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

[Mark Lord <lkml@rtr.ca>]

Mark Lord wrote:
> Linus Torvalds wrote:
..
>> There's a different (but related to all the same freezer problems) 
>> patch by Rafael Wysocki <rjw@sisk.pl>:
>>
>>     Re: [RFC][PATCH][EXPERIMENTAL] Make kernel threads nonfreezable by 
>> default
>>     Message-Id: <200705291415.31970.rjw@sisk.pl>
> 
> I'll hunt for Rafael's patch next.

Mmm.. Rafael's patch appears to be part of a large series of 15 patches,
and it's not totally clear to me how to test just that part of it,
so I think I'll leave things as is (working!) for now.

Here's my current fix:

-- snip --

Regarding commit 7ed92f1a149dddc3cb537ccd7441e98adac12c3e
"USB: make the autosuspend workqueue thread freezable":

This commit causes USB lockups on resume from Suspend-to-RAM
on my Core2duo notebook.  Applying the patch below to revert
the commit fixes the problem for me in 2.6.22-rc3.

Signed-off-by:  Mark Lord <mlord@pobox.com>
---
--- b0rken/drivers/usb/core/usb.c	2007-05-29 13:32:32.000000000 -0400
+++ linux/drivers/usb/core/usb.c	2007-05-29 13:27:10.000000000 -0400
@@ -205,11 +205,7 @@
 
 static int ksuspend_usb_init(void)
 {
-	/* This workqueue is supposed to be both freezable and
-	 * singlethreaded.  Its job doesn't justify running on more
-	 * than one CPU.
-	 */
-	ksuspend_usb_wq = create_freezeable_workqueue("ksuspend_usbd");
+	ksuspend_usb_wq = create_singlethread_workqueue("ksuspend_usbd");
 	if (!ksuspend_usb_wq)
 		return -ENOMEM;
 	return 0;