lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120425144452.32a803fe@notabene.brown>
Date:	Wed, 25 Apr 2012 14:44:52 +1000
From:	NeilBrown <neilb@...e.de>
To:	Chris Jones <chrisjones@...n.net.au>
Cc:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: 3.1+ kernels unbootable

On Wed, 25 Apr 2012 14:20:01 +1000 Chris Jones <chrisjones@...n.net.au> wrote:

> On Wed, 25 Apr 2012 13:21:22 +1000
> NeilBrown <neilb@...e.de> wrote:
> 
> > You might like to try 
> >  a/ describing your current system in more detail.
> >  b/ describe what actually happens when you try to boot a 3.1+ kernel.
> >     It is rare that absolutely nothing happens.
> >     Also describe the provenance of these 3.1+ kernels (do you compile
> >     yourself, or get them from a distro or....)
> > 
> > NeilBrown
> 
> 
> Thanks Neil. I use standard x86_64 Intel CPU system with 2.0GB DDRII
> RAM on a Gigabyte GA-8VM800PMD-775-RH motherboard.
> 
> This happens on both Ubuntu and Fedora kernels. And in fact, any kernel
> 3.1 or above. It's not only very odd, but also very annoying.
> 
> I get to the boot screen and when I select ENTER to boot into the
> system, my screen goes black and then the system reboots. It's
> definitely a kernel issue as anything (as mentioned already) with
> kernel 3.0 or below works just fine.
> 

You have a couple of options here.

One is to use git-bisect to narrow down where the breakage is.  This means
building about a dozen or a score of kernels and testing each one and then
trying again.  If you are happy building your own kernels and have an
afternoon to spare this is probably a good idea.  There should be plenty of
instruction on the web about how to do this but if you cannot find any feel
free to ask.

The other is to try turning features off and debugging on.
Many distros have some sort of "fail-safe" boot option which disables things
like ACPI and known-problematic drivers... though with it failing so early
most drives won't have even tried to run.  I'd guess an ACPI problem, but
that is largely because I know almost nothing about ACPI so it is easy to
blame it.  So try adding "acpi=off" to the boot args.

Linux has a thing called 'early_printk' which allows messages to be displayed
even before the normal drivers are loaded.  I don't know much about enabling
that on an x86 system (I use it a lot on ARM though).  You need it enabled
when the kernel is compiled, and you need a boot arg to enable it too.  Maybe
if you manage to enable  that you might get some message printed.

Or maybe there is some other much more useful thing you can try and someone
else will chime in soon and tell me I don't know what I'm talking about and
explain in detail the right way so solve this problem - that would be awesome.

NeilBrown

Download attachment "signature.asc" of type "application/pgp-signature" (829 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