[PATCH] log machine check errors

Linas Vepstas linas at austin.ibm.com
Fri Aug 27 08:37:36 EST 2004

On Thu, Aug 26, 2004 at 05:19:48PM -0500, Jake Moilanen was heard to remark:
> > Actually, there is one thing we probably should do, and that's to
> > change the buffer to 4K now, so that if/when firmware moves to 4K,
> > older kernels (i.e. the ones we are creating now) will continue
> > to work correctly.
> I personally disagree w/ doing this.  Memory is cheap, but
> RTAS_ERROR_LOG_MAX is used to define at least 4 static buffers that I
> can think of off the top of my head.  So that is an extra 8K for
> something that _may_ happen in the future.

yes.  BTW, they're statically allocated so that they end up in RMO
so firmware can get at them.

> Plus we've seen bugs in the
> Firmware where they do not handle us passing the error log max in
> correctly.  We may regress boxes.

I think I've fixed all of those now, pending yesterday's patch.


** Sent via the linuxppc64-dev mail list. See http://lists.linuxppc.org/

More information about the Linuxppc64-dev mailing list