Loading...
--- Libc/Libc-320/ppc/string/strlen.s
+++ Libc/Libc-391/ppc/string/strlen.s
@@ -1,117 +1,100 @@
/*
- * Copyright (c) 2000-2001 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2003 Apple Computer, Inc. All rights reserved.
*
* @APPLE_LICENSE_HEADER_START@
*
- * Copyright (c) 1999-2003 Apple Computer, Inc. All Rights Reserved.
+ * The contents of this file constitute Original Code as defined in and
+ * are subject to the Apple Public Source License Version 1.1 (the
+ * "License"). You may not use this file except in compliance with the
+ * License. Please obtain a copy of the License at
+ * http://www.apple.com/publicsource and read it before using this file.
*
- * This file contains Original Code and/or Modifications of Original Code
- * as defined in and that are subject to the Apple Public Source License
- * Version 2.0 (the 'License'). You may not use this file except in
- * compliance with the License. Please obtain a copy of the License at
- * http://www.opensource.apple.com/apsl/ and read it before using this
- * file.
- *
- * The Original Code and all software distributed under the License are
- * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
+ * This Original Code and all software distributed under the License are
+ * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
- * Please see the License for the specific language governing rights and
- * limitations under the License.
+ * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT. Please see the
+ * License for the specific language governing rights and limitations
+ * under the License.
*
* @APPLE_LICENSE_HEADER_END@
*/
-#define ASSEMBLER
+
+/* We use mode-independent "g" opcodes such as "srgi". These expand
+ * into word operations when targeting __ppc__, and into doubleword
+ * operations when targeting __ppc64__.
+ */
+#include <architecture/ppc/mode_independent_asm.h>
+
#include <mach/ppc/asm.h>
-#undef ASSEMBLER
-;
-;
-; Strlen, optimized for PPC. The routine we use is 2-3x faster
-; then the simple loop which checks each byte for zero.
-; For 0- and 1-byte strings, the simple routine is faster, but
-; only by a few cycles. The algorithm used was adapted from the
-; Mac OS 9 stdCLib strcopy routine, which was originally
-; written by Gary Davidian. It relies on the following rather
-; inobvious but very efficient test:
-;
-; y = dataWord + 0xFEFEFEFF
-; z = ~dataWord & 0x80808080
-; if ( y & z ) = 0 then all bytes in dataWord are non-zero
-;
-; The test maps any non-zero byte to zeros and any zero byte to 0x80,
-; with one exception: 0x01 bytes preceeding the first zero are also
-; mapped to 0x80.
-;
-;
-; int strlen(ptr)
-;
-;
+#define __APPLE_API_PRIVATE
+#include <machine/cpu_capabilities.h>
+#undef __APPLE_API_PRIVATE
- .align 5
- .globl EXT(strlen)
-LEXT(strlen)
- andi. r4,r3,0x03 ; test alignment first
- mr r9,r3 ; store the original address for later use....
- bne LalignSource ; align the source addr if not already aligned
-Llentry:
- lis r5,hi16(0xFEFEFEFF)
- lis r6,hi16(0x80808080)
- subi r3,r3,0x04 ; pre-decrement r3 for the lwzu
- ori r5,r5,lo16(0xFEFEFEFF) ; r5=0xFEFEFEFF
- ori r6,r6,lo16(0x80808080) ; r6=0x80808080
+// Strlen, optimized for PPC. We use an inobvious but very efficient
+// word-parallel test for 0-bytes:
+//
+// y = dataWord + 0xFEFEFEFF
+// z = ~dataWord & 0x80808080
+// if ( y & z ) = 0 then all bytes in dataWord are non-zero
+//
+// The test maps any non-zero byte to zeros and any zero byte to 0x80,
+// with one exception: 0x01 bytes preceeding the first zero are also
+// mapped to 0x80. Using altivec is another possibility, but it turns
+// out that the overhead of maintaining VRSAVE and dealing with edge
+// cases pushes the crossover point out to around 30 bytes... longer
+// the the "typical" operand length.
+//
+// In 64-bit mode, the algorithm is doubleword parallel.
