crash after resume with USB mouse attached

[Martin Pitt]

With kernel 2.6.12, when I resume from STR, I get a kernel crash if I leave my
external USB mouse attached. That didn't happen with the Hoary kernel (well,
that didn't properly reinitialize the USB module so that you had to reload
[eo]hci_hcd, but at least it didn't crash).

Platform: Apple iBook G4, [EO]HCI

Jul 1 00:07:40 localhost kernel: [ 5084.247726] PCI: Enabling device
0001:10:1b.0 (0000 -> 0002)
Jul 1 00:07:40 localhost kernel: [ 5084.308705] PCI: Enabling device
0001:10:1b.1 (0000 -> 0002)
Jul 1 00:07:40 localhost kernel: [ 5084.370710] PCI: Enabling device
0001:10:1b.2 (0000 -> 0002)
Jul 1 00:07:40 localhost kernel: [ 5084.391710] ehci_hcd 0001:10:1b.2: park 0
Jul 1 00:07:40 localhost kernel: [ 5084.392192] ehci_hcd 0001:10:1b.2: USB 2.0
restarted, EHCI 1.00, driver 10 Dec 2004
Jul 1 00:07:40 localhost kernel: [ 5084.599755] usb 2-1: USB disconnect, address 2
Jul 1 00:07:40 localhost kernel: [ 5085.425740] eth0: resuming
Jul 1 00:07:40 localhost kernel: [ 5085.488015] PHY ID: 4061e4, addr: 0
Jul 1 00:07:40 localhost kernel: [ 5086.502814] hda: Enabling Ultra DMA 5
Jul 1 00:07:40 localhost kernel: [ 5086.506374] hdc: Enabling MultiWord DMA 2
Jul 1 00:07:40 localhost kernel: [ 5086.508007] Oops: kernel access of bad
area, sig: 11 [#1]
Jul 1 00:07:40 localhost kernel: [ 5086.508652] NIP: D2746EE4 LR: D24A14FC SP:
CB645D40 REGS: cb645c90 TRAP: 0300 Not tainted
Jul 1 00:07:40 localhost kernel: [ 5086.509640] MSR: 0200b032 EE: 1 PR: 0 FP: 1
ME: 1 IR/DR: 11
Jul 1 00:07:40 localhost kernel: [ 5086.510295] DAR: 000018A8, DSISR: 40000000
Jul 1 00:07:40 localhost kernel: [ 5086.510781] TASK = cfe48090[3520]
'pbbuttonsd' THREAD: cb644000
Jul 1 00:07:40 localhost kernel: [ 5086.511476] Last syscall: 54
Jul 1 00:07:40 localhost kernel: [ 5086.511830] GPR00: 00000000 CB645D40
CFE48090 00000000 00000010 C7C263D0 C02ED95C CA1760B8
Jul 1 00:07:40 localhost kernel: [ 5086.513073] GPR08: 00000000 D24C1618
D2750CC8 00000000 22000288 10026C18 100C0000 100CB618
Jul 1 00:07:40 localhost kernel: [ 5086.514317] GPR16: 100C0000 100A0000
00000000 100C0000 100D5408 10010000 C02E0000 10010000
Jul 1 00:07:40 localhost kernel: [ 5086.515560] GPR24: C0270000 C028AE28
C0290000 C0290000 C028AE30 C7C2640C C026A4A4 C7C26378
Jul 1 00:07:40 localhost kernel: [ 5086.516832] NIP [d2746ee4]
hid_resume+0x20/0x54 [usbhid]
Jul 1 00:07:40 localhost kernel: [ 5086.517510] LR [d24a14fc]
usb_generic_resume+0x78/0x9c [usbcore]
Jul 1 00:07:40 localhost kernel: [ 5086.518331] Call trace:
Jul 1 00:07:40 localhost kernel: [ 5086.518651] [d24a14fc]
usb_generic_resume+0x78/0x9c [usbcore]
Jul 1 00:07:40 localhost kernel: [ 5086.519421] [c0139b04] resume_device+0xa4/0xc0
Jul 1 00:07:40 localhost kernel: [ 5086.520026] [c0139c14] dpm_resume+0xf4/0x184
Jul 1 00:07:40 localhost kernel: [ 5086.520600] [c0139cdc] device_resume+0x38/0xa4
Jul 1 00:07:40 localhost kernel: [ 5086.521192] [c02c7118]
pmac_wakeup_devices+0xc0/0xe0
Jul 1 00:07:40 localhost kernel: [ 5086.521857] [c02c8a98] pmu_ioctl+0x8d4/0x944
Jul 1 00:07:40 localhost kernel: [ 5086.522430] [c007b5cc] do_ioctl+0x68/0x9c
Jul 1 00:07:40 localhost kernel: [ 5086.522980] [c007b6b8] vfs_ioctl+0xb8/0x400
Jul 1 00:07:40 localhost kernel: [ 5086.523543] [c007ba40] sys_ioctl+0x40/0x74
Jul 1 00:07:40 localhost kernel: [ 5086.524096] [c00047f0]
ret_from_syscall+0x0/0x4c
Jul 1 00:07:40 localhost kernel: [ 5087.289729] usb 2-1: new low speed USB
device using ohci_hcd and address 3
Jul 1 00:07:41 localhost kernel: [ 5087.600050] input: USB HID v1.00 Mouse
[1241:1177] on usb-0001:10:1b.0-1

