diff mbox series

[v2,3/8] arm64: Import updated version of Cortex Strings' strlen

Message ID 32e3489398a24b23ae6e996935ac4818f8fd9dfd.1622128527.git.robin.murphy@arm.com (mailing list archive)
State New, archived
Headers show
Series arm64: String function updates | expand

Commit Message

Robin Murphy May 27, 2021, 3:34 p.m. UTC
From: Sam Tebbs <sam.tebbs@arm.com>

Import an updated version of the former Cortex Strings - now Arm
Optimized Routines - strcmp function. The latest version introduces
Advanced SIMD usage which rules it out for our purposes, but we can
still pick an intermediate improvement from the previous version,
namely string/aarch64/strlen.S at commit 98e4d6a from
https://github.com/ARM-software/optimized-routines

Note that for simplicity Arm have chosen to contribute this code
to Linux under GPLv2 rather than the original MIT license.

Signed-off-by: Sam Tebbs <sam.tebbs@arm.com>
[ rm: update attribution and commit message ]
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
---
 arch/arm64/lib/strlen.S | 262 +++++++++++++++++++++++++++-------------
 1 file changed, 175 insertions(+), 87 deletions(-)
diff mbox series

Patch

diff --git a/arch/arm64/lib/strlen.S b/arch/arm64/lib/strlen.S
index ee3ed882dd79..b557185b54a5 100644
--- a/arch/arm64/lib/strlen.S
+++ b/arch/arm64/lib/strlen.S
@@ -1,115 +1,203 @@ 
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (C) 2013 ARM Ltd.
- * Copyright (C) 2013 Linaro.
+ * Copyright (c) 2013, Arm Limited.
  *
- * This code is based on glibc cortex strings work originally authored by Linaro
- * be found @
- *
- * http://bazaar.launchpad.net/~linaro-toolchain-dev/cortex-strings/trunk/
- * files/head:/src/aarch64/
+ * Adapted from the original at:
+ * https://github.com/ARM-software/optimized-routines/blob/master/string/aarch64/strlen.S
  */
 
 #include <linux/linkage.h>
 #include <asm/assembler.h>
 
-/*
- * calculate the length of a string
+/* Assumptions:
  *
- * Parameters:
- *	x0 - const string pointer
- * Returns:
- *	x0 - the return length of specific string
+ * ARMv8-a, AArch64, unaligned accesses, min page size 4k.
  */
 
+#define L(label) .L ## label
+
 /* Arguments and results.  */
-srcin		.req	x0
-len		.req	x0
+#define srcin		x0
+#define len		x0
 
 /* Locals and temporaries.  */
-src		.req	x1
-data1		.req	x2
-data2		.req	x3
-data2a		.req	x4
-has_nul1	.req	x5
-has_nul2	.req	x6
-tmp1		.req	x7
-tmp2		.req	x8
-tmp3		.req	x9
-tmp4		.req	x10
-zeroones	.req	x11
-pos		.req	x12
+#define src		x1
+#define data1		x2
+#define data2		x3
+#define has_nul1	x4
+#define has_nul2	x5
+#define tmp1		x4
+#define tmp2		x5
+#define tmp3		x6
+#define tmp4		x7
+#define zeroones	x8
+
+	/* NUL detection works on the principle that (X - 1) & (~X) & 0x80
+	   (=> (X - 1) & ~(X | 0x7f)) is non-zero iff a byte is zero, and
+	   can be done in parallel across the entire word. A faster check
+	   (X - 1) & 0x80 is zero for non-NUL ASCII characters, but gives
+	   false hits for characters 129..255.	*/
 
 #define REP8_01 0x0101010101010101
 #define REP8_7f 0x7f7f7f7f7f7f7f7f
 #define REP8_80 0x8080808080808080
 
+#define MIN_PAGE_SIZE 4096
+
+	/* Since strings are short on average, we check the first 16 bytes
+	   of the string for a NUL character.  In order to do an unaligned ldp
+	   safely we have to do a page cross check first.  If there is a NUL
+	   byte we calculate the length from the 2 8-byte words using
+	   conditional select to reduce branch mispredictions (it is unlikely
+	   strlen will be repeatedly called on strings with the same length).
+
+	   If the string is longer than 16 bytes, we align src so don't need
+	   further page cross checks, and process 32 bytes per iteration
+	   using the fast NUL check.  If we encounter non-ASCII characters,
+	   fallback to a second loop using the full NUL check.
+
+	   If the page cross check fails, we read 16 bytes from an aligned
+	   address, remove any characters before the string, and continue
+	   in the main loop using aligned loads.  Since strings crossing a
+	   page in the first 16 bytes are rare (probability of
+	   16/MIN_PAGE_SIZE ~= 0.4%), this case does not need to be optimized.
+
+	   AArch64 systems have a minimum page size of 4k.  We don't bother
+	   checking for larger page sizes - the cost of setting up the correct
+	   page size is just not worth the extra gain from a small reduction in
+	   the cases taking the slow path.  Note that we only care about
+	   whether the first fetch, which may be misaligned, crosses a page
+	   boundary.  */
+
 SYM_FUNC_START_WEAK_PI(strlen)
-	mov	zeroones, #REP8_01
-	bic	src, srcin, #15
-	ands	tmp1, srcin, #15
-	b.ne	.Lmisaligned
-	/*
-	* NUL detection works on the principle that (X - 1) & (~X) & 0x80
-	* (=> (X - 1) & ~(X | 0x7f)) is non-zero iff a byte is zero, and
-	* can be done in parallel across the entire word.
-	*/
-	/*
-	* The inner loop deals with two Dwords at a time. This has a
-	* slightly higher start-up cost, but we should win quite quickly,
-	* especially on cores with a high number of issue slots per
-	* cycle, as we get much better parallelism out of the operations.
-	*/
-.Lloop:
-	ldp	data1, data2, [src], #16
-.Lrealigned:
+	and	tmp1, srcin, MIN_PAGE_SIZE - 1
+	mov	zeroones, REP8_01
+	cmp	tmp1, MIN_PAGE_SIZE - 16
+	b.gt	L(page_cross)
+	ldp	data1, data2, [srcin]
+#ifdef __AARCH64EB__
+	/* For big-endian, carry propagation (if the final byte in the
+	   string is 0x01) means we cannot use has_nul1/2 directly.
+	   Since we expect strings to be small and early-exit,
+	   byte-swap the data now so has_null1/2 will be correct.  */
+	rev	data1, data1
+	rev	data2, data2
+#endif
 	sub	tmp1, data1, zeroones
-	orr	tmp2, data1, #REP8_7f
+	orr	tmp2, data1, REP8_7f
 	sub	tmp3, data2, zeroones
-	orr	tmp4, data2, #REP8_7f
-	bic	has_nul1, tmp1, tmp2
-	bics	has_nul2, tmp3, tmp4
-	ccmp	has_nul1, #0, #0, eq	/* NZCV = 0000  */
-	b.eq	.Lloop
+	orr	tmp4, data2, REP8_7f
+	bics	has_nul1, tmp1, tmp2
+	bic	has_nul2, tmp3, tmp4
+	ccmp	has_nul2, 0, 0, eq
+	beq	L(main_loop_entry)
 
-	sub	len, src, srcin
-	cbz	has_nul1, .Lnul_in_data2
-CPU_BE(	mov	data2, data1 )	/*prepare data to re-calculate the syndrome*/
-	sub	len, len, #8
-	mov	has_nul2, has_nul1
-.Lnul_in_data2:
-	/*
-	* For big-endian, carry propagation (if the final byte in the
-	* string is 0x01) means we cannot use has_nul directly.  The
-	* easiest way to get the correct byte is to byte-swap the data
-	* and calculate the syndrome a second time.
-	*/
-CPU_BE( rev	data2, data2 )
-CPU_BE( sub	tmp1, data2, zeroones )
-CPU_BE( orr	tmp2, data2, #REP8_7f )
-CPU_BE( bic	has_nul2, tmp1, tmp2 )
-
-	sub	len, len, #8
-	rev	has_nul2, has_nul2
-	clz	pos, has_nul2
-	add	len, len, pos, lsr #3		/* Bits to bytes.  */
+	/* Enter with C = has_nul1 == 0.  */
+	csel	has_nul1, has_nul1, has_nul2, cc
+	mov	len, 8
+	rev	has_nul1, has_nul1
+	clz	tmp1, has_nul1
+	csel	len, xzr, len, cc
+	add	len, len, tmp1, lsr 3
 	ret
 
