Messages in this thread | | | Date | Wed, 3 Feb 1999 21:12:08 +0000 | From | Philipp Rumpf <> | Subject | Re: [patch] Katmai/PIII support for Linux pre-2.2.2-1 |
| |
> please take a look at my patch, it pretty much shows that are lots of > other things to be taken care of, not only context switching. (ptrace, > signals, ELF compatibility)
Did you provide an interface to get the complete FPU state yet ? I don't think developing code will be much fun when you cannot see the XMMX registers...
> I dont think dynamic usage is achievable in a sane way, and i dont think it's desirable.
I cannot think of anything that could not be handled in a clean and not-to-slow way (AFAIK, ptrace FPU access is mostly used when debugging which is not normally performance-critical).
> we _have_ a common kernel, which is called CONFIG_M386. PII kernels
This common kernel does not have security holes (that can be fixed).
Even though it does not use 486+-specific instructions, it has no serious performance impact. For games, we are talking about 30% - 60% here IMO.
> It makes tons of sense.
Yes. Nevertheless, the main reason (I think) for splitting up the architecture is to achieve maximal performance for a given processor.
> no Philipp, this is one of the reasons why i didnt use the approach in > your patch. It's just so ugly to do boot-time linking that way. Also
IA32 _is_ ugly.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |