Re: iwl4965: "MAC is in deep sleep!" freezes

[Christoph Anton Mitterer <calestyo@scientia.net>]

[-- Attachment #1: Type: text/plain, Size: 1996 bytes --]

Hi.

Sorry for completely forgetting about this... (for so long)

I've tried your patch now against 3.1, and it seems to not help.

Not sure whether trying with 2.6.38 helps.
It "stopped" working within the very same kernel.

The old mainboard got broken and was replaced by a new one (of the same
type),... but perhaps the wifi controller is of a new revision or so?!


Cheers,
Chris.


On Thu, 2011-09-08 at 10:49 +0200, Stanislaw Gruszka wrote:
> On Mon, Sep 05, 2011 at 09:42:01AM +0000, Christoph Anton Mitterer wrote:
> > On Mon, 5 Sep 2011 11:19:18 +0200, Stanislaw Gruszka <sgruszka@redhat.com>
> > wrote:
> > > this look like firmware hang. This happen just after module load, that's
> > > good news, because it allow to log relative small amount debug messages
> > > to see what possibly driver do wrong to crash firmware.
> > Note, that the effects of - usually but not always - frozen system while
> > the errors are printed (sometimes keyboard input seems still be possible
> > during that) ... happen over and over again... say about every  10 minutes
> > or so.
> > Not just the first time the modules is loaded.
> > Is the firmware loaded over and over again?
> I think is not, but even if is, kernel should not crash. Can you grab
> crash logs using photo camera, serial cable or netconsole, to allow to 
> see where exactly crash happen and fix it.
> 
> > > Please configure syslog to log kernel debug messages.
> > Will mail it later.
> 
> Firmware hang seems to happen when REPLY_LEDS_CMD command is send.
> No idea if attached patch could help or not, but is worth to try it.
> It prevent to send LED command asynchronously, when possibly some
> other commands are pending in firmware.
> 
> If it does not help, please provide the same way captured debug
> messages from 2.6.38 kernel, where drivers works. I'll compare
> it with broken driver logs to figure out what broken driver
> do differently.
> 
> Thanks
> Stanislaw


[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5677 bytes --]