kernel: bzeront() - Switch out loop instruction in i386 bzero for sub/jnz.

Much as I like loop, it has a ~7-9 cycle latency on AMD CPUs. Lets spend
idlezero time actually zeroing...

diff --git a/sys/cpu/i386/misc/bzeront.s b/sys/cpu/i386/misc/bzeront.s
index bae1ec7..5759338 100644 (file)
--- a/sys/cpu/i386/misc/bzeront.s
+++ b/sys/cpu/i386/misc/bzeront.s
@@ -53,6 +53,7 @@ ENTRY(bzeront)
        shrl $2, %ecx
 1:
        movnti %edx, (%eax)
-       add $4, %eax
-       loop 1b
+       addl $4, %eax
+       subl $1, %ecx
+       jnz 1b
        ret