[PATCH 0/4] arm64: wire CRC32 instructions into core crc32 routines

[PATCH 0/4] arm64: wire CRC32 instructions into core crc32 routines

[Ard Biesheuvel]

There are many crc32 users in the kernel that call the library routine
rather than the crypto API wrapper, and so none of these callers use the
accelerated arm64 instructions when available.

While this is not known to cause performance issues, calling a table based
time variant implementation with a non-negligible D-cache footprint (8 KB)
is wasteful in any case, and now that the crc32 instructions have been made
mandatory in the architecture, let's wire them up into the core crc routines.

This also means that they will be exposed to the crypto API via the generic
CRC32 driver, and so we can remove the scalar routines from the crypto API
driver. This leaves the PMULL code, which will only be useful on systems
that implement 64x64 PMULL but not the CRC32 instructions. Given that no
such systems are known to exist, this driver is removed entirely in patch #4.

Ard Biesheuvel (4):
  lib/crc32: make core crc32() routines weak so they can be overridden
  arm64: cpufeature: add feature for CRC32 instructions
  arm64/lib: add accelerated crc32 routines
  crypto: arm64/crc32 - remove PMULL based CRC32 driver

 arch/arm64/Kconfig                |   1 +
 arch/arm64/configs/defconfig      |   1 -
 arch/arm64/crypto/Kconfig         |   5 -
 arch/arm64/crypto/Makefile        |   3 -
 arch/arm64/crypto/crc32-ce-core.S | 287 --------------------
 arch/arm64/crypto/crc32-ce-glue.c | 244 -----------------
 arch/arm64/include/asm/cpucaps.h  |   3 +-
 arch/arm64/kernel/cpufeature.c    |   9 +
 arch/arm64/lib/Makefile           |   2 +
 arch/arm64/lib/crc32.S            |  60 ++++
 lib/crc32.c                       |  11 +-
 11 files changed, 81 insertions(+), 545 deletions(-)
 delete mode 100644 arch/arm64/crypto/crc32-ce-core.S
 delete mode 100644 arch/arm64/crypto/crc32-ce-glue.c
 create mode 100644 arch/arm64/lib/crc32.S

-- 
2.18.0

[PATCH 1/4] lib/crc32: make core crc32() routines weak so they can be overridden

[Ard Biesheuvel]

Allow architectures to drop in accelerated CRC32 routines by making
the crc32_le/__crc32c_le entry points weak, and exposing non-weak
aliases for them that may be used by the accelerated versions as
fallbacks in case the instructions they rely upon are not available.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 lib/crc32.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/lib/crc32.c b/lib/crc32.c
index a6c9afafc8c8..45b1d67a1767 100644
--- a/lib/crc32.c
+++ b/lib/crc32.c
@@ -183,21 +183,21 @@ static inline u32 __pure crc32_le_generic(u32 crc, unsigned char const *p,
 }
 
 #if CRC_LE_BITS == 1
-u32 __pure crc32_le(u32 crc, unsigned char const *p, size_t len)
+u32 __pure __weak crc32_le(u32 crc, unsigned char const *p, size_t len)
 {
 	return crc32_le_generic(crc, p, len, NULL, CRC32_POLY_LE);
 }
-u32 __pure __crc32c_le(u32 crc, unsigned char const *p, size_t len)
+u32 __pure __weak __crc32c_le(u32 crc, unsigned char const *p, size_t len)
 {
 	return crc32_le_generic(crc, p, len, NULL, CRC32C_POLY_LE);
 }
 #else
-u32 __pure crc32_le(u32 crc, unsigned char const *p, size_t len)
+u32 __pure __weak crc32_le(u32 crc, unsigned char const *p, size_t len)
 {
 	return crc32_le_generic(crc, p, len,
 			(const u32 (*)[256])crc32table_le, CRC32_POLY_LE);
 }
-u32 __pure __crc32c_le(u32 crc, unsigned char const *p, size_t len)
+u32 __pure __weak __crc32c_le(u32 crc, unsigned char const *p, size_t len)
 {
 	return crc32_le_generic(crc, p, len,
 			(const u32 (*)[256])crc32ctable_le, CRC32C_POLY_LE);
@@ -206,6 +206,9 @@ u32 __pure __crc32c_le(u32 crc, unsigned char const *p, size_t len)
 EXPORT_SYMBOL(crc32_le);
 EXPORT_SYMBOL(__crc32c_le);
 
+u32 crc32_le_base(u32, unsigned char const *, size_t) __alias(crc32_le);
+u32 __crc32c_le_base(u32, unsigned char const *, size_t) __alias(__crc32c_le);
+
 /*
  * This multiplies the polynomials x and y modulo the given modulus.
  * This follows the "little-endian" CRC convention that the lsbit
-- 
2.18.0

[PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Ard Biesheuvel]

Add a CRC32 feature bit and wire it up to the CPU id register so we
will be able to use alternatives patching for CRC32 operations.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/arm64/include/asm/cpucaps.h | 3 ++-
 arch/arm64/kernel/cpufeature.c   | 9 +++++++++
 2 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
index ae1f70450fb2..9932aca9704b 100644
--- a/arch/arm64/include/asm/cpucaps.h
+++ b/arch/arm64/include/asm/cpucaps.h
@@ -51,7 +51,8 @@
 #define ARM64_SSBD				30
 #define ARM64_MISMATCHED_CACHE_TYPE		31
 #define ARM64_HAS_STAGE2_FWB			32
+#define ARM64_HAS_CRC32				33
 
-#define ARM64_NCAPS				33
+#define ARM64_NCAPS				34
 
 #endif /* __ASM_CPUCAPS_H */
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index e238b7932096..7626b80128f5 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1222,6 +1222,15 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
 		.cpu_enable = cpu_enable_hw_dbm,
 	},
 #endif
+	{
+		.desc = "CRC32 instructions",
+		.capability = ARM64_HAS_CRC32,
+		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
+		.matches = has_cpuid_feature,
+		.sys_reg = SYS_ID_AA64ISAR0_EL1,
+		.field_pos = ID_AA64ISAR0_CRC32_SHIFT,
+		.min_field_value = 1,
+	},
 	{},
 };
 
-- 
2.18.0

[PATCH 3/4] arm64/lib: add accelerated crc32 routines

[Ard Biesheuvel]

Unlike crc32c(), which is wired up to the crypto API internally so the
optimal driver is selected based on the platform's capabilities,
crc32_le() is implemented as a library function using a slice-by-8 table
based C implementation. Even though few of the call sites may be
bottlenecks, calling a time variant implementation with a non-negligible
D-cache footprint is a bit of a waste, given that ARMv8.1 and up mandates
support for the CRC32 instructions that were optional in ARMv8.0, but are
already widely available, even on the Cortex-A53 based Raspberry Pi.

