i386: inline restore_fpu
-=> Linus Torvalds wrote to All and Chuck Ebbert <=-
[..]
Chuck>
> "m" ((tsk)->thread.i387.fxsave))
>
> because that's the largest possible operand that could end up > being read?
LT> Yes, I ended up fixing that already (along with handling the fxsave LT> case too).
[..]
2.6.13-rc3-git5 lags -git4 on nbench's bitfield test, a part of its
memory index, eg typical runs
-git5
BITFIELD : 1.1589e+07 : 1.99 : 0.42
-git4
BITFIELD : 2.1892e+07 : 3.76 : 0.78
lmbench has slower context switch numbers for -git5.
--- MultiMail/Linux v0.46