x86/crc32: update prototype for crc_pcl()

- Change the len parameter from unsigned int to size_t, so that the
  library function which takes a size_t can safely use this code.

- Rename to crc32c_x86_3way() which is much clearer.

- Move the crc parameter to the front, as this is the usual convention.

Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20241202010844.144356-12-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>

arch/x86/crypto/crc32c-intel_glue.c

@@ -41,8 +41,7 @@
  */
 #define CRC32C_PCL_BREAKEVEN	512
 
-asmlinkage unsigned int crc_pcl(const u8 *buffer, unsigned int len,
-				unsigned int crc_init);
+asmlinkage u32 crc32c_x86_3way(u32 crc, const u8 *buffer, size_t len);
 #endif /* CONFIG_X86_64 */
 
 static u32 crc32c_intel_le_hw_byte(u32 crc, unsigned char const *data, size_t length)
@@ -159,7 +158,7 @@ static int crc32c_pcl_intel_update(struct shash_desc *desc, const u8 *data,
 	 */
 	if (len >= CRC32C_PCL_BREAKEVEN && crypto_simd_usable()) {
 		kernel_fpu_begin();
-		*crcp = crc_pcl(data, len, *crcp);
+		*crcp = crc32c_x86_3way(*crcp, data, len);
 		kernel_fpu_end();
 	} else
 		*crcp = crc32c_intel_le_hw(*crcp, data, len);
@@ -171,7 +170,7 @@ static int __crc32c_pcl_intel_finup(u32 *crcp, const u8 *data, unsigned int len,
 {
 	if (len >= CRC32C_PCL_BREAKEVEN && crypto_simd_usable()) {
 		kernel_fpu_begin();
-		*(__le32 *)out = ~cpu_to_le32(crc_pcl(data, len, *crcp));
+		*(__le32 *)out = ~cpu_to_le32(crc32c_x86_3way(*crcp, data, len));
 		kernel_fpu_end();
 	} else
 		*(__le32 *)out =

arch/x86/crypto/crc32c-pcl-intel-asm_64.S

@@ -52,15 +52,16 @@
 # regular CRC code that does not interleave the CRC instructions.
 #define SMALL_SIZE 200
 
-# unsigned int crc_pcl(const u8 *buffer, unsigned int len, unsigned int crc_init);
+# u32 crc32c_x86_3way(u32 crc, const u8 *buffer, size_t len);
 
 .text
-SYM_FUNC_START(crc_pcl)
-#define    bufp		  %rdi
-#define    bufp_d	  %edi
-#define    len		  %esi
-#define    crc_init	  %edx
-#define    crc_init_q	  %rdx
+SYM_FUNC_START(crc32c_x86_3way)
+#define    crc0		  %edi
+#define    crc0_q	  %rdi
+#define    bufp		  %rsi
+#define    bufp_d	  %esi
+#define    len		  %rdx
+#define    len_dw	  %edx
 #define    n_misaligned	  %ecx /* overlaps chunk_bytes! */
 #define    n_misaligned_q %rcx
 #define    chunk_bytes	  %ecx /* overlaps n_misaligned! */
@@ -85,9 +86,9 @@ SYM_FUNC_START(crc_pcl)
 .Ldo_align:
 	movq	(bufp), %rax
 	add	n_misaligned_q, bufp
-	sub	n_misaligned, len
+	sub	n_misaligned_q, len
 .Lalign_loop:
-	crc32b	%al, crc_init		# compute crc32 of 1-byte
+	crc32b	%al, crc0		# compute crc32 of 1-byte
 	shr	$8, %rax		# get next byte
 	dec	n_misaligned
 	jne     .Lalign_loop
@@ -102,7 +103,7 @@ SYM_FUNC_START(crc_pcl)
 
 .Lpartial_block:
 	# Compute floor(len / 24) to get num qwords to process from each lane.
-	imul	$2731, len, %eax	# 2731 = ceil(2^16 / 24)
+	imul	$2731, len_dw, %eax	# 2731 = ceil(2^16 / 24)
 	shr	$16, %eax
 	jmp	.Lcrc_3lanes
 