So implement routines that use these instructions if available, and fall
back to the existing generic routines otherwise. The selection is based
on alternatives patching.

Note that this unconditionally selects CONFIG_CRC32 as a builtin. Since
CRC32 is relied upon by core functionality such as CONFIG_OF_FLATTREE,
this just codifies the status quo.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/arm64/Kconfig      |  1 +
 arch/arm64/lib/Makefile |  2 +
 arch/arm64/lib/crc32.S  | 60 ++++++++++++++++++++
 3 files changed, 63 insertions(+)

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 29e75b47becd..0625355f12fa 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -75,6 +75,7 @@ config ARM64
 	select CLONE_BACKWARDS
 	select COMMON_CLK
 	select CPU_PM if (SUSPEND || CPU_IDLE)
+	select CRC32
 	select DCACHE_WORD_ACCESS
 	select DMA_DIRECT_OPS
 	select EDAC_SUPPORT
diff --git a/arch/arm64/lib/Makefile b/arch/arm64/lib/Makefile
index 68755fd70dcf..f28f91fd96a2 100644
--- a/arch/arm64/lib/Makefile
+++ b/arch/arm64/lib/Makefile
@@ -25,3 +25,5 @@ KCOV_INSTRUMENT_atomic_ll_sc.o	:= n
 UBSAN_SANITIZE_atomic_ll_sc.o	:= n
 
 lib-$(CONFIG_ARCH_HAS_UACCESS_FLUSHCACHE) += uaccess_flushcache.o
+
+obj-$(CONFIG_CRC32) += crc32.o
diff --git a/arch/arm64/lib/crc32.S b/arch/arm64/lib/crc32.S
new file mode 100644
index 000000000000..5bc1e85b4e1c
--- /dev/null
+++ b/arch/arm64/lib/crc32.S
@@ -0,0 +1,60 @@
+/*
+ * Accelerated CRC32(C) using AArch64 CRC instructions
+ *
+ * Copyright (C) 2016 - 2018 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/alternative.h>
+#include <asm/assembler.h>
+
+	.cpu		generic+crc
+
+	.macro		__crc32, c
+0:	subs		x2, x2, #16
+	b.mi		8f
+	ldp		x3, x4, [x1], #16
+CPU_BE(	rev		x3, x3		)
+CPU_BE(	rev		x4, x4		)
+	crc32\c\()x	w0, w0, x3
+	crc32\c\()x	w0, w0, x4
+	b.ne		0b
+	ret
+
+8:	tbz		x2, #3, 4f
+	ldr		x3, [x1], #8
+CPU_BE(	rev		x3, x3		)
+	crc32\c\()x	w0, w0, x3
+4:	tbz		x2, #2, 2f
+	ldr		w3, [x1], #4
+CPU_BE(	rev		w3, w3		)
+	crc32\c\()w	w0, w0, w3
+2:	tbz		x2, #1, 1f
+	ldrh		w3, [x1], #2
+CPU_BE(	rev16		w3, w3		)
+	crc32\c\()h	w0, w0, w3
+1:	tbz		x2, #0, 0f
+	ldrb		w3, [x1]
+	crc32\c\()b	w0, w0, w3
+0:	ret
+	.endm
+
+	.align		5
+ENTRY(crc32_le)
+alternative_if_not ARM64_HAS_CRC32
+	b		crc32_le_base
+alternative_else_nop_endif
+	__crc32
+ENDPROC(crc32_le)
+
+	.align		5
+ENTRY(__crc32c_le)
+alternative_if_not ARM64_HAS_CRC32
+	b		__crc32c_le_base
+alternative_else_nop_endif
+	__crc32		c
+ENDPROC(__crc32c_le)
-- 
2.18.0

[PATCH 4/4] crypto: arm64/crc32 - remove PMULL based CRC32 driver

[Ard Biesheuvel]

Now that the scalar fallbacks have been moved out of this driver into
the core crc32()/crc32c() routines, we are left with a CRC32 crypto API
driver for arm64 that is based only on 64x64 polynomial multiplication,
which is an optional instruction in the ARMv8 architecture, and is less
and less likely to be available on cores that do not also implement the
CRC32 instructions, given that those are mandatory in the architecture
as of ARMv8.1.

Since the scalar instructions do not require the special handling that
SIMD instructions do, and since they turn out to be considerably faster
on some cores (Cortex-A53) as well, there is really no point in keeping
this code around so let's just remove it.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
 arch/arm64/configs/defconfig      |   1 -
 arch/arm64/crypto/Kconfig         |   5 -
 arch/arm64/crypto/Makefile        |   3 -
 arch/arm64/crypto/crc32-ce-core.S | 287 --------------------
 arch/arm64/crypto/crc32-ce-glue.c | 244 -----------------
 5 files changed, 540 deletions(-)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index f67e8d5e93ad..323da306e9f4 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -703,7 +703,6 @@ CONFIG_CRYPTO_SHA3_ARM64=m
 CONFIG_CRYPTO_SM3_ARM64_CE=m
 CONFIG_CRYPTO_GHASH_ARM64_CE=y
 CONFIG_CRYPTO_CRCT10DIF_ARM64_CE=m
-CONFIG_CRYPTO_CRC32_ARM64_CE=m
 CONFIG_CRYPTO_AES_ARM64_CE_CCM=y
 CONFIG_CRYPTO_AES_ARM64_CE_BLK=y
 CONFIG_CRYPTO_CHACHA20_NEON=m
diff --git a/arch/arm64/crypto/Kconfig b/arch/arm64/crypto/Kconfig
index e3fdb0fd6f70..63dc00423ca0 100644
--- a/arch/arm64/crypto/Kconfig
+++ b/arch/arm64/crypto/Kconfig
@@ -66,11 +66,6 @@ config CRYPTO_CRCT10DIF_ARM64_CE
 	depends on KERNEL_MODE_NEON && CRC_T10DIF
 	select CRYPTO_HASH
 
-config CRYPTO_CRC32_ARM64_CE
-	tristate "CRC32 and CRC32C digest algorithms using ARMv8 extensions"
-	depends on CRC32
-	select CRYPTO_HASH
-
 config CRYPTO_AES_ARM64
 	tristate "AES core cipher using scalar instructions"
 	select CRYPTO_AES
diff --git a/arch/arm64/crypto/Makefile b/arch/arm64/crypto/Makefile
index bcafd016618e..776357a3be35 100644
--- a/arch/arm64/crypto/Makefile
+++ b/arch/arm64/crypto/Makefile
@@ -32,9 +32,6 @@ ghash-ce-y := ghash-ce-glue.o ghash-ce-core.o
 obj-$(CONFIG_CRYPTO_CRCT10DIF_ARM64_CE) += crct10dif-ce.o
 crct10dif-ce-y := crct10dif-ce-core.o crct10dif-ce-glue.o
 
-obj-$(CONFIG_CRYPTO_CRC32_ARM64_CE) += crc32-ce.o
-crc32-ce-y:= crc32-ce-core.o crc32-ce-glue.o
-
 obj-$(CONFIG_CRYPTO_AES_ARM64_CE) += aes-ce-cipher.o
 aes-ce-cipher-y := aes-ce-core.o aes-ce-glue.o
 
diff --git a/arch/arm64/crypto/crc32-ce-core.S b/arch/arm64/crypto/crc32-ce-core.S
deleted file mode 100644
index 8061bf0f9c66..000000000000
--- a/arch/arm64/crypto/crc32-ce-core.S
+++ /dev/null
@@ -1,287 +0,0 @@
-/*
- * Accelerated CRC32(C) using arm64 CRC, NEON and Crypto Extensions instructions
- *
- * Copyright (C) 2016 Linaro Ltd <ard.biesheuvel@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-/* GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see http://www.gnu.org/licenses
- *
- * Please  visit http://www.xyratex.com/contact if you need additional
- * information or have any questions.
- *
- * GPL HEADER END
- */
-
-/*
- * Copyright 2012 Xyratex Technology Limited
- *
- * Using hardware provided PCLMULQDQ instruction to accelerate the CRC32
- * calculation.
- * CRC32 polynomial:0x04c11db7(BE)/0xEDB88320(LE)
- * PCLMULQDQ is a new instruction in Intel SSE4.2, the reference can be found
- * at:
- * http://www.intel.com/products/processor/manuals/
- * Intel(R) 64 and IA-32 Architectures Software Developer's Manual
- * Volume 2B: Instruction Set Reference, N-Z
- *
- * Authors:   Gregory Prestas <Gregory_Prestas@us.xyratex.com>
- *	      Alexander Boyko <Alexander_Boyko@xyratex.com>
- */
-
-#include <linux/linkage.h>
-#include <asm/assembler.h>
-
-	.section	".rodata", "a"
-	.align		6
-	.cpu		generic+crypto+crc
-
-.Lcrc32_constants:
-	/*
-	 * [x4*128+32 mod P(x) << 32)]'  << 1   = 0x154442bd4
-	 * #define CONSTANT_R1  0x154442bd4LL
-	 *
-	 * [(x4*128-32 mod P(x) << 32)]' << 1   = 0x1c6e41596
-	 * #define CONSTANT_R2  0x1c6e41596LL
-	 */
-	.octa		0x00000001c6e415960000000154442bd4
-
-	/*
-	 * [(x128+32 mod P(x) << 32)]'   << 1   = 0x1751997d0
-	 * #define CONSTANT_R3  0x1751997d0LL
-	 *
-	 * [(x128-32 mod P(x) << 32)]'   << 1   = 0x0ccaa009e
-	 * #define CONSTANT_R4  0x0ccaa009eLL
-	 */
-	.octa		0x00000000ccaa009e00000001751997d0
-
-	/*
-	 * [(x64 mod P(x) << 32)]'       << 1   = 0x163cd6124
-	 * #define CONSTANT_R5  0x163cd6124LL
-	 */
-	.quad		0x0000000163cd6124
-	.quad		0x00000000FFFFFFFF
-
-	/*
-	 * #define CRCPOLY_TRUE_LE_FULL 0x1DB710641LL
-	 *
-	 * Barrett Reduction constant (u64`) = u` = (x**64 / P(x))`
-	 *                                                      = 0x1F7011641LL
-	 * #define CONSTANT_RU  0x1F7011641LL
-	 */
-	.octa		0x00000001F701164100000001DB710641
-
-.Lcrc32c_constants:
-	.octa		0x000000009e4addf800000000740eef02
-	.octa		0x000000014cd00bd600000000f20c0dfe
-	.quad		0x00000000dd45aab8
-	.quad		0x00000000FFFFFFFF
-	.octa		0x00000000dea713f10000000105ec76f0
-
-	vCONSTANT	.req	v0
-	dCONSTANT	.req	d0
-	qCONSTANT	.req	q0
-
-	BUF		.req	x19
-	LEN		.req	x20
-	CRC		.req	x21
-	CONST		.req	x22
-
-	vzr		.req	v9
-
-	/**
-	 * Calculate crc32
-	 * BUF - buffer
-	 * LEN - sizeof buffer (multiple of 16 bytes), LEN should be > 63
-	 * CRC - initial crc32
-	 * return %eax crc32
-	 * uint crc32_pmull_le(unsigned char const *buffer,
-	 *                     size_t len, uint crc32)
-	 */
-	.text
-ENTRY(crc32_pmull_le)
-	adr_l		x3, .Lcrc32_constants
-	b		0f
-
-ENTRY(crc32c_pmull_le)
-	adr_l		x3, .Lcrc32c_constants
-
-0:	frame_push	4, 64
-
-	mov		BUF, x0
-	mov		LEN, x1
-	mov		CRC, x2
-	mov		CONST, x3
-
-	bic		LEN, LEN, #15
-	ld1		{v1.16b-v4.16b}, [BUF], #0x40
-	movi		vzr.16b, #0
-	fmov		dCONSTANT, CRC
-	eor		v1.16b, v1.16b, vCONSTANT.16b
-	sub		LEN, LEN, #0x40
-	cmp		LEN, #0x40
-	b.lt		less_64
-
-	ldr		qCONSTANT, [CONST]
-
-loop_64:		/* 64 bytes Full cache line folding */
-	sub		LEN, LEN, #0x40
-
-	pmull2		v5.1q, v1.2d, vCONSTANT.2d
-	pmull2		v6.1q, v2.2d, vCONSTANT.2d
-	pmull2		v7.1q, v3.2d, vCONSTANT.2d
-	pmull2		v8.1q, v4.2d, vCONSTANT.2d
-
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	pmull		v2.1q, v2.1d, vCONSTANT.1d
-	pmull		v3.1q, v3.1d, vCONSTANT.1d
-	pmull		v4.1q, v4.1d, vCONSTANT.1d
-
-	eor		v1.16b, v1.16b, v5.16b
-	ld1		{v5.16b}, [BUF], #0x10
-	eor		v2.16b, v2.16b, v6.16b
-	ld1		{v6.16b}, [BUF], #0x10
-	eor		v3.16b, v3.16b, v7.16b
-	ld1		{v7.16b}, [BUF], #0x10
-	eor		v4.16b, v4.16b, v8.16b
-	ld1		{v8.16b}, [BUF], #0x10
-
-	eor		v1.16b, v1.16b, v5.16b
-	eor		v2.16b, v2.16b, v6.16b
-	eor		v3.16b, v3.16b, v7.16b
-	eor		v4.16b, v4.16b, v8.16b
-
-	cmp		LEN, #0x40
-	b.lt		less_64
-
-	if_will_cond_yield_neon
-	stp		q1, q2, [sp, #.Lframe_local_offset]
-	stp		q3, q4, [sp, #.Lframe_local_offset + 32]
-	do_cond_yield_neon
-	ldp		q1, q2, [sp, #.Lframe_local_offset]
-	ldp		q3, q4, [sp, #.Lframe_local_offset + 32]
-	ldr		qCONSTANT, [CONST]
-	movi		vzr.16b, #0
-	endif_yield_neon
-	b		loop_64
-
-less_64:		/* Folding cache line into 128bit */
-	ldr		qCONSTANT, [CONST, #16]
-
-	pmull2		v5.1q, v1.2d, vCONSTANT.2d
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v5.16b
-	eor		v1.16b, v1.16b, v2.16b
-
-	pmull2		v5.1q, v1.2d, vCONSTANT.2d
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v5.16b
-	eor		v1.16b, v1.16b, v3.16b
-
-	pmull2		v5.1q, v1.2d, vCONSTANT.2d
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v5.16b
-	eor		v1.16b, v1.16b, v4.16b
-
-	cbz		LEN, fold_64
-
-loop_16:		/* Folding rest buffer into 128bit */
-	subs		LEN, LEN, #0x10
-
-	ld1		{v2.16b}, [BUF], #0x10
-	pmull2		v5.1q, v1.2d, vCONSTANT.2d
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v5.16b
-	eor		v1.16b, v1.16b, v2.16b
-
-	b.ne		loop_16
-
-fold_64:
-	/* perform the last 64 bit fold, also adds 32 zeroes
-	 * to the input stream */
-	ext		v2.16b, v1.16b, v1.16b, #8
-	pmull2		v2.1q, v2.2d, vCONSTANT.2d
-	ext		v1.16b, v1.16b, vzr.16b, #8
-	eor		v1.16b, v1.16b, v2.16b
-
-	/* final 32-bit fold */
-	ldr		dCONSTANT, [CONST, #32]
-	ldr		d3, [CONST, #40]
-
-	ext		v2.16b, v1.16b, vzr.16b, #4
-	and		v1.16b, v1.16b, v3.16b
-	pmull		v1.1q, v1.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v2.16b
-
-	/* Finish up with the bit-reversed barrett reduction 64 ==> 32 bits */
-	ldr		qCONSTANT, [CONST, #48]
-
-	and		v2.16b, v1.16b, v3.16b
-	ext		v2.16b, vzr.16b, v2.16b, #8
-	pmull2		v2.1q, v2.2d, vCONSTANT.2d
-	and		v2.16b, v2.16b, v3.16b
-	pmull		v2.1q, v2.1d, vCONSTANT.1d
-	eor		v1.16b, v1.16b, v2.16b
-	mov		w0, v1.s[1]
-
-	frame_pop
-	ret
-ENDPROC(crc32_pmull_le)
-ENDPROC(crc32c_pmull_le)
-
-	.macro		__crc32, c
-0:	subs		x2, x2, #16
-	b.mi		8f
-	ldp		x3, x4, [x1], #16
-CPU_BE(	rev		x3, x3		)
-CPU_BE(	rev		x4, x4		)
-	crc32\c\()x	w0, w0, x3
-	crc32\c\()x	w0, w0, x4
-	b.ne		0b
-	ret
-
-8:	tbz		x2, #3, 4f
-	ldr		x3, [x1], #8
-CPU_BE(	rev		x3, x3		)
-	crc32\c\()x	w0, w0, x3
-4:	tbz		x2, #2, 2f
-	ldr		w3, [x1], #4
-CPU_BE(	rev		w3, w3		)
