[PATCH 5/6] crypto: sha512-mb - Crypto computation (x4 AVX2)

[Megha Dey]

From: Megha Dey 

This patch introduces the assembly routines to do SHA512 computation on
buffers belonging to several jobs at once. The assembly routines are
optimized with AVX2 instructions that have 4 data lanes and using AVX2
registers.

Signed-off-by: Megha Dey 
Reviewed-by: Fenghua Yu 
Reviewed-by: Tim Chen 
---
 arch/x86/crypto/sha512-mb/sha512_x4_avx2.S | 529 +
 1 file changed, 529 insertions(+)
 create mode 100644 arch/x86/crypto/sha512-mb/sha512_x4_avx2.S

diff --git a/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S 
b/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S
new file mode 100644
index 000..31ab1ef
--- /dev/null
+++ b/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S
@@ -0,0 +1,529 @@
+/*
+ * Multi-buffer SHA512 algorithm hash compute routine
+ *
+ * This file is provided under a dual BSD/GPLv2 license.  When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License 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 for more details.
+ *
+ * Contact Information:
+ * Megha Dey 
+ *
+ * BSD LICENSE
+ *
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+# code to compute quad SHA512 using AVX2
+# use YMMs to tackle the larger digest size
+# outer calling routine takes care of save and restore of XMM registers
+# Logic designed/laid out by JDG
+
+# Function clobbers: rax, rcx, rdx, rbx, rsi, rdi, r9-r15; ymm0-15
+# Stack must be aligned to 32 bytes before call
+# Linux clobbers: rax rbx rcx rsi r8 r9 r10 r11 r12
+# Linux preserves: rcx rdx rdi rbp r13 r14 r15
+# clobbers ymm0-15
+
+#include 
+#include "sha512_mb_mgr_datastruct.S"
+
+arg1 = %rdi
+arg2 = %rsi
+
+# Common definitions
+STATE = arg1
+INP_SIZE = arg2
+
+IDX = %rax
+ROUND = %rbx
+TBL = %r8
+
+inp0 = %r9
+inp1 = %r10
+inp2 = %r11
+inp3 = %r12
+
+a = %ymm0
+b = %ymm1
+c = %ymm2
+d = %ymm3
+e = %ymm4
+f = %ymm5
+g = %ymm6
+h = %ymm7
+
+a0 = %ymm8
+a1 = %ymm9
+a2 = %ymm10
+
+TT0 = %ymm14
+TT1 = %ymm13
+TT2 = %ymm12
+TT3 = %ymm11
+TT4 = %ymm10
+TT5 = %ymm9
+
+T1 = %ymm14
+TMP = %ymm15
+
+# Define stack usage
+STACK_SPACE1 = SZ4*16 + NUM_SHA512_DIGEST_WORDS*SZ4 + 24
+
+#define VMOVPD vmovupd
+_digest = SZ4*16
+
+# transpose r0, r1, r2, r3, t0, t1
+# "transpose" data in {r0..r3} using temps {t0..t3}
+# Input looks like: {r0 r1 r2 r3}
+# r0 = {a7 a6 a5 a4 a3 a2 a1 a0}
+# r1 = {b7 b6 b5 b4 b3 b2 b1 b0}
+# r2 = {c7 c6 c5 c4 c3 c2 c1 c0}
+# r3 = {d7 d6 d5 d4 d3 d2 d1 d0}
+#
+# output looks like: {t0 r1 r0 r3}
+# t0 = {d1 d0 c1 c0 b1 b0 a1 a0}
+# r1 = {d3 d2 c3 c2 b3 b2 a3 a2}
+# r0 = {d5 d4 c5 c4 b5 b4 a5 a4}
+# r3 = {d7 d6 c7 c6 b7 b6 a7 a6}
+
+.macro TRANSPOSE r0 r1 r2 r3 t0 t1
+   vshufps  $0x44, \r1, \r0, \t0 # t0 = {b5 b4 a5 a4   b1 b0 a1 a0}
+vshufps  $0xEE, \r1, \r0, \r0 # r0 = {b7 b6 a7 a6   b3 b2 a3 a2}
+vshufps  $0x44, \r3, \r2, \t1 # t1 = {d5 d4 c5 c4   d1 d0 c1 c0}
+vshufps  

[PATCH 5/6] crypto: sha512-mb - Crypto computation (x4 AVX2)

[megha . dey]

From: Megha Dey 

This patch introduces the assembly routines to do SHA512 computation on
buffers belonging to several jobs at once. The assembly routines are
optimized with AVX2 instructions that have 4 data lanes and using AVX2
registers.