-.Lmisaligned:
-	cmp	tmp1, #8
-	neg	tmp1, tmp1
-	ldp	data1, data2, [src], #16
-	lsl	tmp1, tmp1, #3		/* Bytes beyond alignment -> bits.  */
-	mov	tmp2, #~0
-	/* Big-endian.  Early bytes are at MSB.  */
-CPU_BE( lsl	tmp2, tmp2, tmp1 )	/* Shift (tmp1 & 63).  */
-	/* Little-endian.  Early bytes are at LSB.  */
-CPU_LE( lsr	tmp2, tmp2, tmp1 )	/* Shift (tmp1 & 63).  */
+	/* The inner loop processes 32 bytes per iteration and uses the fast
+	   NUL check.  If we encounter non-ASCII characters, use a second
+	   loop with the accurate NUL check.  */
+	.p2align 4
+L(main_loop_entry):
+	bic	src, srcin, 15
+	sub	src, src, 16
+L(main_loop):
+	ldp	data1, data2, [src, 32]!
+L(page_cross_entry):
+	sub	tmp1, data1, zeroones
+	sub	tmp3, data2, zeroones
+	orr	tmp2, tmp1, tmp3
+	tst	tmp2, zeroones, lsl 7
+	bne	1f
+	ldp	data1, data2, [src, 16]
+	sub	tmp1, data1, zeroones
+	sub	tmp3, data2, zeroones
+	orr	tmp2, tmp1, tmp3
+	tst	tmp2, zeroones, lsl 7
+	beq	L(main_loop)
+	add	src, src, 16
+1:
+	/* The fast check failed, so do the slower, accurate NUL check.	 */
+	orr	tmp2, data1, REP8_7f
+	orr	tmp4, data2, REP8_7f
+	bics	has_nul1, tmp1, tmp2
+	bic	has_nul2, tmp3, tmp4
+	ccmp	has_nul2, 0, 0, eq
+	beq	L(nonascii_loop)
+
+	/* Enter with C = has_nul1 == 0.  */
+L(tail):
+#ifdef __AARCH64EB__
+	/* For big-endian, carry propagation (if the final byte in the
+	   string is 0x01) means we cannot use has_nul1/2 directly.  The
+	   easiest way to get the correct byte is to byte-swap the data
+	   and calculate the syndrome a second time.  */
+	csel	data1, data1, data2, cc
+	rev	data1, data1
+	sub	tmp1, data1, zeroones
+	orr	tmp2, data1, REP8_7f
+	bic	has_nul1, tmp1, tmp2
+#else
+	csel	has_nul1, has_nul1, has_nul2, cc
+#endif
+	sub	len, src, srcin
+	rev	has_nul1, has_nul1
+	add	tmp2, len, 8
+	clz	tmp1, has_nul1
+	csel	len, len, tmp2, cc
+	add	len, len, tmp1, lsr 3
+	ret
+
+L(nonascii_loop):
+	ldp	data1, data2, [src, 16]!
+	sub	tmp1, data1, zeroones
+	orr	tmp2, data1, REP8_7f
+	sub	tmp3, data2, zeroones
+	orr	tmp4, data2, REP8_7f
+	bics	has_nul1, tmp1, tmp2
+	bic	has_nul2, tmp3, tmp4
+	ccmp	has_nul2, 0, 0, eq
+	bne	L(tail)
+	ldp	data1, data2, [src, 16]!
+	sub	tmp1, data1, zeroones
+	orr	tmp2, data1, REP8_7f
+	sub	tmp3, data2, zeroones
+	orr	tmp4, data2, REP8_7f
+	bics	has_nul1, tmp1, tmp2
+	bic	has_nul2, tmp3, tmp4
+	ccmp	has_nul2, 0, 0, eq
+	beq	L(nonascii_loop)
+	b	L(tail)
+
+	/* Load 16 bytes from [srcin & ~15] and force the bytes that precede
+	   srcin to 0x7f, so we ignore any NUL bytes before the string.
+	   Then continue in the aligned loop.  */
+L(page_cross):
+	bic	src, srcin, 15
+	ldp	data1, data2, [src]
+	lsl	tmp1, srcin, 3
+	mov	tmp4, -1
+#ifdef __AARCH64EB__
+	/* Big-endian.	Early bytes are at MSB.	 */
+	lsr	tmp1, tmp4, tmp1	/* Shift (tmp1 & 63).  */
+#else
+	/* Little-endian.  Early bytes are at LSB.  */
+	lsl	tmp1, tmp4, tmp1	/* Shift (tmp1 & 63).  */
+#endif
+	orr	tmp1, tmp1, REP8_80
+	orn	data1, data1, tmp1
+	orn	tmp2, data2, tmp1
+	tst	srcin, 8
+	csel	data1, data1, tmp4, eq
+	csel	data2, data2, tmp2, eq
+	b	L(page_cross_entry)
 
-	orr	data1, data1, tmp2
-	orr	data2a, data2, tmp2
-	csinv	data1, data1, xzr, le
-	csel	data2, data2, data2a, le
-	b	.Lrealigned
 SYM_FUNC_END_PI(strlen)
 EXPORT_SYMBOL_NOKASAN(strlen)