* [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512
@ 2022-11-21 0:39 Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 1/4] crypto: aria: add keystream array into request ctx Taehee Yoo
` (4 more replies)
0 siblings, 5 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 0:39 UTC (permalink / raw)
To: linux-crypto, herbert, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Cc: ap420073
This patchset is to implement aria-avx2 and aria-avx512.
There are some differences between aria-avx, aria-avx2, and aria-avx512,
but they are not core logic(s-box, diffusion layer).
ARIA-AVX2
It supports 32way parallel processing using 256bit registers.
Like ARIA-AVX, it supports both AES-NI based s-box layer algorithm and
GFNI based s-box layer algorithm.
These algorithms are the same as ARIA-AVX except that AES-NI doesn't
support 256bit registers, so it is used twice.
ARIA-AVX512
It supports 64way parallel processing using 512bit registers.
It supports only GFNI based s-box layer algorithm.
Benchmarks with i3-12100
commands: modprobe tcrypt mode=610 num_mb=8192
ARIA-AVX512(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx512) encryption
tcrypt: 1 operation in 1504 cycles (1024 bytes)
tcrypt: 1 operation in 4595 cycles (4096 bytes)
tcrypt: 1 operation in 1763 cycles (1024 bytes)
tcrypt: 1 operation in 5540 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx512) decryption
tcrypt: 1 operation in 1502 cycles (1024 bytes)
tcrypt: 1 operation in 4615 cycles (4096 bytes)
tcrypt: 1 operation in 1759 cycles (1024 bytes)
tcrypt: 1 operation in 5554 cycles (4096 bytes)
ARIA-AVX2 with GFNI(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) encryption
tcrypt: 1 operation in 2003 cycles (1024 bytes)
tcrypt: 1 operation in 5867 cycles (4096 bytes)
tcrypt: 1 operation in 2358 cycles (1024 bytes)
tcrypt: 1 operation in 7295 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) decryption
tcrypt: 1 operation in 2004 cycles (1024 bytes)
tcrypt: 1 operation in 5956 cycles (4096 bytes)
tcrypt: 1 operation in 2409 cycles (1024 bytes)
tcrypt: 1 operation in 7564 cycles (4096 bytes)
ARIA-AVX with GFNI(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx) encryption
tcrypt: 1 operation in 2761 cycles (1024 bytes)
tcrypt: 1 operation in 9390 cycles (4096 bytes)
tcrypt: 1 operation in 3401 cycles (1024 bytes)
tcrypt: 1 operation in 11876 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx) decryption
tcrypt: 1 operation in 2735 cycles (1024 bytes)
tcrypt: 1 operation in 9424 cycles (4096 bytes)
tcrypt: 1 operation in 3369 cycles (1024 bytes)
tcrypt: 1 operation in 11954 cycles (4096 bytes)
This patchset will not successfully be merged before applying CFI fixes
because of conflict.
v6:
- Rebase for "CFI fixes" patchset.
https://lore.kernel.org/linux-crypto/20221118194421.160414-1-ebiggers@kernel.org/T/#t
- Use SYM_TYPED_FUNC_START instead of SYM_FUNC_START.
v5:
- Set CRYPTO_ALG_SKCIPHER_REQSIZE_LARGE flag to avx2, and avx512.
v4:
- Use keystream array in the request ctx.
v3:
- Use ARIA_CTX_enc_key, ARIA_CTX_dec_key, and ARIA_CTX_rounds defines.
v2:
- Add new "add keystream array into struct aria_ctx" patch.
- Use keystream array in the aria_ctx instead of stack memory
Taehee Yoo (4):
crypto: aria: add keystream array into request ctx
crypto: aria: do not use magic number offsets of aria_ctx
crypto: aria: implement aria-avx2
crypto: aria: implement aria-avx512
arch/x86/crypto/Kconfig | 38 +
arch/x86/crypto/Makefile | 6 +
arch/x86/crypto/aria-aesni-avx-asm_64.S | 26 +-
arch/x86/crypto/aria-aesni-avx2-asm_64.S | 1433 +++++++++++++++++++++
arch/x86/crypto/aria-avx.h | 46 +
arch/x86/crypto/aria-gfni-avx512-asm_64.S | 1020 +++++++++++++++
arch/x86/crypto/aria_aesni_avx2_glue.c | 252 ++++
arch/x86/crypto/aria_aesni_avx_glue.c | 45 +-
arch/x86/crypto/aria_gfni_avx512_glue.c | 250 ++++
arch/x86/kernel/asm-offsets.c | 11 +
crypto/aria_generic.c | 4 +
11 files changed, 3103 insertions(+), 28 deletions(-)
create mode 100644 arch/x86/crypto/aria-aesni-avx2-asm_64.S
create mode 100644 arch/x86/crypto/aria-gfni-avx512-asm_64.S
create mode 100644 arch/x86/crypto/aria_aesni_avx2_glue.c
create mode 100644 arch/x86/crypto/aria_gfni_avx512_glue.c
--
2.34.1
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH v6 1/4] crypto: aria: add keystream array into request ctx
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
@ 2022-11-21 0:39 ` Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx Taehee Yoo
` (3 subsequent siblings)
4 siblings, 0 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 0:39 UTC (permalink / raw)
To: linux-crypto, herbert, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Cc: ap420073
avx accelerated aria module used local keystream array.
But, keystream array size is too big.
So, it puts the keystream array into request ctx.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
---
v6:
- No changes.
v5:
- No changes.
v4:
- Add aria_avx_request ctx for keystream array
v3:
- No changes.
v2:
- Patch introduced.
