← Back to team overview

hybrid-graphics-linux team mailing list archive

Re: ASUS N53Jn report -- debug info on shutdown freeze with card OFF

 

Le lundi 28 février 2011 à 14:47 +0100, Romain Knezevic a écrit :
> Le lundi 28 février 2011 à 11:29 +0000, Peter Goodall a écrit : 
> > On Mon, Feb 28, 2011 at 3:14 AM, Romain Knezevic
> > <romainknezevic@xxxxxxxxx> wrote:
> > > Running on an ASUS N53Jn w/ Ubuntu 10.10
> > >
> > > The test_off.sh script works fine, but I get 2 freezes
> > >        - from shuting down if I do not turn the card back on beforehand
> > >
> > 
> > I think I get the same behaviour with Ubuntu 11.04, but need to
> > verify.  Would like to know how to get useful debugging information if
> > this is the case.
> > 
> 
> Yeah agree with that, some kind of dump ? Don't really know unix's crash
> policy ! I'm thinking maybe one of the scripts in /etc/rc*.d is
> responsible for the crash ? ---- OK writing this made me realize that it
> wouldnt take long to test this out myself, so I did it. 
> Here is my current working /etc/rc6.d folder :
>         K74bluetooth -> ../init.d/bluetooth
>         README
>         S10rebootrasp -> ../init.d/rebootrasp
>         S10unattended-upgrades -> ../init.d/unattended-upgrades
>         S20sendsigs -> ../init.d/sendsigs
>         S30urandom -> ../init.d/urandom
>         S31umountnfs.sh -> ../init.d/umountnfs.sh
>         S35networking -> ../init.d/networking
>         S40umountfs -> ../init.d/umountfs
>         S60umountroot -> ../init.d/umountroot
>         S90reboot -> ../init.d/reboot
> As I understood it while trying to get to run my script that turns the
> card back on before shutdown, the order of execution is kill
> K00daemon...kill K99daemon...run S00script...run S99script.
> 
> So that is my working configuration above, the link S10rebootrasp
> -> /etc/init.d/rebootrasp is mine, it only contains 2 lines, one turns
> the card on and the other deletes a guard file that prevents my .bashrc
> to run test_off.sh script more than once, so it's turned back OFF on
> next fresh start when I open a terminal. Ok so with this I reboot with
> no problems, I just tried "mv S10rebootrasp S91rebootrasp" so my script
> is executed last, and it's not at all. It crashes and after hardreboot I
> went to check if my guard file has been deleted and it was still here,
> so the crash comes from one of the scripts above, I'm guessing
> S90reboot. 
> I could try placing my script between every pair but lazy
> right now maybe later.
> 

Ok, tried it, it's S20sendsigs that screws up. Also after 2 more
freezes, the message described in my previous mail was the same,
except the line of the error and sometimes the proc # changed.
I had

BUG: soft lockup - CPU#3 [Xorg:1416]
for one freeze
and

BUG: soft lockup - CPU#0 [Xorg:1437]
for the other, and the first one in my previous mail

BUG: soft lockup - CPU#0 [Xorg:1438]




References