-	crc32\c\()w	w0, w0, w3
-2:	tbz		x2, #1, 1f
-	ldrh		w3, [x1], #2
-CPU_BE(	rev16		w3, w3		)
-	crc32\c\()h	w0, w0, w3
-1:	tbz		x2, #0, 0f
-	ldrb		w3, [x1]
-	crc32\c\()b	w0, w0, w3
-0:	ret
-	.endm
-
-	.align		5
-ENTRY(crc32_armv8_le)
-	__crc32
-ENDPROC(crc32_armv8_le)
-
-	.align		5
-ENTRY(crc32c_armv8_le)
-	__crc32		c
-ENDPROC(crc32c_armv8_le)
diff --git a/arch/arm64/crypto/crc32-ce-glue.c b/arch/arm64/crypto/crc32-ce-glue.c
deleted file mode 100644
index 34b4e3d46aab..000000000000
--- a/arch/arm64/crypto/crc32-ce-glue.c
+++ /dev/null
@@ -1,244 +0,0 @@
-/*
- * Accelerated CRC32(C) using arm64 NEON and Crypto Extensions instructions
- *
- * Copyright (C) 2016 - 2017 Linaro Ltd <ard.biesheuvel@linaro.org>
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-#include <linux/cpufeature.h>
-#include <linux/crc32.h>
-#include <linux/init.h>
-#include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/string.h>
-
-#include <crypto/internal/hash.h>
-
-#include <asm/hwcap.h>
-#include <asm/neon.h>
-#include <asm/simd.h>
-#include <asm/unaligned.h>
-
-#define PMULL_MIN_LEN		64L	/* minimum size of buffer
-					 * for crc32_pmull_le_16 */
-#define SCALE_F			16L	/* size of NEON register */
-
-asmlinkage u32 crc32_pmull_le(const u8 buf[], u64 len, u32 init_crc);
-asmlinkage u32 crc32_armv8_le(u32 init_crc, const u8 buf[], size_t len);
-
-asmlinkage u32 crc32c_pmull_le(const u8 buf[], u64 len, u32 init_crc);
-asmlinkage u32 crc32c_armv8_le(u32 init_crc, const u8 buf[], size_t len);
-
-static u32 (*fallback_crc32)(u32 init_crc, const u8 buf[], size_t len);
-static u32 (*fallback_crc32c)(u32 init_crc, const u8 buf[], size_t len);
-
-static int crc32_pmull_cra_init(struct crypto_tfm *tfm)
-{
-	u32 *key = crypto_tfm_ctx(tfm);
-
-	*key = 0;
-	return 0;
-}
-
-static int crc32c_pmull_cra_init(struct crypto_tfm *tfm)
-{
-	u32 *key = crypto_tfm_ctx(tfm);
-
-	*key = ~0;
-	return 0;
-}
-
-static int crc32_pmull_setkey(struct crypto_shash *hash, const u8 *key,
-			      unsigned int keylen)
-{
-	u32 *mctx = crypto_shash_ctx(hash);
-
-	if (keylen != sizeof(u32)) {
-		crypto_shash_set_flags(hash, CRYPTO_TFM_RES_BAD_KEY_LEN);
-		return -EINVAL;
-	}
-	*mctx = le32_to_cpup((__le32 *)key);
-	return 0;
-}
-
-static int crc32_pmull_init(struct shash_desc *desc)
-{
-	u32 *mctx = crypto_shash_ctx(desc->tfm);
-	u32 *crc = shash_desc_ctx(desc);
-
-	*crc = *mctx;
-	return 0;
-}
-
-static int crc32_update(struct shash_desc *desc, const u8 *data,
-			unsigned int length)
-{
-	u32 *crc = shash_desc_ctx(desc);
-
-	*crc = crc32_armv8_le(*crc, data, length);
-	return 0;
-}
-
-static int crc32c_update(struct shash_desc *desc, const u8 *data,
-			 unsigned int length)
-{
-	u32 *crc = shash_desc_ctx(desc);
-
-	*crc = crc32c_armv8_le(*crc, data, length);
-	return 0;
-}
-
-static int crc32_pmull_update(struct shash_desc *desc, const u8 *data,
-			 unsigned int length)
-{
-	u32 *crc = shash_desc_ctx(desc);
-	unsigned int l;
-
-	if ((u64)data % SCALE_F) {
-		l = min_t(u32, length, SCALE_F - ((u64)data % SCALE_F));
-
-		*crc = fallback_crc32(*crc, data, l);
-
-		data += l;
-		length -= l;
-	}
-
-	if (length >= PMULL_MIN_LEN && may_use_simd()) {
-		l = round_down(length, SCALE_F);
-
-		kernel_neon_begin();
-		*crc = crc32_pmull_le(data, l, *crc);
-		kernel_neon_end();
-
-		data += l;
-		length -= l;
-	}
-
-	if (length > 0)
-		*crc = fallback_crc32(*crc, data, length);
-
-	return 0;
-}
-
-static int crc32c_pmull_update(struct shash_desc *desc, const u8 *data,
-			 unsigned int length)
-{
-	u32 *crc = shash_desc_ctx(desc);
-	unsigned int l;
-
-	if ((u64)data % SCALE_F) {
-		l = min_t(u32, length, SCALE_F - ((u64)data % SCALE_F));
-
-		*crc = fallback_crc32c(*crc, data, l);
-
-		data += l;
-		length -= l;
-	}
-
-	if (length >= PMULL_MIN_LEN && may_use_simd()) {
-		l = round_down(length, SCALE_F);
-
-		kernel_neon_begin();
-		*crc = crc32c_pmull_le(data, l, *crc);
-		kernel_neon_end();
-
-		data += l;
-		length -= l;
-	}
-
-	if (length > 0) {
-		*crc = fallback_crc32c(*crc, data, length);
-	}
-
-	return 0;
-}
-
-static int crc32_pmull_final(struct shash_desc *desc, u8 *out)
-{
-	u32 *crc = shash_desc_ctx(desc);
-
-	put_unaligned_le32(*crc, out);
-	return 0;
-}
-
-static int crc32c_pmull_final(struct shash_desc *desc, u8 *out)
-{
-	u32 *crc = shash_desc_ctx(desc);
-
-	put_unaligned_le32(~*crc, out);
-	return 0;
-}
-
-static struct shash_alg crc32_pmull_algs[] = { {
-	.setkey			= crc32_pmull_setkey,
-	.init			= crc32_pmull_init,
-	.update			= crc32_update,
-	.final			= crc32_pmull_final,
-	.descsize		= sizeof(u32),
-	.digestsize		= sizeof(u32),
-
-	.base.cra_ctxsize	= sizeof(u32),
-	.base.cra_init		= crc32_pmull_cra_init,
-	.base.cra_name		= "crc32",
-	.base.cra_driver_name	= "crc32-arm64-ce",
-	.base.cra_priority	= 200,
-	.base.cra_flags		= CRYPTO_ALG_OPTIONAL_KEY,
-	.base.cra_blocksize	= 1,
-	.base.cra_module	= THIS_MODULE,
-}, {
-	.setkey			= crc32_pmull_setkey,
-	.init			= crc32_pmull_init,
-	.update			= crc32c_update,
-	.final			= crc32c_pmull_final,
-	.descsize		= sizeof(u32),
-	.digestsize		= sizeof(u32),
-
-	.base.cra_ctxsize	= sizeof(u32),
-	.base.cra_init		= crc32c_pmull_cra_init,
-	.base.cra_name		= "crc32c",
-	.base.cra_driver_name	= "crc32c-arm64-ce",
-	.base.cra_priority	= 200,
-	.base.cra_flags		= CRYPTO_ALG_OPTIONAL_KEY,
-	.base.cra_blocksize	= 1,
-	.base.cra_module	= THIS_MODULE,
-} };
-
-static int __init crc32_pmull_mod_init(void)
-{
-	if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) && (elf_hwcap & HWCAP_PMULL)) {
-		crc32_pmull_algs[0].update = crc32_pmull_update;
-		crc32_pmull_algs[1].update = crc32c_pmull_update;
-
-		if (elf_hwcap & HWCAP_CRC32) {
-			fallback_crc32 = crc32_armv8_le;
-			fallback_crc32c = crc32c_armv8_le;
-		} else {
-			fallback_crc32 = crc32_le;
-			fallback_crc32c = __crc32c_le;
-		}
-	} else if (!(elf_hwcap & HWCAP_CRC32)) {
-		return -ENODEV;
-	}
-	return crypto_register_shashes(crc32_pmull_algs,
-				       ARRAY_SIZE(crc32_pmull_algs));
-}
-
-static void __exit crc32_pmull_mod_exit(void)
-{
-	crypto_unregister_shashes(crc32_pmull_algs,
-				  ARRAY_SIZE(crc32_pmull_algs));
-}
-
-static const struct cpu_feature crc32_cpu_feature[] = {
-	{ cpu_feature(CRC32) }, { cpu_feature(PMULL) }, { }
-};
-MODULE_DEVICE_TABLE(cpu, crc32_cpu_feature);
-
-module_init(crc32_pmull_mod_init);
-module_exit(crc32_pmull_mod_exit);
-
-MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
-MODULE_LICENSE("GPL v2");
-- 
2.18.0

