Bug 23312

Summary:	asix usb network driver non-functional in 2.6.36, and breaks system suspend
Product:	Power Management	Reporter:	Armando Di Cianno (armando)
Component:	Hibernation/Suspend	Assignee:	power-management_other
Status:	CLOSED DUPLICATE
Severity:	high	CC:	rjw
Priority:	P1
Hardware:	All
OS:	Linux
Kernel Version:	2.6.36	Subsystem:
Regression:	No	Bisected commit-id:

Description Armando Di Cianno 2010-11-19 21:02:17 UTC

I'm running a vanilla 2.6.36 kernel on a Gentoo system. I've adding a small number of patches (mainly PCI ID additions) to support hardware on a MacBook Air 11'' (MacBookAir3,2).

In Gentoo's 2.6.35-gentoo-r12, I was loading ehci_hcd and ohci_hcd as modules (automatically from the system initramfs), and everything was working as expected, concerning the asix driver. Upon upgrading the 2.6.36, USB wasn't really working, but I read the release notes and compiled ehci_hcd and ohci_hcd into the kernel, so that I wouldn't have a load order issue.

Other devices (that I've tested) seem to work with 2.6.36. The asix module is loaded at boot time, and suspend works. Once I plug in my usb ethernet adaptor (Apple brand), I see the following in dmesg:

[ 1413.936568] usb 2-2: new high speed USB device using ehci_hcd and address 2
[ 1414.178721] usb 2-2: unable to read config index 0 descriptor/start: -71
[ 1414.178725] usb 2-2: chopping to 0 config(s)
[ 1414.242694] usb 2-2: string descriptor 0 read error: -71
[ 1414.242701] usb 2-2: New USB device found, idVendor=05ac, idProduct=1402
[ 1414.242705] usb 2-2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 1414.242842] usb 2-2: no configuration chosen from 0 choices

No extra data outputs on disconnect.

 * The driver no longer creates eth0
 * The system can no longer suspend


-----

$ uname -a
Linux the-hint 2.6.36-gentoo-r1-adc0 #12 SMP PREEMPT Fri Nov 19 10:59:05 EST 2010 x86_64 Intel(R) Core(TM)2 Duo CPU U9600 @ 1.60GHz GenuineIntel GNU/Linux

Comment 1 Armando Di Cianno 2010-11-19 21:11:25 UTC

dmesg output on suspend attempt that fails:


[ 1939.804951] PM: Syncing filesystems ... done.
[ 1942.760146] CE: hpet increased min_delta_ns to 16875 nsec
[ 1942.760278] CE: hpet increased min_delta_ns to 25312 nsec
[ 1943.150693] Freezing user space processes ... (elapsed 0.01 seconds) done.
[ 1943.169528] Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
[ 1943.179906] Suspending console(s) (use no_console_suspend to debug)
[ 1943.180384] pm_op(): usb_dev_suspend+0x0/0x10 returns -32
[ 1943.180388] PM: Device 2-2 failed to suspend async: error -32
[ 1943.180792] PM: Some devices failed to suspend
[ 1943.181280] PM: resume of devices complete after 0.483 msecs
[ 1943.181436] Restarting tasks ... done.
[ 1943.182370] video LNXVIDEO:00: Restoring backlight state


NOTE: The usb ethernet adaptor is unplugged at this point.

Comment 2 Armando Di Cianno 2010-11-19 21:49:27 UTC

Please note that the driver also does NOT working in 2.6.36. The power/suspend issue is more grave, but the driver not working is an issue as well.

If there's anything I can do to test, please just let me know! Cheers.

Comment 3 Armando Di Cianno 2010-11-24 16:38:43 UTC

I've also tested 2.6.36.1

Two further points:

1) All devices that were working in 2.6.35 now do not work -- hard drives, this asix usb adaptor, etc

2) The other devices, like an HD, fail, but do not break sleep; asix fails and breaks sleep

**** So this bug is definitely not only a power issue but a USB issue. Should I notify the USB maintainers as well, or create or a new bug and leave this one be?

Comment 4 Armando Di Cianno 2010-11-27 06:01:30 UTC

The latest patch in ticket #21182 fixes this issue. USB devices once again work. The asix module recognizes the hardware, and no longer breaks suspend.

If the patch in that ticket has made it into mainline, then this can be considered closed.

Comment 5 Rafael J. Wysocki 2010-11-27 20:17:48 UTC


*** This bug has been marked as a duplicate of bug 21182 ***

Comment 6 Rafael J. Wysocki 2010-11-27 20:18:27 UTC

Please use bug #21182 for tracking this issue from now on.