From: NIIBE Y. <gn...@ch...> - 2000-08-11 09:09:46
|
Greg Banks wrote: > Sorry, I was going to do this Monday but instead I spent all day > trying to track down an NFS problem. Maybe tonight. You don't need to feel sorry. It's not large change, you could postpone it. * * * I've been chasing cache issue (I guess) of SH-4. I'm sure there's some problem around swap code. I don't know if it's in generic code or in SuperH specific code yet. With swap, process occasionaly segfaults or aborts on SH-4. You can reproduce it with setting the memory, say, 16MB. When I disable operand cache ("data cache" in terms of SuperH), it works fine (quite slow though). On SH-3, it works fine. So far, I couldn't find the bug yet. -- |