[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: 2.4.20 kernel panics on boot w/ AS2100



On Thu, Feb 27, 2003 at 07:07:27PM -0700, Soren Harward wrote:
> I have an Alphaserver 2100 (aka "Sable") revision 4/200.  I installed
> Debian 3.0r1 on it a few days ago and updated everything to "testing".
> It's a dual-processor machine, so yesterday I built a new 2.4.20 kernel
> with SMP turned on.

I assume the installation was using a GENERIC uniprocessor kernel, and
what was installed was a GENERIC SMP kernel, but old (2.2.x)?

> However, immediately on boot, the kernel panics

Unfortunately, a common occurrence when building SMP kernels with GCC
2.95.x, IIRC.

> Since it happenned so early in the boot, I thought it might have been a
> problem with being compiled for the wrong processor type (I've seen
> similar errors trying to boot an i686 kernel on an i586 machine).  So I
> went back to the kernel config and made sure that CONFIG_ALPHA_SABLE was
> set, but that CONFIG_ALPHA_EV5 was *not* set and that CONFIG_ALPHA_EV4
> was.  See, my AS2100 uses the EV4 processor and most AS2100's use the EV5
> processor.  So I recompiled, and still got a kernel panic.

It probably wasn't the CONFIG options.

> But during this
> next compile, I noticed that the Makefile was still using the -Wa,-mev6
> switch, which passes "-mev6" (optimize for EV6) to the assembler.  So I
> commented that out of the Makefile, and re-re-compiled.  Still got a
> kernel panic.

That's NOT optimize for EV6, it's an assmebler (-Wa,) directive to
tell it to allow/process EV6 instructions if it finds them in the
assembler code passed to it. Otherwise it would flag them as errors,
or try to emulate them with older instructions, both of which are bad.
There are places in the kernel code where EV6 instructions appear,
especially for GENERIC kernels, as you'd expect, but otherwise as well.

> I'm in the middle of re-re-re-compiling using
> CONFIG_ALPHA_GENERIC; I'll see if that works.

That may work, but maybe only non-SMP, I'm not sure I remember the
limitations of GCC 2.95.x...

> Anyone have any idea what's going on?  Is it indeed that my kernel is
> getting compiled for the wrong processor, or am I completely on the
> wrong track?  What else should I try?  Different kernel version?
> Different compiler?  Different kernel options?
> 
> Again, this is kernel 2.4.20, compiling with gcc 2.95.4, running on
> Debian 3.0r1/testing.

I think the problem is a long-standing one associated with GCC 2.95.x.
Recent postings to this list, IIRC, talked about better results after
updating to GCC 3.2.

Good luck.

 --Jay++

-----------------------------------------------------------------------------
Jay A Estabrook                            HPTC - LINUX support
Hewlett-Packard Company - MRO1-2/K15       (508) 467-2080
200 Forest Street, Marlboro MA 01752       Jay.Estabrook@hp.com
-----------------------------------------------------------------------------



Reply to: