Re: Regression: USB is nfg after suspend/resume(RAM) cycle on Intel chipset

[Mark Lord <lkml@rtr.ca>]

Mark Lord wrote:
> I just "upgraded" from 2.6.21.3 to 2.6.22-rc3,
> but will be rebooting back into 2.6.21.xx shortly.
> 
> Suspend/Resume (RAM) works perfectly on this Dell i9400 dual-core notebook.
> Except with 2.6.22-rc3, all USB devices are non-functional on resume.
> Most of the time.
> Once in five reboots, my mouse worked after the first suspend/resume cycle,
> but not thereafter.  'lsusb' also hangs after resume.
> 
> I'm attaching a full syslog with USB_DEBUG=y.
> I cannot use PM_DEBUG=y due to an unresolved race condition
> in the HRTIMER stuff (previously reported/discussed with no attempt at 
> resolution).
> 
> This notebook uses very mainstream Intel chips for most stuff:
..
> There are a zillion USB patches in 2.6.22-rc*.
> Greg: got any good suggestions on which one to revert first?

Okay, I used the "targeted revert" method instead of "git-bisect",
and nailed it in one try by reverting these two commits by Alan Stern:

7ed92f1a149dddc3cb537ccd7441e98adac12c3e USB: make the autosuspend workqueue thread freezable
ef7f6c7084b333c7524dcd297e0578d43733a2a2 USB: more autosuspend timer stuff

I yanked them both, as they appeared to be releated based on the titles.
Reverting this pair of commits fixes the USB suspend/resume regression.

Alan Stern:  your ball now.

-ml