PIRQ problem (Please try using pci=biosirq)

Deepak Kotian dkotian3 at vsnl.net
Sun May 18 01:17:01 CEST 2003


Thanks, this is a good one.
One more question. How does one enable Shadow Ram.

Regards
Deepak


----- Original Message -----
From: "Eric W. Biederman" <ebiederman at lnxi.com>
To: "Deepak Kotian" <dkotian3 at vsnl.net>
Cc: "ron minnich" <rminnich at lanl.gov>; <linuxbios at clustermatic.org>
Sent: Sunday, May 18, 2003 12:01 AM
Subject: Re: PIRQ problem (Please try using pci=biosirq)


> "Deepak Kotian" <dkotian3 at vsnl.net> writes:
>
> > I do not have acces to my system now, so cannot tell for sure but what
> > it the signifigance  of this paramater CONFIG_COMPRESS = 0 or 1
> > I have seen this in couple of other responses as well which talks about
this
> > one.
> >
> > Can on please elaborate on this please and how this paramerter would
lead
> > to this error.
>
> LinuxBIOS is naturally split into 2 components.
> The part that runs after ram is believed to be working.
> And the part that is used to initialize ram.
>
> During ram initialization the code must run out of
> the ROM chip.  After that however the code can live
> in RAM.  And we take advantage of that fact by compressing
> that code in the ROM chip.
>
> Currently the PIRQ tables are just copied from the LinuxBIOS
> image in the ROM chip into RAM.  Around 0xf0000 - 0xfffff.
> If ``shadow'' ram has not been enabled there the copy fails.
>
> However if LinuxBIOS is not compressed in the ROM chip, the
> uncompressed table can still be found in the last 64K before 1M.
> And so the pirq table can still be found that way.
>
> The real fix is to enable shadow RAM.  Disabling compression
> is an effective short term work around.
>
> Eric
>
> _______________________________________________
> Linuxbios mailing list
> Linuxbios at clustermatic.org
> http://www.clustermatic.org/mailman/listinfo/linuxbios
>





More information about the coreboot mailing list