Re: [PATCH 0/4] arm64: wire CRC32 instructions into core crc32 routines

[Theodore Y. Ts'o]

On Mon, Aug 27, 2018 at 01:02:41PM +0200, Ard Biesheuvel wrote:
> While this is not known to cause performance issues, calling a table based
> time variant implementation with a non-negligible D-cache footprint (8 KB)
> is wasteful in any case, and now that the crc32 instructions have been made
> mandatory in the architecture, let's wire them up into the core crc routines.

Stupid question --- are there any arm64 SOC's out there which do *not*
have the crc32 instructions?  Presumably there won't be in the future,
because it's now mandatory --- but where there any in the past?

	     	 	       	   	 - Ted

Re: [PATCH 0/4] arm64: wire CRC32 instructions into core crc32 routines

[Ard Biesheuvel]

On 27 August 2018 at 16:53, Theodore Y. Ts'o <tytso@mit.edu> wrote:
> On Mon, Aug 27, 2018 at 01:02:41PM +0200, Ard Biesheuvel wrote:
>> While this is not known to cause performance issues, calling a table based
>> time variant implementation with a non-negligible D-cache footprint (8 KB)
>> is wasteful in any case, and now that the crc32 instructions have been made
>> mandatory in the architecture, let's wire them up into the core crc routines.
>
> Stupid question --- are there any arm64 SOC's out there which do *not*
> have the crc32 instructions?  Presumably there won't be in the future,
> because it's now mandatory --- but where there any in the past?
>

Yes, the APM Xgene for instance.

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Will Deacon]

On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
> Add a CRC32 feature bit and wire it up to the CPU id register so we
> will be able to use alternatives patching for CRC32 operations.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> ---
>  arch/arm64/include/asm/cpucaps.h | 3 ++-
>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
>  2 files changed, 11 insertions(+), 1 deletion(-)

Acked-by: Will Deacon <will.deacon@arm.com>

With the minor caveat below...

> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
> index ae1f70450fb2..9932aca9704b 100644
> --- a/arch/arm64/include/asm/cpucaps.h
> +++ b/arch/arm64/include/asm/cpucaps.h
> @@ -51,7 +51,8 @@
>  #define ARM64_SSBD				30
>  #define ARM64_MISMATCHED_CACHE_TYPE		31
>  #define ARM64_HAS_STAGE2_FWB			32
> +#define ARM64_HAS_CRC32				33
>  
> -#define ARM64_NCAPS				33
> +#define ARM64_NCAPS				34


... if this goes via crypto, you'll almost certainly get a (trivial)
conflict with arm64, since these numbers get bumped all the time.