arch/x86/crypto/aria_aesni_avx_glue.c | 39 ++++++++++++++++++---------
1 file changed, 26 insertions(+), 13 deletions(-)
diff --git a/arch/x86/crypto/aria_aesni_avx_glue.c b/arch/x86/crypto/aria_aesni_avx_glue.c
index c561ea4fefa5..5f97e442349f 100644
--- a/arch/x86/crypto/aria_aesni_avx_glue.c
+++ b/arch/x86/crypto/aria_aesni_avx_glue.c
@@ -33,6 +33,10 @@ asmlinkage void aria_aesni_avx_gfni_ctr_crypt_16way(const void *ctx, u8 *dst,
static struct aria_avx_ops aria_ops;
+struct aria_avx_request_ctx {
+ u8 keystream[ARIA_AESNI_PARALLEL_BLOCK_SIZE];
+};
+
static int ecb_do_encrypt(struct skcipher_request *req, const u32 *rkey)
{
ECB_WALK_START(req, ARIA_BLOCK_SIZE, ARIA_AESNI_PARALLEL_BLOCKS);
@@ -73,6 +77,7 @@ static int aria_avx_set_key(struct crypto_skcipher *tfm, const u8 *key,
static int aria_avx_ctr_encrypt(struct skcipher_request *req)
{
+ struct aria_avx_request_ctx *req_ctx = skcipher_request_ctx(req);
struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
struct skcipher_walk walk;
@@ -86,10 +91,9 @@ static int aria_avx_ctr_encrypt(struct skcipher_request *req)
u8 *dst = walk.dst.virt.addr;
while (nbytes >= ARIA_AESNI_PARALLEL_BLOCK_SIZE) {
- u8 keystream[ARIA_AESNI_PARALLEL_BLOCK_SIZE];
-
kernel_fpu_begin();
- aria_ops.aria_ctr_crypt_16way(ctx, dst, src, keystream,
+ aria_ops.aria_ctr_crypt_16way(ctx, dst, src,
+ &req_ctx->keystream[0],
walk.iv);
kernel_fpu_end();
dst += ARIA_AESNI_PARALLEL_BLOCK_SIZE;
@@ -98,28 +102,29 @@ static int aria_avx_ctr_encrypt(struct skcipher_request *req)
}
while (nbytes >= ARIA_BLOCK_SIZE) {
- u8 keystream[ARIA_BLOCK_SIZE];
-
- memcpy(keystream, walk.iv, ARIA_BLOCK_SIZE);
+ memcpy(&req_ctx->keystream[0], walk.iv, ARIA_BLOCK_SIZE);
crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
- aria_encrypt(ctx, keystream, keystream);
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
- crypto_xor_cpy(dst, src, keystream, ARIA_BLOCK_SIZE);
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ ARIA_BLOCK_SIZE);
dst += ARIA_BLOCK_SIZE;
src += ARIA_BLOCK_SIZE;
nbytes -= ARIA_BLOCK_SIZE;
}
if (walk.nbytes == walk.total && nbytes > 0) {
- u8 keystream[ARIA_BLOCK_SIZE];
-
- memcpy(keystream, walk.iv, ARIA_BLOCK_SIZE);
+ memcpy(&req_ctx->keystream[0], walk.iv,
+ ARIA_BLOCK_SIZE);
crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
- aria_encrypt(ctx, keystream, keystream);
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
- crypto_xor_cpy(dst, src, keystream, nbytes);
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ nbytes);
dst += nbytes;
src += nbytes;
nbytes = 0;
@@ -130,6 +135,13 @@ static int aria_avx_ctr_encrypt(struct skcipher_request *req)
return err;
}
+static int aria_avx_init_tfm(struct crypto_skcipher *tfm)
+{
+ crypto_skcipher_set_reqsize(tfm, sizeof(struct aria_avx_request_ctx));
+
+ return 0;
+}
+
static struct skcipher_alg aria_algs[] = {
{
.base.cra_name = "__ecb(aria)",
@@ -160,6 +172,7 @@ static struct skcipher_alg aria_algs[] = {
.setkey = aria_avx_set_key,
.encrypt = aria_avx_ctr_encrypt,
.decrypt = aria_avx_ctr_encrypt,
+ .init = aria_avx_init_tfm,
}
};
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 1/4] crypto: aria: add keystream array into request ctx Taehee Yoo
@ 2022-11-21 0:39 ` Taehee Yoo
2022-12-02 10:11 ` Herbert Xu
2022-11-21 0:39 ` [PATCH v6 3/4] crypto: aria: implement aria-avx2 Taehee Yoo
` (2 subsequent siblings)
4 siblings, 1 reply; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 0:39 UTC (permalink / raw)
To: linux-crypto, herbert, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Cc: ap420073
aria-avx assembly code accesses members of aria_ctx with magic number
offset. If the shape of struct aria_ctx is changed carelessly,
aria-avx will not work.
So, we need to ensure accessing members of aria_ctx with correct
offset values, not with magic numbers.
It adds ARIA_CTX_enc_key, ARIA_CTX_dec_key, and ARIA_CTX_rounds in the
asm-offsets.c So, correct offset definitions will be generated.
aria-avx assembly code can access members of aria_ctx safely with
these definitions.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
---
v6:
- Rebase for "CFI fixes" patchset.
https://lore.kernel.org/linux-crypto/20221118194421.160414-1-ebiggers@kernel.org/T/#t
v5:
- No changes.
v4:
- Add BUILD_BUG_ON() to check size of fields of aria_ctx.
v3:
- Patch introduced.
arch/x86/crypto/aria-aesni-avx-asm_64.S | 26 +++++++++++--------------
arch/x86/kernel/asm-offsets.c | 11 +++++++++++
crypto/aria_generic.c | 4 ++++
3 files changed, 26 insertions(+), 15 deletions(-)
diff --git a/arch/x86/crypto/aria-aesni-avx-asm_64.S b/arch/x86/crypto/aria-aesni-avx-asm_64.S
index 03ae4cd1d976..be6adc6e7458 100644
--- a/arch/x86/crypto/aria-aesni-avx-asm_64.S
+++ b/arch/x86/crypto/aria-aesni-avx-asm_64.S
@@ -8,13 +8,9 @@
#include <linux/linkage.h>
#include <linux/cfi_types.h>
+#include <asm/asm-offsets.h>
#include <asm/frame.h>
-/* struct aria_ctx: */
-#define enc_key 0
-#define dec_key 272
-#define rounds 544
-
/* register macros */
#define CTX %rdi
@@ -874,7 +870,7 @@ SYM_FUNC_START_LOCAL(__aria_aesni_avx_crypt_16way)
aria_fo(%xmm9, %xmm8, %xmm11, %xmm10, %xmm12, %xmm13, %xmm14, %xmm15,
%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%rax, %r9, 10);
- cmpl $12, rounds(CTX);
+ cmpl $12, ARIA_CTX_rounds(CTX);
jne .Laria_192;
aria_ff(%xmm1, %xmm0, %xmm3, %xmm2, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -887,7 +883,7 @@ SYM_FUNC_START_LOCAL(__aria_aesni_avx_crypt_16way)
aria_fo(%xmm9, %xmm8, %xmm11, %xmm10, %xmm12, %xmm13, %xmm14, %xmm15,
%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%rax, %r9, 12);
- cmpl $14, rounds(CTX);
+ cmpl $14, ARIA_CTX_rounds(CTX);
jne .Laria_256;
aria_ff(%xmm1, %xmm0, %xmm3, %xmm2, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -923,7 +919,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_encrypt_16way)
FRAME_BEGIN
- leaq enc_key(CTX), %r9;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
inpack16_pre(%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -948,7 +944,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_decrypt_16way)
FRAME_BEGIN
- leaq dec_key(CTX), %r9;
+ leaq ARIA_CTX_dec_key(CTX), %r9;
inpack16_pre(%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -1056,7 +1052,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_ctr_crypt_16way)
leaq (%rdx), %r11;
leaq (%rcx), %rsi;
leaq (%rcx), %rdx;
- leaq enc_key(CTX), %r9;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
call __aria_aesni_avx_crypt_16way;
@@ -1157,7 +1153,7 @@ SYM_FUNC_START_LOCAL(__aria_aesni_avx_gfni_crypt_16way)
%xmm0, %xmm1, %xmm2, %xmm3,
%xmm4, %xmm5, %xmm6, %xmm7,
%rax, %r9, 10);
- cmpl $12, rounds(CTX);
+ cmpl $12, ARIA_CTX_rounds(CTX);
jne .Laria_gfni_192;
