x175/0x2c0 [<ffffffff88116d80>] :usbcore:usb_hcd_irq+0x2b/0x63 [<ffffffff80271810>] handle_IRQ_event+0x2e/0x65 [<ffffffff80272a29>] handle_fasteoi_irq+0x9d/0xdd [<ffffffff8020ebae>] do_IRQ+0x6e/0xd8 [<ffffffff8020c59d>] ret_from_intr+0x0/0x19 [<ffffffff802219d5>] native_irq_enable+0x6/0x7 [<ffffffff803d20f5>] cpuidle_idle_call+0x8d/0xc1 [<ffffffff8020b08a>] cpu_idle+0x92/0xda [<ffffffff80460c7f>] start_secondary+0x40e/0x41d I did some research and my best guess is that uhci-q.c:uhci_giveback_urb() misses calls to spin_lock_irqsave/restore(), to make sure IRQs are locally disabled when usb_hcd_unlink_urb_from_ep() is called. I wasn't able to verify if this is already always the case when uhci_giveback_urb() is called. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c4 Leonardo Chiquitto <lchiquitto@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Resolution|INVALID | --- Comment #4 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-11 14:54:05 MST --- I forgot to change the status when adding the last comment. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User gregkh@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c5 Greg Kroah-Hartman <gregkh@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |NEEDINFO Info Provider| |lchiquitto@novell.com --- Comment #5 from Greg Kroah-Hartman <gregkh@novell.com> 2008-05-12 09:27:42 MST --- odd thing is, this is an ohci device, not a uhci, device, right? Can you attach the output of 'hwinfo' to this bug so I can see what the usb topology looks like here? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c6 Leonardo Chiquitto <lchiquitto@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #207105|0 |1 is obsolete| | --- Comment #6 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-12 11:05:17 MST --- Created an attachment (id=214476) --> (https://bugzilla.novell.com/attachment.cgi?id=214476) Logged kernel call trace Correct call trace -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c7 Leonardo Chiquitto <lchiquitto@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |REOPENED Info Provider|lchiquitto@novell.com | --- Comment #7 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-12 11:09:19 MST --- Created an attachment (id=214477) --> (https://bugzilla.novell.com/attachment.cgi?id=214477) Hardware information Output of hwinfo -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c8 --- Comment #8 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-12 11:20:37 MST --- Yes, it's an OHCI device. Following the :ohci_hcd:* functions from the trace, I couldn't find where interruptions are being disabled. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 Greg Kroah-Hartman <gregkh@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gregkh@novell.com AssignedTo|kernel-maintainers@forge.provo.novell.com |oneukum@novell.com Status|REOPENED |NEW -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User oneukum@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c9 Oliver Neukum <oneukum@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |jkosina@novell.com --- Comment #9 from Oliver Neukum <oneukum@novell.com> 2008-05-17 01:40:50 MST --- Jiri, this looks like the spinlock issue you are chasing on lkml. Any idea what enables the interrupt? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c10 --- Comment #10 from Jiri Kosina <jkosina@novell.com> 2008-05-21 05:44:20 MST --- Created an attachment (id=217221) --> (https://bugzilla.novell.com/attachment.cgi?id=217221) fix hcd deadlock Could you please try whether the problem goes away if you use the kernel from http://labs.suse.cz/jikos/download/bug-378509/ It has the attached patch incorporated. I still don't see from the code how this bug could be causing the traces you are seeing, but it should fix a possible deadlock in HCD IRQ path, so is worth trying. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 Jiri Kosina <jkosina@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jkosina@novell.com Info Provider|jkosina@novell.com |lchiquitto@novell.com -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c11 --- Comment #11 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-21 06:00:41 MST --- Hi Jiri, I'm using a kernel with your patch applied for some days now, and the machine doesn't hang anymore. I also double checked the path from the call trace, and couldn't find where interrupts are being enabled. As your patch reverts the changes made by commit e9df41c5c58, don't you think it's worth raising the issue on LKML, or privately with the author, before adding the patch to the OpenSUSE kernel? Thanks, Leonardo -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c12 Leonardo Chiquitto <lchiquitto@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|lchiquitto@novell.com | --- Comment #12 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-21 06:03:37 MST --- Forgot to remove the NEEDINFO flag (again). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c13 --- Comment #13 from Jiri Kosina <jkosina@novell.com> 2008-05-21 06:04:52 MST --- (In reply to comment #11 from Leonardo Chiquitto)
I'm using a kernel with your patch applied for some days now, and the machine doesn't hang anymore.