Will

>  #endif /* __ASM_CPUCAPS_H */
> diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
> index e238b7932096..7626b80128f5 100644
> --- a/arch/arm64/kernel/cpufeature.c
> +++ b/arch/arm64/kernel/cpufeature.c
> @@ -1222,6 +1222,15 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
>  		.cpu_enable = cpu_enable_hw_dbm,
>  	},
>  #endif
> +	{
> +		.desc = "CRC32 instructions",
> +		.capability = ARM64_HAS_CRC32,
> +		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
> +		.matches = has_cpuid_feature,
> +		.sys_reg = SYS_ID_AA64ISAR0_EL1,
> +		.field_pos = ID_AA64ISAR0_CRC32_SHIFT,
> +		.min_field_value = 1,
> +	},
>  	{},
>  };
>  
> -- 
> 2.18.0
> 

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Ard Biesheuvel]

On 28 August 2018 at 19:01, Will Deacon <will.deacon@arm.com> wrote:
> On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
>> Add a CRC32 feature bit and wire it up to the CPU id register so we
>> will be able to use alternatives patching for CRC32 operations.
>>
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> ---
>>  arch/arm64/include/asm/cpucaps.h | 3 ++-
>>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
>>  2 files changed, 11 insertions(+), 1 deletion(-)
>
> Acked-by: Will Deacon <will.deacon@arm.com>
>
> With the minor caveat below...
>
>> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
>> index ae1f70450fb2..9932aca9704b 100644
>> --- a/arch/arm64/include/asm/cpucaps.h
>> +++ b/arch/arm64/include/asm/cpucaps.h
>> @@ -51,7 +51,8 @@
>>  #define ARM64_SSBD                           30
>>  #define ARM64_MISMATCHED_CACHE_TYPE          31
>>  #define ARM64_HAS_STAGE2_FWB                 32
>> +#define ARM64_HAS_CRC32                              33
>>
>> -#define ARM64_NCAPS                          33
>> +#define ARM64_NCAPS                          34
>
>
> ... if this goes via crypto, you'll almost certainly get a (trivial)
> conflict with arm64, since these numbers get bumped all the time.
>

I think the first three patches should go through the arm64 tree. The
last one just removes the now redundant crc32 SIMD driver, and Herbert
could pick that up separately, i.e., it should be totally independent.


>>  #endif /* __ASM_CPUCAPS_H */
>> diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
>> index e238b7932096..7626b80128f5 100644
>> --- a/arch/arm64/kernel/cpufeature.c
>> +++ b/arch/arm64/kernel/cpufeature.c
>> @@ -1222,6 +1222,15 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
>>               .cpu_enable = cpu_enable_hw_dbm,
>>       },
>>  #endif
>> +     {
>> +             .desc = "CRC32 instructions",
>> +             .capability = ARM64_HAS_CRC32,
>> +             .type = ARM64_CPUCAP_SYSTEM_FEATURE,
>> +             .matches = has_cpuid_feature,
>> +             .sys_reg = SYS_ID_AA64ISAR0_EL1,
>> +             .field_pos = ID_AA64ISAR0_CRC32_SHIFT,
>> +             .min_field_value = 1,
>> +     },
>>       {},
>>  };
>>
>> --
>> 2.18.0
>>

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Herbert Xu]

On Tue, Aug 28, 2018 at 08:43:35PM +0200, Ard Biesheuvel wrote:
> On 28 August 2018 at 19:01, Will Deacon <will.deacon@arm.com> wrote:
> > On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
> >> Add a CRC32 feature bit and wire it up to the CPU id register so we
> >> will be able to use alternatives patching for CRC32 operations.
> >>
> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> >> ---
> >>  arch/arm64/include/asm/cpucaps.h | 3 ++-
> >>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
> >>  2 files changed, 11 insertions(+), 1 deletion(-)
> >
> > Acked-by: Will Deacon <will.deacon@arm.com>
> >
> > With the minor caveat below...
> >
> >> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
> >> index ae1f70450fb2..9932aca9704b 100644
> >> --- a/arch/arm64/include/asm/cpucaps.h
> >> +++ b/arch/arm64/include/asm/cpucaps.h
> >> @@ -51,7 +51,8 @@
> >>  #define ARM64_SSBD                           30
> >>  #define ARM64_MISMATCHED_CACHE_TYPE          31
> >>  #define ARM64_HAS_STAGE2_FWB                 32
> >> +#define ARM64_HAS_CRC32                              33
> >>
> >> -#define ARM64_NCAPS                          33
> >> +#define ARM64_NCAPS                          34
> >
> >
> > ... if this goes via crypto, you'll almost certainly get a (trivial)
> > conflict with arm64, since these numbers get bumped all the time.
> >
> 
> I think the first three patches should go through the arm64 tree. The
> last one just removes the now redundant crc32 SIMD driver, and Herbert
> could pick that up separately, i.e., it should be totally independent.

Yes let's do that.

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH 4/4] crypto: arm64/crc32 - remove PMULL based CRC32 driver

[Herbert Xu]

On Mon, Aug 27, 2018 at 01:02:45PM +0200, Ard Biesheuvel wrote:
> Now that the scalar fallbacks have been moved out of this driver into
> the core crc32()/crc32c() routines, we are left with a CRC32 crypto API
> driver for arm64 that is based only on 64x64 polynomial multiplication,
> which is an optional instruction in the ARMv8 architecture, and is less
> and less likely to be available on cores that do not also implement the
> CRC32 instructions, given that those are mandatory in the architecture
> as of ARMv8.1.
> 
> Since the scalar instructions do not require the special handling that
> SIMD instructions do, and since they turn out to be considerably faster
> on some cores (Cortex-A53) as well, there is really no point in keeping
> this code around so let's just remove it.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Will Deacon]

