Bug#664725: linux-image-3.2.0-2-rt-686-pae: Real-time kernel freezes immediately at boot most of the time
tags 664725 - lfs
thanks
Hello Gennady,
On Tue, Mar 20, 2012 at 11:19:23AM +0100, Gennady N. Uraltsev wrote:
> Package: linux-2.6
> Version: 3.2.10-1
> Severity: important
> Tags: lfs
>
> When trying to boot
> the 3.2.0-2-rt-686 kernel on my machine it freezes most of the time. In
> particular just after grub finishes loading the kernel and the initrd and
> proceeds to boot the kernel the message "Loading Linux" or something of that
> sort appears, the cursor blinks a couple of times (so far everything as usual)
> but then the cursor stops blinking and disappears completely and the system is
> completely unresponsive. Caps Lock, alt+sysrq+... and other keys don't have any
> effect.
Can you please remove "quiet" from your kernel command line and follow
up with the kernel logs. (You can get them via a serial port, netconsole
or as a last resort take a picture of your monitor.)
Additionally you can test 3.2.12 which is in unstable.
> Normally, the message that appears immediately after this on my pc, when a
> normal kernel is run, is a debug message from udev (I think) described in:
>
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=654282
>
> but that is NOT fatal and is present on all kernels (even the non rt ones).
> However in this case nothing of the sort is displayed.
>
> This problem seems very similar to the one in bug report 63269:
>
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=663269
>
> that has been said to be fixed. As a matter of fact I have noticed one
> difference:
>
> with version 3.2.10-1 (current), if I load a normal, non-rt kernel first then
> restart my pc and load the rt kernel the machine usually starts and doesn't
> hang and works perfectly, while normally (for example running the rt kernel for
> two times in a row) causes the machine to freeze. On the other hand, with
> version 3.2.9-1 there was absolutely nothing to be done to make the machine
> boot with the rt-kernel.
Yeah, 3.2.9-1 rt is known to be badly broken.
> This strange behaviour makes me suspect that something may depend on the way
> grub is booting the kernel (maybe some options that persist across boots).
Does it make a difference if you do cold booting?
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Reply to: