[coreboot] [PATCH] fix 'AMD Fam10 code breaks with gcc 4.5.0'

[Paul Menzel]

Dear Scott,


Am Freitag, den 03.09.2010, 00:45 +0200 schrieb Paul Menzel:

> Am Donnerstag, den 02.09.2010, 16:06 -0500 schrieb Scott:
> > -----Original Message-----
> > From: coreboot-bounces at coreboot.org [mailto:coreboot-bounces at coreboot.org] On Behalf Of Paul Menzel
> > Sent: Thursday, September 02, 2010 04:11 AM
> > To: coreboot at coreboot.org
> > Subject: Re: [coreboot] [PATCH] fix 'AMD Fam10 code breaks with gcc 4.5.0'
> 
> > Am Donnerstag, den 02.09.2010, 10:59 +0200 schrieb Paul Menzel:
> > > Am Donnerstag, den 02.09.2010, 00:40 -0500 schrieb Scott:
> > > 
> > > > The subversion comment for -r 5571 states:
> > > > 
> > > >    The AMD Fam10 code breaks with coreboot 4.5.0.
> > > >    Potentially caused by reordering. Going back to 4.4.4
> > > >    which is known working on Fam10 until gcc or the Fam10 code is fixed.
> > > > 
> > > >    Signed-off-by: Stefan Reinauer <stepan at coresystems.de>
> > > >    Acked-by: Stefan Reinauer <stepan at coresystems.de>
> > > > 
> > > > 
> > > > I encountered the same problem and debugged it. The AP code that disables
> > > > cache as ram before the final halt has to be all inline. Function calls
> > > > require a valid stack, and the stack is kept in the very cache as ram that
> > > > the code is disabling. I found with gcc 450, the code for rdmsr, disable_cache,
> > > > and enable_cache and not getting inlined as intended. Function calls are
> > > 
> > > s/and/are/
> > > 
> > > > generated, and the first one after the AP clears msr 268 fails. The solution
> > > > is to force these functions to generate inline code by adding
> > > > __attribute__((always_inline)) to their declarations:
> > > 
> > > great find!!! Could you please send a patch according to the development
> > > guidelines. Especially do not forget to add your Signed-off-by line.
> > 
> > They are documented in the Wiki [1].
> > 
> > > Just to leave no doubt, could you please add that you tested your fix
> > > using GCC 4.5.0 with the used hardware.
> > > 
> > > [.]
> 
> > OK, I resubmitted it. Currently tested on simnow only. I will make a
> > Tilapia binary and ask the AMD guys to test it. By the way, the patch
> > file has Windows style line endings. I wonder if others can use it easily...
> 
> I guess next time this message should go to the list since it contains
> information interesting to the others too.

[I hit »Sent« too early. I am sorry!]

… therefore I am replying to the list again.

I hope other will have time to test this. We will see if the developers
have a problem when committing your patch. If yes, I guess Patrick can
give you a hint on how to circumvent this problem.


Thanks,

Paul


PS: Do you know if you can configure Outlook to use a style when
replying similar to everyone else is using on this list, i. e. use a
line »Joe Doe wrote on …« and afterward quoting everything by putting
»>« in front of every line?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://www.coreboot.org/pipermail/coreboot/attachments/20100903/6f2c8281/attachment.sig>