-LLoop:
- lwzu r8,4(r3) ; get the first 4 bytes and increment address
- add r4,r5,r8 ; r4= data + 0xFEFEFEFF
- andc r7,r6,r8 ; r7= ~data & 0x80808080
- and. r4,r4,r7 ; r4= r4 & r7
- beq LLoop ; if r4 is zero, then all bytes are non-zero
+ .text
+ .align 5
+ .globl EXT(strlen)
+LEXT(strlen) // int strlen(ptr)
+ clrrgi r9,r3,LOG2_GPR_BYTES// align pointer by zeroing right LOG2_GPR_BYTES bits
+ li r7,-1 // get 0xFFs
+ lg r8,0(r9) // get word or doubleword with 1st operand byte
+ rlwinm r4,r3,3,(GPR_BYTES-1)*8 // get starting bit position of operand
+#if defined(__ppc__)
+ lis r5,hi16(0xFEFEFEFF) // start to generate 32-bit magic constants
+ lis r6,hi16(0x80808080)
+ srw r7,r7,r4 // create a mask of 0xFF bytes for operand in r8
+ ori r5,r5,lo16(0xFEFEFEFF)
+ ori r6,r6,lo16(0x80808080)
+#else
+ ld r5,_COMM_PAGE_MAGIC_FE(0) // get 0xFEFEFEFE FEFEFEFF from commpage
+ ld r6,_COMM_PAGE_MAGIC_80(0) // get 0x80808080 80808080 from commpage
+ srd r7,r7,r4 // create a mask of 0xFF bytes for operand in r8
+#endif
+ orc r8,r8,r7 // make sure bytes preceeding operand are 0xFF
+ b Lloop1 // enter loop
+
+// Loop over words or doublewords.
+// r3 = original address
+// r5 = 0xFEFEFEFE FEFEFEFF
+// r6 = 0x80808080 80808080
+// r9 = address (aligned)
-; Now we know one of the bytes in r8 is zero,
-; we just have to figure out which one.
-; We have mapped 0 bytes to 0x80, and nonzero bytes to 0x00,
-; with one exception:
-; 0x01 bytes preceeding the first zero are also mapped to 0x80.
-; So we have to mask out the 0x80s caused by 0x01s before
-; counting leading zeroes to get the bytes in last word.
+ .align 5
+Lloop:
+ lgu r8,GPR_BYTES(r9) // get next word or doubleword
+Lloop1: // initial entry
+ add r4,r5,r8 // r4 = data + 0xFEFEFEFF
+ andc r7,r6,r8 // r7 = ~data & 0x80808080
+ and. r4,r4,r7 // r4 = r4 & r7
+ beq Lloop // if r4 is zero, then all bytes are non-zero
- rlwinm r5,r8,7,0,31 ; move 0x01 bits to 0x80 position
- subf r3,r9,r3 ; start to compute string length
- andc r4,r4,r5 ; turn off false hits from 0x0100 worst case
- cntlzw r7,r4 ; now we can count leading 0s
- srwi r7,r7,3 ; convert 0,8,16,24 to 0,1,2,3
- add r3,r3,r7 ; add in nonzero bytes in last word
- blr
+// Now we know one of the bytes in r8 is zero, we just have to figure out which one.
+// We have mapped 0 bytes to 0x80, and nonzero bytes to 0x00, with one exception:
+// 0x01 bytes preceeding the first zero are also mapped to 0x80. So we have to mask
+// out the 0x80s caused by 0x01s before searching for the 0x80 byte.
-; We must align the source address for two reasons: to avoid spurious page
-; faults, and for speed.
-; r4 = low 2 bits of address (1,2, or 3)
-; r3 = address
-; r9 = original address (still same as r3)
-
-LalignSource:
- lbz r5,0(r3) ; get the first byte...
- subic. r4,r4,2 ; test for 1, 2 or 3 bytes
- addi r3,r3,1 ; increment address
- addi r6,r9,1 ; now r6==r3
- cmpwi cr1,r5,0 ; zero?
- beq cr1,Lreturn ; if its zero return zero
- bgt Llentry ; address is aligned now if low bits were 3
-
- lbz r5,0(r3) ; get the next byte...
- addi r3,r3,1 ; increment address
- cmpwi cr1,r5,0 ; zero?
- beq cr1,Lreturn ; if its zero return one
- beq Llentry ; addr is aligned now if low bits were 2
-
- lbz r5,0(r3) ; get the next byte...
- addi r3,r3,1 ; increment address
- cmpwi cr1,r5,0 ; zero?
- bne cr1,Llentry ; not zero, continue check (now aligned)
-Lreturn:
- sub r3,r3,r6 ; get string length (0, 1, or 2)
- blr
-
+ slgi r5,r8,7 // move 0x01 bits to 0x80 position
+ sub r3,r9,r3 // start to compute string length
+ andc r4,r4,r5 // turn off false hits from 0x0100 worst case
+ cntlzg r7,r4 // now we can count leading 0s
+ srwi r7,r7,3 // convert 0,8,16,24 to 0,1,2,3, etc
+ add r3,r3,r7 // add in nonzero bytes in last word
+ blr