[Matthew Garrett (mjg59)]

Can you try with the latest acpi-support and acpid packages?

[Benjamin Herrenschmidt (benh-kernel)]

Usual USB crap. ACPI stuff won't help. Go to linux-usb-devel and ask them why
they can't get it working :)

It might help to enable CONFIG_USB_SUSPEND or disable it if it's enabled. This
specific bug has been
reported for MONTHES and they still haven't fixed it. It's a quite nice NULL
dereference though,
it shouldn't be too hard.

[Matthew Garrett (mjg59)]

Martin: ping?

[Martin Pitt (pitti)]

(In reply to comment #3)
> Martin: ping?

Oh, sorry, I completely ignored your question since ACPI is so non-powerpc-ish...

[Martin Pitt (pitti)]

I just suspended and resumed with the USB mouse attached two times, it worked
fine, so I'm closing this.

(Of course I did not install any acpi stuff, these packages are !powerpc).

[Benjamin Herrenschmidt (benh-kernel)]

Might be a bit quick to close that one, this is a recurring issue as far as I'm
concerned (that is I'm getting reports regularty). There seem to be a race in
the USB input driver vs. suspend/resume, though it's not reproduceable on all
platforms.

[Martin Pitt (pitti)]

(In reply to comment #6)
> Might be a bit quick to close that one, this is a recurring issue as far as I'm
> concerned (that is I'm getting reports regularty). There seem to be a race in
> the USB input driver vs. suspend/resume, though it's not reproduceable on all
> platforms.

I did not see any dups and other reporters here, so I assumed I could close it.
However, we can always reopen it if somebody encounters it again.

[John Steele Scott (toojays)]

Please reopen, I think just hit this tonight. However, to really exacerbate the
situation, I have a USB mouse attached via a USB hub.

Unfortunately my brain wasn't working fast enough to take a photo of the stack
trace, but I did write down usb_generic_resume before I rebooted. This bug is
consistent with what I reported to linux-usb-devel back in May:
http://thread.gmane.org/gmane.linux.usb.devel/34760

Back then I found that compiling with CONFIG_USB_SUSPEND set fixed the problem
for me. I haven't yet tried this with the Ubuntu kernel. I've just upgraded to
breezy and I was hoping you would've already taken care of all the nasty details
for me ;)

[Martin Pitt (pitti)]

(In reply to comment #8)
> Please reopen, I think just hit this tonight. However, to really exacerbate the
> situation, I have a USB mouse attached via a USB hub.

[Ben Collins (ben-collins)]

If possible, please upgrade to Dapper's 2.6.15-8 kernel. If you do not want to
upgrade to Dapper, then you can also wait for the Dapper Flight 2 CD's, which
are due out within the next few days.

Let me know if this bug still exists with this kernel.

[John Steele Scott (toojays)]

I've upgraded to Dapper a few weeks ago, and this bug seems to be fixed now.

[Chuck Short (zulcss)]

Great closing