Signed-off-by: Megha Dey 
Reviewed-by: Fenghua Yu 
Reviewed-by: Tim Chen 
---
 arch/x86/crypto/sha512-mb/sha512_x4_avx2.S | 518 +
 1 file changed, 518 insertions(+)
 create mode 100644 arch/x86/crypto/sha512-mb/sha512_x4_avx2.S

diff --git a/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S 
b/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S
new file mode 100644
index 000..22cb53f
--- /dev/null
+++ b/arch/x86/crypto/sha512-mb/sha512_x4_avx2.S
@@ -0,0 +1,518 @@
+/*
+ * Multi-buffer SHA512 algorithm hash compute routine
+ *
+ * This file is provided under a dual BSD/GPLv2 license.  When using or
+ * redistributing this file, you may do so under either license.
+ *
+ * GPL LICENSE SUMMARY
+ *
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License 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 for more details.
+ *
+ * Contact Information:
+ * Megha Dey 
+ *
+ * BSD LICENSE
+ *
+ * Copyright(c) 2016 Intel Corporation.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+# code to compute quad SHA512 using AVX2
+# use YMMs to tackle the larger digest size
+# outer calling routine takes care of save and restore of XMM registers
+# Logic designed/laid out by JDG
+
+# Function clobbers: rax, rcx, rdx, rbx, rsi, rdi, r9-r15; ymm0-15
+# Stack must be aligned to 32 bytes before call
+# Linux clobbers: rax rbx rcx rsi r8 r9 r10 r11 r12
+# Linux preserves: rcx rdx rdi rbp r13 r14 r15
+# clobbers ymm0-15
+
+#include 
+#include "sha512_mb_mgr_datastruct.S"
+
+arg1 = %rdi
+arg2 = %rsi
+
+# Common definitions
+STATE = arg1
+INP_SIZE = arg2
+
+IDX = %rax
+ROUND = %rbx
+TBL = %r8
+
+inp0 = %r9
+inp1 = %r10
+inp2 = %r11
+inp3 = %r12
+
+a = %ymm0
+b = %ymm1
+c = %ymm2
+d = %ymm3
+e = %ymm4
+f = %ymm5
+g = %ymm6
+h = %ymm7
+
+a0 = %ymm8
+a1 = %ymm9
+a2 = %ymm10
+
+TT0 = %ymm14
+TT1 = %ymm13
+TT2 = %ymm12
+TT3 = %ymm11
+TT4 = %ymm10
+TT5 = %ymm9
+
+T1 = %ymm14
+TMP = %ymm15
+
+# Define stack usage
+STACK_SPACE1 = SZ4*16 + NUM_SHA512_DIGEST_WORDS*SZ4 + 24
+#define VMOVPD vmovupd
+_digest = SZ4*16
+
+# transpose r0, r1, r2, r3, t0, t1
+# "transpose" data in {r0..r3} using temps {t0..t3}
+# Input looks like: {r0 r1 r2 r3}
+# r0 = {a7 a6 a5 a4 a3 a2 a1 a0}
+# r1 = {b7 b6 b5 b4 b3 b2 b1 b0}
+# r2 = {c7 c6 c5 c4 c3 c2 c1 c0}
+# r3 = {d7 d6 d5 d4 d3 d2 d1 d0}
+#
+# output looks like: {t0 r1 r0 r3}
+# t0 = {d1 d0 c1 c0 b1 b0 a1 a0}
+# r1 = {d3 d2 c3 c2 b3 b2 a3 a2}
+# r0 = {d5 d4 c5 c4 b5 b4 a5 a4}
+# r3 = {d7 d6 c7 c6 b7 b6 a7 a6}
+
+.macro TRANSPOSE r0 r1 r2 r3 t0 t1
+   vshufps  $0x44, \r1, \r0, \t0 # t0 = {b5 b4 a5 a4   b1 b0 a1 a0}
+vshufps  $0xEE, \r1, \r0, \r0 # r0 = {b7 b6 a7 a6   b3 b2 a3 a2}
+vshufps  $0x44, \r3, \r2, \t1 # t1 = {d5 d4 c5 c4   d1 d0 c1 c0}
+vshufps