aria_ff_gfni(%xmm1, %xmm0, %xmm3, %xmm2, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -1174,7 +1170,7 @@ SYM_FUNC_START_LOCAL(__aria_aesni_avx_gfni_crypt_16way)
%xmm0, %xmm1, %xmm2, %xmm3,
%xmm4, %xmm5, %xmm6, %xmm7,
%rax, %r9, 12);
- cmpl $14, rounds(CTX);
+ cmpl $14, ARIA_CTX_rounds(CTX);
jne .Laria_gfni_256;
aria_ff_gfni(%xmm1, %xmm0, %xmm3, %xmm2,
%xmm4, %xmm5, %xmm6, %xmm7,
@@ -1218,7 +1214,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_gfni_encrypt_16way)
FRAME_BEGIN
- leaq enc_key(CTX), %r9;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
inpack16_pre(%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -1243,7 +1239,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_gfni_decrypt_16way)
FRAME_BEGIN
- leaq dec_key(CTX), %r9;
+ leaq ARIA_CTX_dec_key(CTX), %r9;
inpack16_pre(%xmm0, %xmm1, %xmm2, %xmm3, %xmm4, %xmm5, %xmm6, %xmm7,
%xmm8, %xmm9, %xmm10, %xmm11, %xmm12, %xmm13, %xmm14,
@@ -1275,7 +1271,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx_gfni_ctr_crypt_16way)
leaq (%rdx), %r11;
leaq (%rcx), %rsi;
leaq (%rcx), %rdx;
- leaq enc_key(CTX), %r9;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
call __aria_aesni_avx_gfni_crypt_16way;
diff --git a/arch/x86/kernel/asm-offsets.c b/arch/x86/kernel/asm-offsets.c
index cb50589a7102..32192a91c65b 100644
--- a/arch/x86/kernel/asm-offsets.c
+++ b/arch/x86/kernel/asm-offsets.c
@@ -7,6 +7,7 @@
#define COMPILE_OFFSETS
#include <linux/crypto.h>
+#include <crypto/aria.h>
#include <linux/sched.h>
#include <linux/stddef.h>
#include <linux/hardirq.h>
@@ -109,6 +110,16 @@ static void __used common(void)
OFFSET(TSS_sp1, tss_struct, x86_tss.sp1);
OFFSET(TSS_sp2, tss_struct, x86_tss.sp2);
+#if defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64) || \
+ defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64_MODULE)
+
+ /* Offset for fields in aria_ctx */
+ BLANK();
+ OFFSET(ARIA_CTX_enc_key, aria_ctx, enc_key);
+ OFFSET(ARIA_CTX_dec_key, aria_ctx, dec_key);
+ OFFSET(ARIA_CTX_rounds, aria_ctx, rounds);
+#endif
+
if (IS_ENABLED(CONFIG_KVM_INTEL)) {
BLANK();
OFFSET(VMX_spec_ctrl, vcpu_vmx, spec_ctrl);
diff --git a/crypto/aria_generic.c b/crypto/aria_generic.c
index 4cc29b82b99d..d96dfc4fdde6 100644
--- a/crypto/aria_generic.c
+++ b/crypto/aria_generic.c
@@ -178,6 +178,10 @@ int aria_set_key(struct crypto_tfm *tfm, const u8 *in_key, unsigned int key_len)
if (key_len != 16 && key_len != 24 && key_len != 32)
return -EINVAL;
+ BUILD_BUG_ON(sizeof(ctx->enc_key) != 272);
+ BUILD_BUG_ON(sizeof(ctx->dec_key) != 272);
+ BUILD_BUG_ON(sizeof(int) != sizeof(ctx->rounds));
+
ctx->key_length = key_len;
ctx->rounds = (key_len + 32) / 4;
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH v6 3/4] crypto: aria: implement aria-avx2
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 1/4] crypto: aria: add keystream array into request ctx Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx Taehee Yoo
@ 2022-11-21 0:39 ` Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 4/4] crypto: aria: implement aria-avx512 Taehee Yoo
2022-11-21 9:47 ` [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Borislav Petkov
4 siblings, 0 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 0:39 UTC (permalink / raw)
To: linux-crypto, herbert, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Cc: ap420073
aria-avx2 implementation uses AVX2, AES-NI, and GFNI.
It supports 32way parallel processing.
So, byteslicing code is changed to support 32way parallel.
And it exports some aria-avx functions such as encrypt() and decrypt().
There are two main logics, s-box layer and diffusion layer.
These codes are the same as aria-avx implementation.
But some instruction are exchanged because they don't support 256bit
registers.
Also, AES-NI doesn't support 256bit register.
So, aesenclast and aesdeclast are used twice like below:
vextracti128 $1, ymm0, xmm6;
vaesenclast xmm7, xmm0, xmm0;
vaesenclast xmm7, xmm6, xmm6;
vinserti128 $1, xmm6, ymm0, ymm0;
Benchmark with modprobe tcrypt mode=610 num_mb=8192, i3-12100:
ARIA-AVX2 with GFNI(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) encryption
tcrypt: 1 operation in 2003 cycles (1024 bytes)
tcrypt: 1 operation in 5867 cycles (4096 bytes)
tcrypt: 1 operation in 2358 cycles (1024 bytes)
tcrypt: 1 operation in 7295 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) decryption
tcrypt: 1 operation in 2004 cycles (1024 bytes)
tcrypt: 1 operation in 5956 cycles (4096 bytes)
tcrypt: 1 operation in 2409 cycles (1024 bytes)
tcrypt: 1 operation in 7564 cycles (4096 bytes)
ARIA-AVX with GFNI(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx) encryption
tcrypt: 1 operation in 2761 cycles (1024 bytes)
tcrypt: 1 operation in 9390 cycles (4096 bytes)
tcrypt: 1 operation in 3401 cycles (1024 bytes)
tcrypt: 1 operation in 11876 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx) decryption
tcrypt: 1 operation in 2735 cycles (1024 bytes)
tcrypt: 1 operation in 9424 cycles (4096 bytes)
tcrypt: 1 operation in 3369 cycles (1024 bytes)
tcrypt: 1 operation in 11954 cycles (4096 bytes)
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
---
v6:
- Use SYM_TYPED_FUNC_START instead of SYM_FUNC_START.
v5:
- Set CRYPTO_ALG_SKCIPHER_REQSIZE_LARGE flag.
v4:
- Add aria_avx2_request_ctx for keystream array.
v3:
- Use ARIA_CTX_enc_key, ARIA_CTX_dec_key, and ARIA_CTX_rounds defines.
v2:
- Add new "add keystream array into struct aria_ctx" patch.
- Use keystream array in the aria_ctx instead of stack memory
arch/x86/crypto/Kconfig | 19 +
arch/x86/crypto/Makefile | 3 +
arch/x86/crypto/aria-aesni-avx2-asm_64.S | 1433 ++++++++++++++++++++++
arch/x86/crypto/aria-avx.h | 38 +
arch/x86/crypto/aria_aesni_avx2_glue.c | 252 ++++
arch/x86/crypto/aria_aesni_avx_glue.c | 6 +
6 files changed, 1751 insertions(+)
create mode 100644 arch/x86/crypto/aria-aesni-avx2-asm_64.S
create mode 100644 arch/x86/crypto/aria_aesni_avx2_glue.c
diff --git a/arch/x86/crypto/Kconfig b/arch/x86/crypto/Kconfig
index 71c4c473d34b..3837ba8b78c5 100644
--- a/arch/x86/crypto/Kconfig
+++ b/arch/x86/crypto/Kconfig
@@ -304,6 +304,25 @@ config CRYPTO_ARIA_AESNI_AVX_X86_64
Processes 16 blocks in parallel.
+config CRYPTO_ARIA_AESNI_AVX2_X86_64
+ tristate "Ciphers: ARIA with modes: ECB, CTR (AES-NI/AVX2/GFNI)"
+ depends on X86 && 64BIT
+ select CRYPTO_SKCIPHER
+ select CRYPTO_SIMD
+ select CRYPTO_ALGAPI
+ select CRYPTO_ARIA
+ select CRYPTO_ARIA_AESNI_AVX_X86_64
+ help
+ Length-preserving cipher: ARIA cipher algorithms
+ (RFC 5794) with ECB and CTR modes
+
+ Architecture: x86_64 using:
+ - AES-NI (AES New Instructions)
+ - AVX2 (Advanced Vector Extensions)
+ - GFNI (Galois Field New Instructions)
+
+ Processes 32 blocks in parallel.
