Subject: Re: VIA C3 woes, revisited?
To: None <port-i386@netbsd.org>
From: Aaron J. Grier <agrier@poofygoof.com>
List: port-i386
Date: 02/27/2007 14:25:12
On Tue, Feb 27, 2007 at 10:19:52AM -0600, John D. Baker wrote:
> The install sets I built work fine on the build host (P-II/266) and
> other i386-class machines I have.  I'm not using any extra flags
> anywhere, so the defaults should apply.

I replaced an athlon with a C3 a while back; I had a bunch of binaries
on it which were optimized for the athlon, which had a few more
instructions on it than the C3:

cpu0: VIA C3 Samuel 2/Ezra (686-class), 796.17 MHz, id 0x673
cpu0: features 80803035<FPU,DE,TSC,MSR,MTRR>
cpu0: features 80803035<PGE,MMX>
cpu0: features 80803035<3DNOW>
cpu0: "VIA Samuel 2"
cpu0: I-cache 64 KB 32B/line 4-way, D-cache 64 KB 32B/line 4-way
cpu0: L2 cache 64 KB 32B/line 4-way
cpu0: ITLB 128 4 KB entries 8-way
cpu0: DTLB 128 4 KB entries 8-way
cpu0: 4 page colors

vs

cpu0: AMD Athlon Model 1 (686-class), 500.05 MHz
cpu0: I-cache 64 KB 64b/line 2-way, D-cache 64 KB 64b/line 2-way
cpu0: L2 cache 512 KB 64b/line 2-way
cpu0: features 81f9ff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR>
cpu0: features 81f9ff<PGE,MCA,CMOV,FGPAT,MMX>

cmove was the offending instruction in my case.

you can find out what's at fault if you can get gdb to run:
http://mail-index.netbsd.org/port-i386/2006/03/23/0002.html

I drove myself crazy for a few days figuring this out.

-- 
  Aaron J. Grier | "Not your ordinary poofy goof." | agrier@poofygoof.com
              "silly brewer, saaz are for pils!"  --  virt