On Tue, Sep 04, 2018 at 11:18:55AM +0800, Herbert Xu wrote:
> On Tue, Aug 28, 2018 at 08:43:35PM +0200, Ard Biesheuvel wrote:
> > On 28 August 2018 at 19:01, Will Deacon <will.deacon@arm.com> wrote:
> > > On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
> > >> Add a CRC32 feature bit and wire it up to the CPU id register so we
> > >> will be able to use alternatives patching for CRC32 operations.
> > >>
> > >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > >> ---
> > >>  arch/arm64/include/asm/cpucaps.h | 3 ++-
> > >>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
> > >>  2 files changed, 11 insertions(+), 1 deletion(-)
> > >
> > > Acked-by: Will Deacon <will.deacon@arm.com>
> > >
> > > With the minor caveat below...
> > >
> > >> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
> > >> index ae1f70450fb2..9932aca9704b 100644
> > >> --- a/arch/arm64/include/asm/cpucaps.h
> > >> +++ b/arch/arm64/include/asm/cpucaps.h
> > >> @@ -51,7 +51,8 @@
> > >>  #define ARM64_SSBD                           30
> > >>  #define ARM64_MISMATCHED_CACHE_TYPE          31
> > >>  #define ARM64_HAS_STAGE2_FWB                 32
> > >> +#define ARM64_HAS_CRC32                              33
> > >>
> > >> -#define ARM64_NCAPS                          33
> > >> +#define ARM64_NCAPS                          34
> > >
> > >
> > > ... if this goes via crypto, you'll almost certainly get a (trivial)
> > > conflict with arm64, since these numbers get bumped all the time.
> > >
> > 
> > I think the first three patches should go through the arm64 tree. The
> > last one just removes the now redundant crc32 SIMD driver, and Herbert
> > could pick that up separately, i.e., it should be totally independent.
> 
> Yes let's do that.

Okey doke! In which case, please can we have your Ack on the first patch?

Cheers,

Will

Re: [PATCH 1/4] lib/crc32: make core crc32() routines weak so they can be overridden

[Herbert Xu]

On Mon, Aug 27, 2018 at 01:02:42PM +0200, Ard Biesheuvel wrote:
> Allow architectures to drop in accelerated CRC32 routines by making
> the crc32_le/__crc32c_le entry points weak, and exposing non-weak
> aliases for them that may be used by the accelerated versions as
> fallbacks in case the instructions they rely upon are not available.
> 
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

Acked-by: Herbert Xu <herbert@gondor.apana.org.au>

Thanks,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Herbert Xu]

On Tue, Sep 04, 2018 at 10:38:45AM +0100, Will Deacon wrote:
> On Tue, Sep 04, 2018 at 11:18:55AM +0800, Herbert Xu wrote:
> > On Tue, Aug 28, 2018 at 08:43:35PM +0200, Ard Biesheuvel wrote:
> > > On 28 August 2018 at 19:01, Will Deacon <will.deacon@arm.com> wrote:
> > > > On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
> > > >> Add a CRC32 feature bit and wire it up to the CPU id register so we
> > > >> will be able to use alternatives patching for CRC32 operations.
> > > >>
> > > >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > > >> ---
> > > >>  arch/arm64/include/asm/cpucaps.h | 3 ++-
> > > >>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
> > > >>  2 files changed, 11 insertions(+), 1 deletion(-)
> > > >
> > > > Acked-by: Will Deacon <will.deacon@arm.com>
> > > >
> > > > With the minor caveat below...
> > > >
> > > >> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
> > > >> index ae1f70450fb2..9932aca9704b 100644
> > > >> --- a/arch/arm64/include/asm/cpucaps.h
> > > >> +++ b/arch/arm64/include/asm/cpucaps.h
> > > >> @@ -51,7 +51,8 @@
> > > >>  #define ARM64_SSBD                           30
> > > >>  #define ARM64_MISMATCHED_CACHE_TYPE          31
> > > >>  #define ARM64_HAS_STAGE2_FWB                 32
> > > >> +#define ARM64_HAS_CRC32                              33
> > > >>
> > > >> -#define ARM64_NCAPS                          33
> > > >> +#define ARM64_NCAPS                          34
> > > >
> > > >
> > > > ... if this goes via crypto, you'll almost certainly get a (trivial)
> > > > conflict with arm64, since these numbers get bumped all the time.
> > > >
> > > 
> > > I think the first three patches should go through the arm64 tree. The
> > > last one just removes the now redundant crc32 SIMD driver, and Herbert
> > > could pick that up separately, i.e., it should be totally independent.
> > 
> > Yes let's do that.
> 
> Okey doke! In which case, please can we have your Ack on the first patch?

Sure, I have just sent an ack for that patch.

Cheers,
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Re: [PATCH 2/4] arm64: cpufeature: add feature for CRC32 instructions

[Catalin Marinas]

On Tue, Sep 04, 2018 at 11:18:55AM +0800, Herbert Xu wrote:
> On Tue, Aug 28, 2018 at 08:43:35PM +0200, Ard Biesheuvel wrote:
> > On 28 August 2018 at 19:01, Will Deacon <will.deacon@arm.com> wrote:
> > > On Mon, Aug 27, 2018 at 01:02:43PM +0200, Ard Biesheuvel wrote:
> > >> Add a CRC32 feature bit and wire it up to the CPU id register so we
> > >> will be able to use alternatives patching for CRC32 operations.
> > >>
> > >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > >> ---
> > >>  arch/arm64/include/asm/cpucaps.h | 3 ++-
> > >>  arch/arm64/kernel/cpufeature.c   | 9 +++++++++
> > >>  2 files changed, 11 insertions(+), 1 deletion(-)
> > >
> > > Acked-by: Will Deacon <will.deacon@arm.com>
> > >
> > > With the minor caveat below...
> > >
> > >> diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
> > >> index ae1f70450fb2..9932aca9704b 100644
> > >> --- a/arch/arm64/include/asm/cpucaps.h
> > >> +++ b/arch/arm64/include/asm/cpucaps.h
> > >> @@ -51,7 +51,8 @@
> > >>  #define ARM64_SSBD                           30
> > >>  #define ARM64_MISMATCHED_CACHE_TYPE          31
> > >>  #define ARM64_HAS_STAGE2_FWB                 32
> > >> +#define ARM64_HAS_CRC32                              33
> > >>
> > >> -#define ARM64_NCAPS                          33
> > >> +#define ARM64_NCAPS                          34
> > >
> > >
> > > ... if this goes via crypto, you'll almost certainly get a (trivial)
> > > conflict with arm64, since these numbers get bumped all the time.
> > >
> > 
> > I think the first three patches should go through the arm64 tree. The
> > last one just removes the now redundant crc32 SIMD driver, and Herbert
> > could pick that up separately, i.e., it should be totally independent.
> 
> Yes let's do that.

I queued the first 3 patches for 4.19. Thanks.

-- 
Catalin