+
config CRYPTO_CHACHA20_X86_64
tristate "Ciphers: ChaCha20, XChaCha20, XChaCha12 (SSSE3/AVX2/AVX-512VL)"
depends on X86 && 64BIT
diff --git a/arch/x86/crypto/Makefile b/arch/x86/crypto/Makefile
index 3b1d701a4f6c..5a349c7a8127 100644
--- a/arch/x86/crypto/Makefile
+++ b/arch/x86/crypto/Makefile
@@ -103,6 +103,9 @@ sm4-aesni-avx2-x86_64-y := sm4-aesni-avx2-asm_64.o sm4_aesni_avx2_glue.o
obj-$(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64) += aria-aesni-avx-x86_64.o
aria-aesni-avx-x86_64-y := aria-aesni-avx-asm_64.o aria_aesni_avx_glue.o
+obj-$(CONFIG_CRYPTO_ARIA_AESNI_AVX2_X86_64) += aria-aesni-avx2-x86_64.o
+aria-aesni-avx2-x86_64-y := aria-aesni-avx2-asm_64.o aria_aesni_avx2_glue.o
+
quiet_cmd_perlasm = PERLASM $@
cmd_perlasm = $(PERL) $< > $@
$(obj)/%.S: $(src)/%.pl FORCE
diff --git a/arch/x86/crypto/aria-aesni-avx2-asm_64.S b/arch/x86/crypto/aria-aesni-avx2-asm_64.S
new file mode 100644
index 000000000000..b6cac9a40f2c
--- /dev/null
+++ b/arch/x86/crypto/aria-aesni-avx2-asm_64.S
@@ -0,0 +1,1433 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * ARIA Cipher 32-way parallel algorithm (AVX2)
+ *
+ * Copyright (c) 2022 Taehee Yoo <ap420073@gmail.com>
+ *
+ */
+
+#include <linux/linkage.h>
+#include <asm/frame.h>
+#include <asm/asm-offsets.h>
+#include <linux/cfi_types.h>
+
+/* register macros */
+#define CTX %rdi
+
+#define ymm0_x xmm0
+#define ymm1_x xmm1
+#define ymm2_x xmm2
+#define ymm3_x xmm3
+#define ymm4_x xmm4
+#define ymm5_x xmm5
+#define ymm6_x xmm6
+#define ymm7_x xmm7
+#define ymm8_x xmm8
+#define ymm9_x xmm9
+#define ymm10_x xmm10
+#define ymm11_x xmm11
+#define ymm12_x xmm12
+#define ymm13_x xmm13
+#define ymm14_x xmm14
+#define ymm15_x xmm15
+
+#define BV8(a0, a1, a2, a3, a4, a5, a6, a7) \
+ ( (((a0) & 1) << 0) | \
+ (((a1) & 1) << 1) | \
+ (((a2) & 1) << 2) | \
+ (((a3) & 1) << 3) | \
+ (((a4) & 1) << 4) | \
+ (((a5) & 1) << 5) | \
+ (((a6) & 1) << 6) | \
+ (((a7) & 1) << 7) )
+
+#define BM8X8(l0, l1, l2, l3, l4, l5, l6, l7) \
+ ( ((l7) << (0 * 8)) | \
+ ((l6) << (1 * 8)) | \
+ ((l5) << (2 * 8)) | \
+ ((l4) << (3 * 8)) | \
+ ((l3) << (4 * 8)) | \
+ ((l2) << (5 * 8)) | \
+ ((l1) << (6 * 8)) | \
+ ((l0) << (7 * 8)) )
+
+#define inc_le128(x, minus_one, tmp) \
+ vpcmpeqq minus_one, x, tmp; \
+ vpsubq minus_one, x, x; \
+ vpslldq $8, tmp, tmp; \
+ vpsubq tmp, x, x;
+
+#define filter_8bit(x, lo_t, hi_t, mask4bit, tmp0) \
+ vpand x, mask4bit, tmp0; \
+ vpandn x, mask4bit, x; \
+ vpsrld $4, x, x; \
+ \
+ vpshufb tmp0, lo_t, tmp0; \
+ vpshufb x, hi_t, x; \
+ vpxor tmp0, x, x;
+
+#define transpose_4x4(x0, x1, x2, x3, t1, t2) \
+ vpunpckhdq x1, x0, t2; \
+ vpunpckldq x1, x0, x0; \
+ \
+ vpunpckldq x3, x2, t1; \
+ vpunpckhdq x3, x2, x2; \
+ \
+ vpunpckhqdq t1, x0, x1; \
+ vpunpcklqdq t1, x0, x0; \
+ \
+ vpunpckhqdq x2, t2, x3; \
+ vpunpcklqdq x2, t2, x2;
+
+#define byteslice_16x16b(a0, b0, c0, d0, \
+ a1, b1, c1, d1, \
+ a2, b2, c2, d2, \
+ a3, b3, c3, d3, \
+ st0, st1) \
+ vmovdqu d2, st0; \
+ vmovdqu d3, st1; \
+ transpose_4x4(a0, a1, a2, a3, d2, d3); \
+ transpose_4x4(b0, b1, b2, b3, d2, d3); \
+ vmovdqu st0, d2; \
+ vmovdqu st1, d3; \
+ \
+ vmovdqu a0, st0; \
+ vmovdqu a1, st1; \
+ transpose_4x4(c0, c1, c2, c3, a0, a1); \
+ transpose_4x4(d0, d1, d2, d3, a0, a1); \
+ \
+ vbroadcasti128 .Lshufb_16x16b, a0; \
+ vmovdqu st1, a1; \
+ vpshufb a0, a2, a2; \
+ vpshufb a0, a3, a3; \
+ vpshufb a0, b0, b0; \
+ vpshufb a0, b1, b1; \
+ vpshufb a0, b2, b2; \
+ vpshufb a0, b3, b3; \
+ vpshufb a0, a1, a1; \
+ vpshufb a0, c0, c0; \
+ vpshufb a0, c1, c1; \
+ vpshufb a0, c2, c2; \
+ vpshufb a0, c3, c3; \
+ vpshufb a0, d0, d0; \
+ vpshufb a0, d1, d1; \
+ vpshufb a0, d2, d2; \
+ vpshufb a0, d3, d3; \
+ vmovdqu d3, st1; \
+ vmovdqu st0, d3; \
+ vpshufb a0, d3, a0; \
+ vmovdqu d2, st0; \
+ \
+ transpose_4x4(a0, b0, c0, d0, d2, d3); \
+ transpose_4x4(a1, b1, c1, d1, d2, d3); \
+ vmovdqu st0, d2; \
+ vmovdqu st1, d3; \
+ \
+ vmovdqu b0, st0; \
+ vmovdqu b1, st1; \
+ transpose_4x4(a2, b2, c2, d2, b0, b1); \
+ transpose_4x4(a3, b3, c3, d3, b0, b1); \
+ vmovdqu st0, b0; \
+ vmovdqu st1, b1; \
+ /* does not adjust output bytes inside vectors */
+
+#define debyteslice_16x16b(a0, b0, c0, d0, \
+ a1, b1, c1, d1, \
+ a2, b2, c2, d2, \
+ a3, b3, c3, d3, \
+ st0, st1) \
+ vmovdqu d2, st0; \
+ vmovdqu d3, st1; \
+ transpose_4x4(a0, a1, a2, a3, d2, d3); \
+ transpose_4x4(b0, b1, b2, b3, d2, d3); \
+ vmovdqu st0, d2; \
+ vmovdqu st1, d3; \
+ \
+ vmovdqu a0, st0; \
+ vmovdqu a1, st1; \
+ transpose_4x4(c0, c1, c2, c3, a0, a1); \
+ transpose_4x4(d0, d1, d2, d3, a0, a1); \
+ \
+ vbroadcasti128 .Lshufb_16x16b, a0; \
+ vmovdqu st1, a1; \
+ vpshufb a0, a2, a2; \
+ vpshufb a0, a3, a3; \
+ vpshufb a0, b0, b0; \
+ vpshufb a0, b1, b1; \
+ vpshufb a0, b2, b2; \
+ vpshufb a0, b3, b3; \
+ vpshufb a0, a1, a1; \
+ vpshufb a0, c0, c0; \
+ vpshufb a0, c1, c1; \
+ vpshufb a0, c2, c2; \
+ vpshufb a0, c3, c3; \
+ vpshufb a0, d0, d0; \
+ vpshufb a0, d1, d1; \
+ vpshufb a0, d2, d2; \
+ vpshufb a0, d3, d3; \
+ vmovdqu d3, st1; \
