Re: MMX for kernel

David S. Miller (davem@jenolan.rutgers.edu)
Fri, 7 Feb 1997 20:56:51 -0500


Date: Fri, 7 Feb 1997 20:34:04 -0500
From: "Theodore Y. Ts'o" <tytso@MIT.EDU>

Yes, that'd be useful; it'd make it much easier to note that a
process has never used the FPU at all. It also avoids disturbing
the FPU registers each time a process starts up.

I don't know, it seems to me the most sensible place to do all this is
indeed in the kernel when the user touches the fpu himself the first
time. We can provide whatever initial state is mandated on a
particular architecture.

This is in fact what we do now on the Sparc, provide a zero floating
point status register, and all fpu registers set to "all 1's"
(aka signaling NAN's).

I am perplexed as to why things aren't done this way on Intel and Libc
does this sort of setup in both libc5 and glibc2. Perhaps there is
some reason for it.

---------------------------------------------////
Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><