Thanks a lot for testing!
I also double checked the path from the call trace, and couldn't find where interrupts are being enabled.
If my patch fixes the problem, then the call trace is bogus and the problem is elsewhere (the lock being used to synchronize between IRQ and non-IRQ contexts, without turning IRQs off, which could easily deadlock).
As your patch reverts the changes made by commit e9df41c5c58, don't you think it's worth raising the issue on LKML, or privately with the author, before adding the patch to the OpenSUSE kernel?
Well it doesn't revert the changes, it just fixes the locking that e9df41c5c58 introduced. I will proceed with the issue upstream. Thanks! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 Jiri Kosina <jkosina@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|oneukum@novell.com |jkosina@novell.com Status|NEW |ASSIGNED -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c14 Jiri Kosina <jkosina@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |lchiquitto@novell.com --- Comment #14 from Jiri Kosina <jkosina@novell.com> 2008-05-28 08:20:32 MDT --- Could you please try the kernel from http://labs.suse.cz/jikos/download/bug-378509/kernel-default-2.6.25.4-8.x86_... It doesn't contain the patch, but has lockdep lock correctness proving option enabled. The output from running this kernel on your system might give us some hint about what is wrong. Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c15 --- Comment #15 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-05-28 08:48:40 MDT --- Downloading it now. Doesn't the -debug kernels come with this option enabled? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c16 --- Comment #16 from Jiri Kosina <jkosina@novell.com> 2008-05-28 09:01:52 MDT --- (In reply to comment #15 from Leonardo Chiquitto)
Downloading it now. Doesn't the -debug kernels come with this option enabled?
Unfortunately they don't, and that's something I have been planning to discuss inside the kernel teams for quite some time already :) Thanks. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c17 --- Comment #17 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-06-04 07:14:04 MDT --- Just to give some feedback on this issue: the machine refuses to hang when using the kernel from comment #14. This is probably Murphy, so I will keep using it for more one week. If it still doesn't hang, I'll revert to default kernel from Factory and see if the problem persists. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c18 --- Comment #18 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-06-30 13:09:22 MDT --- Jiri, I'm sorry for not providing more input on this. I know you follow usb-devel@, but I'd like to register here the following thread that I'm just reading: http://marc.info/?l=linux-kernel&m=121415383913620&w=2 Thanks -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c19 --- Comment #19 from Jiri Kosina <jkosina@novell.com> 2008-07-01 08:31:22 MDT --- For reference: upstream patch that shoudl fix the issue has been posted on http://lkml.org/lkml/2008/6/30/462 Let's wait for upstream testing and merging before we take it into our kernel. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User lchiquitto@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c20 Leonardo Chiquitto <lchiquitto@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|lchiquitto@novell.com | --- Comment #20 from Leonardo Chiquitto <lchiquitto@novell.com> 2008-07-04 06:59:21 MDT --- I just saw that the fix was committed to Linus' tree (de85422b94ddb23c021126815ea49414047c13dc). Jiri, thanks for the help with this ticket. Leonardo -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=378509 User jkosina@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=378509#c21 Jiri Kosina <jkosina@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #21 from Jiri Kosina <jkosina@novell.com> 2008-09-01 03:24:11 MDT --- Closing bug, thanks a lot for testing. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com