@@ -125,16 +126,16 @@ SYM_FUNC_START(crc_pcl)
 	# Unroll the loop by a factor of 4 to reduce the overhead of the loop
 	# bookkeeping instructions, which can compete with crc32q for the ALUs.
 .Lcrc_3lanes_4x_loop:
-	crc32q	(bufp), crc_init_q
+	crc32q	(bufp), crc0_q
 	crc32q	(bufp,chunk_bytes_q), crc1
 	crc32q	(bufp,chunk_bytes_q,2), crc2
-	crc32q	8(bufp), crc_init_q
+	crc32q	8(bufp), crc0_q
 	crc32q	8(bufp,chunk_bytes_q), crc1
 	crc32q	8(bufp,chunk_bytes_q,2), crc2
-	crc32q	16(bufp), crc_init_q
+	crc32q	16(bufp), crc0_q
 	crc32q	16(bufp,chunk_bytes_q), crc1
 	crc32q	16(bufp,chunk_bytes_q,2), crc2
-	crc32q	24(bufp), crc_init_q
+	crc32q	24(bufp), crc0_q
 	crc32q	24(bufp,chunk_bytes_q), crc1
 	crc32q	24(bufp,chunk_bytes_q,2), crc2
 	add	$32, bufp
@@ -146,15 +147,15 @@ SYM_FUNC_START(crc_pcl)
 	jz	.Lcrc_3lanes_last_qword
 
 .Lcrc_3lanes_1x_loop:
-	crc32q	(bufp), crc_init_q
+	crc32q	(bufp), crc0_q
 	crc32q	(bufp,chunk_bytes_q), crc1
 	crc32q	(bufp,chunk_bytes_q,2), crc2
 	add	$8, bufp
 	dec	%eax
 	jnz	.Lcrc_3lanes_1x_loop
 
 .Lcrc_3lanes_last_qword:
-	crc32q	(bufp), crc_init_q
+	crc32q	(bufp), crc0_q
 	crc32q	(bufp,chunk_bytes_q), crc1
 # SKIP  crc32q	(bufp,chunk_bytes_q,2), crc2	; Don't do this one yet
 
@@ -165,9 +166,9 @@ SYM_FUNC_START(crc_pcl)
 	lea	(K_table-8)(%rip), %rax		# first entry is for idx 1
 	pmovzxdq (%rax,chunk_bytes_q), %xmm0	# 2 consts: K1:K2
 	lea	(chunk_bytes,chunk_bytes,2), %eax # chunk_bytes * 3
-	sub	%eax, len			# len -= chunk_bytes * 3
+	sub	%rax, len			# len -= chunk_bytes * 3
 
-	movq	crc_init_q, %xmm1		# CRC for block 1
+	movq	crc0_q, %xmm1			# CRC for block 1
 	pclmulqdq $0x00, %xmm0, %xmm1		# Multiply by K2
 
 	movq    crc1, %xmm2			# CRC for block 2
@@ -176,8 +177,8 @@ SYM_FUNC_START(crc_pcl)
 	pxor    %xmm2,%xmm1
 	movq    %xmm1, %rax
 	xor	(bufp,chunk_bytes_q,2), %rax
-	mov	crc2, crc_init_q
-	crc32	%rax, crc_init_q
+	mov	crc2, crc0_q
+	crc32	%rax, crc0_q
 	lea	8(bufp,chunk_bytes_q,2), bufp
 
 	################################################################
@@ -193,34 +194,34 @@ SYM_FUNC_START(crc_pcl)
 	## 6) Process any remainder without interleaving:
 	#######################################################################
 .Lsmall:
-	test	len, len
+	test	len_dw, len_dw
 	jz	.Ldone
-	mov	len, %eax
+	mov	len_dw, %eax
 	shr	$3, %eax
 	jz	.Ldo_dword
 .Ldo_qwords:
-	crc32q	(bufp), crc_init_q
+	crc32q	(bufp), crc0_q
 	add	$8, bufp
 	dec	%eax
 	jnz	.Ldo_qwords
 .Ldo_dword:
-	test	$4, len
+	test	$4, len_dw
 	jz	.Ldo_word
-	crc32l	(bufp), crc_init
+	crc32l	(bufp), crc0
 	add	$4, bufp
 .Ldo_word:
-	test	$2, len
+	test	$2, len_dw
 	jz	.Ldo_byte
-	crc32w	(bufp), crc_init
+	crc32w	(bufp), crc0
 	add	$2, bufp
 .Ldo_byte:
-	test	$1, len
+	test	$1, len_dw
 	jz	.Ldone
-	crc32b	(bufp), crc_init
+	crc32b	(bufp), crc0
 .Ldone:
-	mov	crc_init, %eax
+	mov	crc0, %eax
         RET
-SYM_FUNC_END(crc_pcl)
+SYM_FUNC_END(crc32c_x86_3way)
 
 .section	.rodata, "a", @progbits
 	################################################################