Re: context switching pessimizations

From: Bruce Evans <bde_at_zeta.org.au>
Date: Sat, 5 Apr 2003 20:52:31 +1000 (EST)
On Sat, 5 Apr 2003, Bruce Evans wrote:

> Recent optimizations of context switching are large pessimizations
> according to lmbench2, at least in the default case (no LAZY_SWITCH).

The ifdefs are just so convoluted that they are broken in the
!LAZY_SWITCH case.

%%%
Index: swtch.s
===================================================================
RCS file: /home/ncvs/src/sys/i386/i386/swtch.s,v
retrieving revision 1.136
diff -u -2 -r1.136 swtch.s
--- swtch.s	2 Apr 2003 23:53:28 -0000	1.136
+++ swtch.s	5 Apr 2003 10:40:20 -0000
_at__at_ -175,4 +175,5 _at__at_
 #endif
 1:
+#endif
 	movl	%cr3,%ebx			/* The same address space? */
 	cmpl	%ebx,%eax
_at__at_ -182,6 +183,4 _at__at_
 	je	sw1
 #endif
-#endif
-
 #ifdef SWTCH_OPTIM_STATS
 	incl	tlb_flush_count
%%%

Bruce
Received on Sat Apr 05 2003 - 00:52:35 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:02 UTC