+ vmovdqu st0, d3; \
+ vpshufb a0, d3, a0; \
+ vmovdqu d2, st0; \
+ \
+ transpose_4x4(c0, d0, a0, b0, d2, d3); \
+ transpose_4x4(c1, d1, a1, b1, d2, d3); \
+ vmovdqu st0, d2; \
+ vmovdqu st1, d3; \
+ \
+ vmovdqu b0, st0; \
+ vmovdqu b1, st1; \
+ transpose_4x4(c2, d2, a2, b2, b0, b1); \
+ transpose_4x4(c3, d3, a3, b3, b0, b1); \
+ vmovdqu st0, b0; \
+ vmovdqu st1, b1; \
+ /* does not adjust output bytes inside vectors */
+
+/* load blocks to registers and apply pre-whitening */
+#define inpack16_pre(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ rio) \
+ vmovdqu (0 * 32)(rio), x0; \
+ vmovdqu (1 * 32)(rio), x1; \
+ vmovdqu (2 * 32)(rio), x2; \
+ vmovdqu (3 * 32)(rio), x3; \
+ vmovdqu (4 * 32)(rio), x4; \
+ vmovdqu (5 * 32)(rio), x5; \
+ vmovdqu (6 * 32)(rio), x6; \
+ vmovdqu (7 * 32)(rio), x7; \
+ vmovdqu (8 * 32)(rio), y0; \
+ vmovdqu (9 * 32)(rio), y1; \
+ vmovdqu (10 * 32)(rio), y2; \
+ vmovdqu (11 * 32)(rio), y3; \
+ vmovdqu (12 * 32)(rio), y4; \
+ vmovdqu (13 * 32)(rio), y5; \
+ vmovdqu (14 * 32)(rio), y6; \
+ vmovdqu (15 * 32)(rio), y7;
+
+/* byteslice pre-whitened blocks and store to temporary memory */
+#define inpack16_post(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_ab, mem_cd) \
+ byteslice_16x16b(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ (mem_ab), (mem_cd)); \
+ \
+ vmovdqu x0, 0 * 32(mem_ab); \
+ vmovdqu x1, 1 * 32(mem_ab); \
+ vmovdqu x2, 2 * 32(mem_ab); \
+ vmovdqu x3, 3 * 32(mem_ab); \
+ vmovdqu x4, 4 * 32(mem_ab); \
+ vmovdqu x5, 5 * 32(mem_ab); \
+ vmovdqu x6, 6 * 32(mem_ab); \
+ vmovdqu x7, 7 * 32(mem_ab); \
+ vmovdqu y0, 0 * 32(mem_cd); \
+ vmovdqu y1, 1 * 32(mem_cd); \
+ vmovdqu y2, 2 * 32(mem_cd); \
+ vmovdqu y3, 3 * 32(mem_cd); \
+ vmovdqu y4, 4 * 32(mem_cd); \
+ vmovdqu y5, 5 * 32(mem_cd); \
+ vmovdqu y6, 6 * 32(mem_cd); \
+ vmovdqu y7, 7 * 32(mem_cd);
+
+#define write_output(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem) \
+ vmovdqu x0, 0 * 32(mem); \
+ vmovdqu x1, 1 * 32(mem); \
+ vmovdqu x2, 2 * 32(mem); \
+ vmovdqu x3, 3 * 32(mem); \
+ vmovdqu x4, 4 * 32(mem); \
+ vmovdqu x5, 5 * 32(mem); \
+ vmovdqu x6, 6 * 32(mem); \
+ vmovdqu x7, 7 * 32(mem); \
+ vmovdqu y0, 8 * 32(mem); \
+ vmovdqu y1, 9 * 32(mem); \
+ vmovdqu y2, 10 * 32(mem); \
+ vmovdqu y3, 11 * 32(mem); \
+ vmovdqu y4, 12 * 32(mem); \
+ vmovdqu y5, 13 * 32(mem); \
+ vmovdqu y6, 14 * 32(mem); \
+ vmovdqu y7, 15 * 32(mem); \
+
+#define aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, idx) \
+ vmovdqu x0, ((idx + 0) * 32)(mem_tmp); \
+ vmovdqu x1, ((idx + 1) * 32)(mem_tmp); \
+ vmovdqu x2, ((idx + 2) * 32)(mem_tmp); \
+ vmovdqu x3, ((idx + 3) * 32)(mem_tmp); \
+ vmovdqu x4, ((idx + 4) * 32)(mem_tmp); \
+ vmovdqu x5, ((idx + 5) * 32)(mem_tmp); \
+ vmovdqu x6, ((idx + 6) * 32)(mem_tmp); \
+ vmovdqu x7, ((idx + 7) * 32)(mem_tmp);
+
+#define aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, idx) \
+ vmovdqu ((idx + 0) * 32)(mem_tmp), x0; \
+ vmovdqu ((idx + 1) * 32)(mem_tmp), x1; \
+ vmovdqu ((idx + 2) * 32)(mem_tmp), x2; \
+ vmovdqu ((idx + 3) * 32)(mem_tmp), x3; \
+ vmovdqu ((idx + 4) * 32)(mem_tmp), x4; \
+ vmovdqu ((idx + 5) * 32)(mem_tmp), x5; \
+ vmovdqu ((idx + 6) * 32)(mem_tmp), x6; \
+ vmovdqu ((idx + 7) * 32)(mem_tmp), x7;
+
+#define aria_ark_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ t0, rk, idx, round) \
+ /* AddRoundKey */ \
+ vpbroadcastb ((round * 16) + idx + 3)(rk), t0; \
+ vpxor t0, x0, x0; \
+ vpbroadcastb ((round * 16) + idx + 2)(rk), t0; \
+ vpxor t0, x1, x1; \
+ vpbroadcastb ((round * 16) + idx + 1)(rk), t0; \
+ vpxor t0, x2, x2; \
+ vpbroadcastb ((round * 16) + idx + 0)(rk), t0; \
+ vpxor t0, x3, x3; \
+ vpbroadcastb ((round * 16) + idx + 7)(rk), t0; \
+ vpxor t0, x4, x4; \
+ vpbroadcastb ((round * 16) + idx + 6)(rk), t0; \
+ vpxor t0, x5, x5; \
+ vpbroadcastb ((round * 16) + idx + 5)(rk), t0; \
+ vpxor t0, x6, x6; \
+ vpbroadcastb ((round * 16) + idx + 4)(rk), t0; \
+ vpxor t0, x7, x7;
+
+#define aria_sbox_8way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ t0, t1, t2, t3, \
+ t4, t5, t6, t7) \
+ vpbroadcastq .Ltf_s2_bitmatrix, t0; \
+ vpbroadcastq .Ltf_inv_bitmatrix, t1; \
+ vpbroadcastq .Ltf_id_bitmatrix, t2; \
+ vpbroadcastq .Ltf_aff_bitmatrix, t3; \
+ vpbroadcastq .Ltf_x2_bitmatrix, t4; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x1, x1; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x5, x5; \
+ vgf2p8affineqb $(tf_inv_const), t1, x2, x2; \
+ vgf2p8affineqb $(tf_inv_const), t1, x6, x6; \
+ vgf2p8affineinvqb $0, t2, x2, x2; \
+ vgf2p8affineinvqb $0, t2, x6, x6; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x0, x0; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x4, x4; \
+ vgf2p8affineqb $(tf_x2_const), t4, x3, x3; \
+ vgf2p8affineqb $(tf_x2_const), t4, x7, x7; \
+ vgf2p8affineinvqb $0, t2, x3, x3; \
+ vgf2p8affineinvqb $0, t2, x7, x7
+
+#define aria_sbox_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ t0, t1, t2, t3, \
+ t4, t5, t6, t7) \
+ vpxor t7, t7, t7; \
+ vpxor t6, t6, t6; \
