[src/trunk]: src/lib/libc/arch/ns32k/string Fix spurious page faults by doubl...

[matthias <matthias%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/d7d64d2fb0c0
branches:  trunk
changeset: 473536:d7d64d2fb0c0
user:      matthias <matthias%NetBSD.org@localhost>
date:      Sun Jun 06 19:49:11 1999 +0000

description:
Fix spurious page faults by double word aligning the source and not the
destination. Thank's to Ian Dall for this fix.

diffstat:

 lib/libc/arch/ns32k/string/strcpy.S |  23 +++++++++++------------
 1 files changed, 11 insertions(+), 12 deletions(-)

diffs (58 lines):

diff -r e40fa05f8eae -r d7d64d2fb0c0 lib/libc/arch/ns32k/string/strcpy.S
--- a/lib/libc/arch/ns32k/string/strcpy.S       Sun Jun 06 19:48:24 1999 +0000
+++ b/lib/libc/arch/ns32k/string/strcpy.S       Sun Jun 06 19:49:11 1999 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: strcpy.S,v 1.3 1998/04/03 22:58:10 matthias Exp $      */
+/*     $NetBSD: strcpy.S,v 1.4 1999/06/06 19:49:11 matthias Exp $      */
 
 /* 
  * Written by Randy Hyde, 1993
@@ -8,7 +8,7 @@
 #include <machine/asm.h>
 
 #if defined(LIBC_SCCS)
-       RCSID("$NetBSD: strcpy.S,v 1.3 1998/04/03 22:58:10 matthias Exp $")
+       RCSID("$NetBSD: strcpy.S,v 1.4 1999/06/06 19:49:11 matthias Exp $")
 #endif
 
 /*
@@ -23,16 +23,15 @@
        movd    B_ARG0,r2
 
        /*
-        * First begin by seeing if we can doubleword align the
-        * pointers. The following code only aligns the pointer in R2.
-        * If the L.O. two bits of R1 do not match, it's going to run
-        * slower but there is nothing we can do about that. Better to
-        * have at least one of them double word aligned rather than
-        * neither.
+        * We need to double word align the src (R1). When we read
+        * by double words we can read potentially up to 3 bytes
+        * past the null terminating byte. We can avoid spurious page
+        * faults by double word aligning the source. It also runs
+        * faster.  IWD
         */
 
        movqd   3,r3
-       andd    r2,r3
+       andd    r1,r3
 
 0:     casew   1f(pc)[r3:w]
 1:     .word   5f-0b
@@ -61,13 +60,13 @@
        addqd   1,r2
 
        /*
-        * Okay, when we get down here R2 points at a double word
-        * aligned destination block of bytes, R1 points at another
+        * Okay, when we get down here R1 points at a double word
+        * aligned source block of bytes, R2 points at the destination
         * block of bytes (typically, though not always double word
         * aligned).
         * This guy processes four bytes at a time and checks for the
         * zero terminating byte amongst the bytes in the double word.
-        * This algorithm is de to Dave Rand.
+        * This algorithm is due to Dave Rand.
         *
         * Sneaky test for zero amongst four bytes:
         *