2.6.24 and suspend

Live forum: http://rt2x00.serialmonkey.com/viewtopic.php?t=4583


26-01-2008 00:26:07

Hi all!

Has anyone tried to compile 2.6.24 with RT2x00 and RT2500PCI enabled?
If yes - does suspend work for you?
For me it doesn't
Actually it never did, I've been trying for almost two years now )


26-01-2008 00:29:42

And could you be more specific about the problem... Things like "it doesn't work" are hardly descriptive.

Is it related to http//bugzilla.kernel.org/show_bug.cgi?id=9585 ?

And does applying the patches posted in there help for you?


26-01-2008 00:47:19

I've posted my issue some time ago and it didn't change since then

It doesn't relate to kernel since setting RT2500PCI to N resolves it.

I will try to produce an updated debug output from RT kernel module.
Will setting RT2x00_DEBUG to Y produce it?


26-01-2008 01:01:54

Debug just taken


26-01-2008 01:07:24

FYI two ACPI error messages are present even with rt2500pci module not loaded and they don't prevent the suspend operation


26-01-2008 10:33:05

Patches from the kernel bugzilla don't change anything.
They produce exactly the same debug output.


26-01-2008 13:00:10

Apparently your bug is completly different then the one from bugzilla.
Please try attached patch.


26-01-2008 17:05:05


Thanks for the patch!
Now my laptop suspends with rt2500pci loaded. But there are problems with resume now.

When I do a clean start of my laptop with no NetworkManager loaded and no other s/w that might use wifi module suspend and resume works just great.

But once I do any operation with rt2500pci, even the simplest one - rmmod followed by modprobe, it still suspends but never resumes - laptop hangs and the caps lock led flashes.

This is not just about module unloading/loading, any other kind of operation attempt leads to resume failures. This includes NetworkManager startup, or trying to bring wlan up with manual NetworkConfiguration.

I'm attaching both successful suspend/resume and failed debug outputs, although the latter doesn't tell much.



29-03-2008 11:33:33

I see that issue http//bugzilla.kernel.org/show_bug.cgi?id=9585 is almost fixed now.
Question is there any way I could test updated rt2x00 with 2.6.24 kernel, or do I have to get rc of 2.6.25?


07-04-2008 08:12:33

I downloaded and built 2.6.25-rc8
It fails to suspend with this message

phy0 -> rt2x00lib_suspend Notice - Going to sleep.
phy0 -> rt2500pci_set_state Notice - Device failed to enter state 1, current device state bbp 3 and rf 3.
pci_device_suspend() rt2x00pci_suspend+0x0/0x49 [rt2x00pci]() returns -16
suspend_device() pci_device_suspend+0x0/0x47() returns -16
Could not suspend device 00000104.0 error -16


07-04-2008 08:24:51

It appears that suspend.diff patch was not included in 2.6.25-rc8
It is a good patch so it is definitely worth inclusion.


07-04-2008 13:17:54

They were pushed to 2.6.25 so perhaps in the next rc release..


07-04-2008 18:38:43

They were pushed to 2.6.25 so perhaps in the next rc release..[/quotewhz3inod]

as of 2.6.25-rc8 (which has all the patches including rfkill) I have exactly the same behavior as described in kernel issue - caps lock blinks on resume.