+ vbroadcasti128 .Linv_shift_row, t0; \
+ vbroadcasti128 .Lshift_row, t1; \
+ vbroadcasti128 .Ltf_lo__inv_aff__and__s2, t2; \
+ vbroadcasti128 .Ltf_hi__inv_aff__and__s2, t3; \
+ vbroadcasti128 .Ltf_lo__x2__and__fwd_aff, t4; \
+ vbroadcasti128 .Ltf_hi__x2__and__fwd_aff, t5; \
+ \
+ vextracti128 $1, x0, t6##_x; \
+ vaesenclast t7##_x, x0##_x, x0##_x; \
+ vaesenclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x0, x0; \
+ \
+ vextracti128 $1, x4, t6##_x; \
+ vaesenclast t7##_x, x4##_x, x4##_x; \
+ vaesenclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x4, x4; \
+ \
+ vextracti128 $1, x1, t6##_x; \
+ vaesenclast t7##_x, x1##_x, x1##_x; \
+ vaesenclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x1, x1; \
+ \
+ vextracti128 $1, x5, t6##_x; \
+ vaesenclast t7##_x, x5##_x, x5##_x; \
+ vaesenclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x5, x5; \
+ \
+ vextracti128 $1, x2, t6##_x; \
+ vaesdeclast t7##_x, x2##_x, x2##_x; \
+ vaesdeclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x2, x2; \
+ \
+ vextracti128 $1, x6, t6##_x; \
+ vaesdeclast t7##_x, x6##_x, x6##_x; \
+ vaesdeclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x6, x6; \
+ \
+ vpbroadcastd .L0f0f0f0f, t6; \
+ \
+ /* AES inverse shift rows */ \
+ vpshufb t0, x0, x0; \
+ vpshufb t0, x4, x4; \
+ vpshufb t0, x1, x1; \
+ vpshufb t0, x5, x5; \
+ vpshufb t1, x3, x3; \
+ vpshufb t1, x7, x7; \
+ vpshufb t1, x2, x2; \
+ vpshufb t1, x6, x6; \
+ \
+ /* affine transformation for S2 */ \
+ filter_8bit(x1, t2, t3, t6, t0); \
+ /* affine transformation for S2 */ \
+ filter_8bit(x5, t2, t3, t6, t0); \
+ \
+ /* affine transformation for X2 */ \
+ filter_8bit(x3, t4, t5, t6, t0); \
+ /* affine transformation for X2 */ \
+ filter_8bit(x7, t4, t5, t6, t0); \
+ \
+ vpxor t6, t6, t6; \
+ vextracti128 $1, x3, t6##_x; \
+ vaesdeclast t7##_x, x3##_x, x3##_x; \
+ vaesdeclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x3, x3; \
+ \
+ vextracti128 $1, x7, t6##_x; \
+ vaesdeclast t7##_x, x7##_x, x7##_x; \
+ vaesdeclast t7##_x, t6##_x, t6##_x; \
+ vinserti128 $1, t6##_x, x7, x7; \
+
+#define aria_diff_m(x0, x1, x2, x3, \
+ t0, t1, t2, t3) \
+ /* T = rotr32(X, 8); */ \
+ /* X ^= T */ \
+ vpxor x0, x3, t0; \
+ vpxor x1, x0, t1; \
+ vpxor x2, x1, t2; \
+ vpxor x3, x2, t3; \
+ /* X = T ^ rotr(X, 16); */ \
+ vpxor t2, x0, x0; \
+ vpxor x1, t3, t3; \
+ vpxor t0, x2, x2; \
+ vpxor t1, x3, x1; \
+ vmovdqu t3, x3;
+
+#define aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7) \
+ /* t1 ^= t2; */ \
+ vpxor y0, x4, x4; \
+ vpxor y1, x5, x5; \
+ vpxor y2, x6, x6; \
+ vpxor y3, x7, x7; \
+ \
+ /* t2 ^= t3; */ \
+ vpxor y4, y0, y0; \
+ vpxor y5, y1, y1; \
+ vpxor y6, y2, y2; \
+ vpxor y7, y3, y3; \
+ \
+ /* t0 ^= t1; */ \
+ vpxor x4, x0, x0; \
+ vpxor x5, x1, x1; \
+ vpxor x6, x2, x2; \
+ vpxor x7, x3, x3; \
+ \
+ /* t3 ^= t1; */ \
+ vpxor x4, y4, y4; \
+ vpxor x5, y5, y5; \
+ vpxor x6, y6, y6; \
+ vpxor x7, y7, y7; \
+ \
+ /* t2 ^= t0; */ \
+ vpxor x0, y0, y0; \
+ vpxor x1, y1, y1; \
+ vpxor x2, y2, y2; \
+ vpxor x3, y3, y3; \
+ \
+ /* t1 ^= t2; */ \
+ vpxor y0, x4, x4; \
+ vpxor y1, x5, x5; \
+ vpxor y2, x6, x6; \
+ vpxor y3, x7, x7;
+
+#define aria_fe(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way(x2, x3, x0, x1, x6, x7, x4, x5, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way(x2, x3, x0, x1, x6, x7, x4, x5, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T3 = ABCD -> BADC \
+ * T3 = y4, y5, y6, y7 -> y5, y4, y7, y6 \
+ * T0 = ABCD -> CDAB \
+ * T0 = x0, x1, x2, x3 -> x2, x3, x0, x1 \
+ * T1 = ABCD -> DCBA \
+ * T1 = x4, x5, x6, x7 -> x7, x6, x5, x4 \
+ */ \
+ aria_diff_word(x2, x3, x0, x1, \
+ x7, x6, x5, x4, \
+ y0, y1, y2, y3, \
+ y5, y4, y7, y6); \
+ aria_store_state_8way(x3, x2, x1, x0, \
+ x6, x7, x4, x5, \
+ mem_tmp, 0);
+
+#define aria_fo(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T1 = ABCD -> BADC \
+ * T1 = x4, x5, x6, x7 -> x5, x4, x7, x6 \
+ * T2 = ABCD -> CDAB \
+ * T2 = y0, y1, y2, y3, -> y2, y3, y0, y1 \
+ * T3 = ABCD -> DCBA \
+ * T3 = y4, y5, y6, y7 -> y7, y6, y5, y4 \
+ */ \
+ aria_diff_word(x0, x1, x2, x3, \
+ x5, x4, x7, x6, \
+ y2, y3, y0, y1, \
+ y7, y6, y5, y4); \
+ aria_store_state_8way(x3, x2, x1, x0, \
+ x6, x7, x4, x5, \
+ mem_tmp, 0);
+
+#define aria_ff(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round, last_round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way(x2, x3, x0, x1, x6, x7, x4, x5, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, last_round); \
+ \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way(x2, x3, x0, x1, x6, x7, x4, x5, \
+ y0, y1, y2, y3, y4, y5, y6, y7); \
+ \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, last_round); \
+ \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8);
+
+#define aria_fe_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T3 = ABCD -> BADC \
+ * T3 = y4, y5, y6, y7 -> y5, y4, y7, y6 \
+ * T0 = ABCD -> CDAB \
+ * T0 = x0, x1, x2, x3 -> x2, x3, x0, x1 \
+ * T1 = ABCD -> DCBA \
+ * T1 = x4, x5, x6, x7 -> x7, x6, x5, x4 \
