[Xorg-driver-geode] [PATCH v2] config: only build on i*86 host cpu.

[Gaetan Nadon]

On Tue, 2010-06-22 at 13:46 -0300, Otavio Salvador wrote:

> I think that a message saying that it is know to work only at i386 is
> fine but completely skiping it looks wrong to me. It would be better
> to fix Xorg docs/scripts to handle it properly.
> 

I can't disagree with that, the usual "error" vs "warning" headache :-)

I can live with stopping the configuration if there is a good message.
However, those who have figured it out on their own open bugs on the
build system(s) claiming
it should not be built. 

There isn't much difference between a package skipped by the build
script and a package
that is skipping itself. The benefit of the later is that there are
multiple build systems and some don't 
have platform checking. Anyone can also cd to /geode and type make.

I think I have explained everything now. If Martin's proposal goes
through,  I can test on AMD64 and Intel 32.

PS: I always forget to show what errors I get on AMD64:

durango.c: Assembler messages:
durango.c:203: Error: suffix or operands invalid for `push'

#define vsa_msr_write(msr,adr,high,low) \
  { int d0, d1, d2, d3, d4;        \
  __asm__ __volatile__(            \
    " push %%ebx\n"                \

In x86_64, the register is called rbx. I am no expert here.
Perhaps there is a compiler flag to handle that. 

Cheers,
Gaetan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg-driver-geode/attachments/20100622/41aa3ff5/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg-driver-geode/attachments/20100622/41aa3ff5/attachment.pgp>