+ */ \
+ aria_diff_word(x2, x3, x0, x1, \
+ x7, x6, x5, x4, \
+ y0, y1, y2, y3, \
+ y5, y4, y7, y6); \
+ aria_store_state_8way(x3, x2, x1, x0, \
+ x6, x7, x4, x5, \
+ mem_tmp, 0);
+
+#define aria_fo_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, y0, y1, y2, y3); \
+ aria_diff_m(x4, x5, x6, x7, y0, y1, y2, y3); \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T1 = ABCD -> BADC \
+ * T1 = x4, x5, x6, x7 -> x5, x4, x7, x6 \
+ * T2 = ABCD -> CDAB \
+ * T2 = y0, y1, y2, y3, -> y2, y3, y0, y1 \
+ * T3 = ABCD -> DCBA \
+ * T3 = y4, y5, y6, y7 -> y7, y6, y5, y4 \
+ */ \
+ aria_diff_word(x0, x1, x2, x3, \
+ x5, x4, x7, x6, \
+ y2, y3, y0, y1, \
+ y7, y6, y5, y4); \
+ aria_store_state_8way(x3, x2, x1, x0, \
+ x6, x7, x4, x5, \
+ mem_tmp, 0);
+
+#define aria_ff_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, rk, round, last_round) \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, round); \
+ \
+ aria_sbox_8way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 8, last_round); \
+ \
+ aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 8); \
+ \
+ aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, 0); \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, round); \
+ \
+ aria_sbox_8way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ \
+ aria_ark_8way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, rk, 0, last_round); \
+ \
+ aria_load_state_8way(y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_tmp, 8);
+
+.section .rodata.cst32.shufb_16x16b, "aM", @progbits, 32
+.align 32
+#define SHUFB_BYTES(idx) \
+ 0 + (idx), 4 + (idx), 8 + (idx), 12 + (idx)
+.Lshufb_16x16b:
+ .byte SHUFB_BYTES(0), SHUFB_BYTES(1), SHUFB_BYTES(2), SHUFB_BYTES(3)
+ .byte SHUFB_BYTES(0), SHUFB_BYTES(1), SHUFB_BYTES(2), SHUFB_BYTES(3)
+
+.section .rodata.cst16, "aM", @progbits, 16
+.align 16
+/* For isolating SubBytes from AESENCLAST, inverse shift row */
+.Linv_shift_row:
+ .byte 0x00, 0x0d, 0x0a, 0x07, 0x04, 0x01, 0x0e, 0x0b
+ .byte 0x08, 0x05, 0x02, 0x0f, 0x0c, 0x09, 0x06, 0x03
+.Lshift_row:
+ .byte 0x00, 0x05, 0x0a, 0x0f, 0x04, 0x09, 0x0e, 0x03
+ .byte 0x08, 0x0d, 0x02, 0x07, 0x0c, 0x01, 0x06, 0x0b
+/* For CTR-mode IV byteswap */
+.Lbswap128_mask:
+ .byte 0x0f, 0x0e, 0x0d, 0x0c, 0x0b, 0x0a, 0x09, 0x08
+ .byte 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00
+
+/* AES inverse affine and S2 combined:
+ * 1 1 0 0 0 0 0 1 x0 0
+ * 0 1 0 0 1 0 0 0 x1 0
+ * 1 1 0 0 1 1 1 1 x2 0
+ * 0 1 1 0 1 0 0 1 x3 1
+ * 0 1 0 0 1 1 0 0 * x4 + 0
+ * 0 1 0 1 1 0 0 0 x5 0
+ * 0 0 0 0 0 1 0 1 x6 0
+ * 1 1 1 0 0 1 1 1 x7 1
+ */
+.Ltf_lo__inv_aff__and__s2:
+ .octa 0x92172DA81A9FA520B2370D883ABF8500
+.Ltf_hi__inv_aff__and__s2:
+ .octa 0x2B15FFC1AF917B45E6D8320C625CB688
+
+/* X2 and AES forward affine combined:
+ * 1 0 1 1 0 0 0 1 x0 0
+ * 0 1 1 1 1 0 1 1 x1 0
+ * 0 0 0 1 1 0 1 0 x2 1
+ * 0 1 0 0 0 1 0 0 x3 0
+ * 0 0 1 1 1 0 1 1 * x4 + 0
+ * 0 1 0 0 1 0 0 0 x5 0
+ * 1 1 0 1 0 0 1 1 x6 0
+ * 0 1 0 0 1 0 1 0 x7 0
+ */
+.Ltf_lo__x2__and__fwd_aff:
+ .octa 0xEFAE0544FCBD1657B8F95213ABEA4100
+.Ltf_hi__x2__and__fwd_aff:
+ .octa 0x3F893781E95FE1576CDA64D2BA0CB204
+
+.section .rodata.cst8, "aM", @progbits, 8
+.align 8
+/* AES affine: */
+#define tf_aff_const BV8(1, 1, 0, 0, 0, 1, 1, 0)
+.Ltf_aff_bitmatrix:
+ .quad BM8X8(BV8(1, 0, 0, 0, 1, 1, 1, 1),
+ BV8(1, 1, 0, 0, 0, 1, 1, 1),
+ BV8(1, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 1, 0, 0, 0, 1),
+ BV8(1, 1, 1, 1, 1, 0, 0, 0),
+ BV8(0, 1, 1, 1, 1, 1, 0, 0),
+ BV8(0, 0, 1, 1, 1, 1, 1, 0),
+ BV8(0, 0, 0, 1, 1, 1, 1, 1))
+
+/* AES inverse affine: */
+#define tf_inv_const BV8(1, 0, 1, 0, 0, 0, 0, 0)
+.Ltf_inv_bitmatrix:
+ .quad BM8X8(BV8(0, 0, 1, 0, 0, 1, 0, 1),
+ BV8(1, 0, 0, 1, 0, 0, 1, 0),
+ BV8(0, 1, 0, 0, 1, 0, 0, 1),
+ BV8(1, 0, 1, 0, 0, 1, 0, 0),
+ BV8(0, 1, 0, 1, 0, 0, 1, 0),
+ BV8(0, 0, 1, 0, 1, 0, 0, 1),
+ BV8(1, 0, 0, 1, 0, 1, 0, 0),
+ BV8(0, 1, 0, 0, 1, 0, 1, 0))
+
+/* S2: */
+#define tf_s2_const BV8(0, 1, 0, 0, 0, 1, 1, 1)
+.Ltf_s2_bitmatrix:
+ .quad BM8X8(BV8(0, 1, 0, 1, 0, 1, 1, 1),
+ BV8(0, 0, 1, 1, 1, 1, 1, 1),
+ BV8(1, 1, 1, 0, 1, 1, 0, 1),
+ BV8(1, 1, 0, 0, 0, 0, 1, 1),
+ BV8(0, 1, 0, 0, 0, 0, 1, 1),
+ BV8(1, 1, 0, 0, 1, 1, 1, 0),
+ BV8(0, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 1, 0, 1, 1, 0))
+
+/* X2: */
+#define tf_x2_const BV8(0, 0, 1, 1, 0, 1, 0, 0)
+.Ltf_x2_bitmatrix:
+ .quad BM8X8(BV8(0, 0, 0, 1, 1, 0, 0, 0),
+ BV8(0, 0, 1, 0, 0, 1, 1, 0),
+ BV8(0, 0, 0, 0, 1, 0, 1, 0),
+ BV8(1, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 0, 1, 1, 0, 0),
+ BV8(0, 1, 1, 0, 1, 0, 1, 1),
+ BV8(1, 0, 1, 1, 1, 1, 0, 1),
+ BV8(1, 0, 0, 1, 0, 0, 1, 1))
+
+/* Identity matrix: */
+.Ltf_id_bitmatrix:
+ .quad BM8X8(BV8(1, 0, 0, 0, 0, 0, 0, 0),
+ BV8(0, 1, 0, 0, 0, 0, 0, 0),
+ BV8(0, 0, 1, 0, 0, 0, 0, 0),
+ BV8(0, 0, 0, 1, 0, 0, 0, 0),
+ BV8(0, 0, 0, 0, 1, 0, 0, 0),
+ BV8(0, 0, 0, 0, 0, 1, 0, 0),
+ BV8(0, 0, 0, 0, 0, 0, 1, 0),
+ BV8(0, 0, 0, 0, 0, 0, 0, 1))
+
+/* 4-bit mask */
+.section .rodata.cst4.L0f0f0f0f, "aM", @progbits, 4
+.align 4
+.L0f0f0f0f:
+ .long 0x0f0f0f0f
+
+.text
+
+SYM_FUNC_START_LOCAL(__aria_aesni_avx2_crypt_32way)
+ /* input:
+ * %r9: rk
+ * %rsi: dst
+ * %rdx: src
+ * %ymm0..%ymm15: byte-sliced blocks
+ */
+
+ FRAME_BEGIN
+
+ movq %rsi, %rax;
+ leaq 8 * 32(%rax), %r8;
+
+ inpack16_post(%ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r8);
+ aria_fo(%ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 0);
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 1);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 2);
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 3);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 4);
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 5);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 6);
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 7);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 8);
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 9);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 10);
+ cmpl $12, ARIA_CTX_rounds(CTX);
+ jne .Laria_192;
+ aria_ff(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 11, 12);
+ jmp .Laria_end;
+.Laria_192:
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 11);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 12);
+ cmpl $14, ARIA_CTX_rounds(CTX);
+ jne .Laria_256;
+ aria_ff(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 13, 14);
+ jmp .Laria_end;
+.Laria_256:
+ aria_fe(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 13);
+ aria_fo(%ymm9, %ymm8, %ymm11, %ymm10, %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 14);
+ aria_ff(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 15, 16);
+.Laria_end:
+ debyteslice_16x16b(%ymm8, %ymm12, %ymm1, %ymm4,
+ %ymm9, %ymm13, %ymm0, %ymm5,
+ %ymm10, %ymm14, %ymm3, %ymm6,
+ %ymm11, %ymm15, %ymm2, %ymm7,
+ (%rax), (%r8));
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(__aria_aesni_avx2_crypt_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_encrypt_32way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ inpack16_pre(%ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rdx);
+
+ call __aria_aesni_avx2_crypt_32way;
+
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_encrypt_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_decrypt_32way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_dec_key(CTX), %r9;
+
+ inpack16_pre(%ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rdx);
+
+ call __aria_aesni_avx2_crypt_32way;
+
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_decrypt_32way)
+
+SYM_FUNC_START_LOCAL(__aria_aesni_avx2_ctr_gen_keystream_32way)
+ /* input:
+ * %rdi: ctx
+ * %rsi: dst
+ * %rdx: src
+ * %rcx: keystream
+ * %r8: iv (big endian, 128bit)
+ */
+
+ FRAME_BEGIN
+ movq 8(%r8), %r11;
+ bswapq %r11;
+
+ vbroadcasti128 .Lbswap128_mask (%rip), %ymm6;
+ vpcmpeqd %ymm0, %ymm0, %ymm0;
+ vpsrldq $8, %ymm0, %ymm0; /* ab: -1:0 ; cd: -1:0 */
+ vpaddq %ymm0, %ymm0, %ymm5; /* ab: -2:0 ; cd: -2:0 */
+
+ /* load IV and byteswap */
+ vmovdqu (%r8), %xmm7;
+ vpshufb %xmm6, %xmm7, %xmm7;
+ vmovdqa %xmm7, %xmm3;
+ inc_le128(%xmm7, %xmm0, %xmm4);
+ vinserti128 $1, %xmm7, %ymm3, %ymm3;
+ vpshufb %ymm6, %ymm3, %ymm8; /* +1 ; +0 */
+
+ /* check need for handling 64-bit overflow and carry */
+ cmpq $(0xffffffffffffffff - 32), %r11;
+ ja .Lhandle_ctr_carry;
+
+ /* construct IVs */
+ vpsubq %ymm5, %ymm3, %ymm3; /* +3 ; +2 */
+ vpshufb %ymm6, %ymm3, %ymm9;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +5 ; +4 */
+ vpshufb %ymm6, %ymm3, %ymm10;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +7 ; +6 */
+ vpshufb %ymm6, %ymm3, %ymm11;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +9 ; +8 */
+ vpshufb %ymm6, %ymm3, %ymm12;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +11 ; +10 */
+ vpshufb %ymm6, %ymm3, %ymm13;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +13 ; +12 */
+ vpshufb %ymm6, %ymm3, %ymm14;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +15 ; +14 */
+ vpshufb %ymm6, %ymm3, %ymm15;
+ vmovdqu %ymm8, (0 * 32)(%rcx);
+ vmovdqu %ymm9, (1 * 32)(%rcx);
+ vmovdqu %ymm10, (2 * 32)(%rcx);
+ vmovdqu %ymm11, (3 * 32)(%rcx);
+ vmovdqu %ymm12, (4 * 32)(%rcx);
+ vmovdqu %ymm13, (5 * 32)(%rcx);
+ vmovdqu %ymm14, (6 * 32)(%rcx);
+ vmovdqu %ymm15, (7 * 32)(%rcx);
+
+ vpsubq %ymm5, %ymm3, %ymm3; /* +17 ; +16 */
+ vpshufb %ymm6, %ymm3, %ymm8;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +19 ; +18 */
+ vpshufb %ymm6, %ymm3, %ymm9;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +21 ; +20 */
+ vpshufb %ymm6, %ymm3, %ymm10;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +23 ; +22 */
+ vpshufb %ymm6, %ymm3, %ymm11;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +25 ; +24 */
+ vpshufb %ymm6, %ymm3, %ymm12;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +27 ; +26 */
+ vpshufb %ymm6, %ymm3, %ymm13;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +29 ; +28 */
+ vpshufb %ymm6, %ymm3, %ymm14;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +31 ; +30 */
+ vpshufb %ymm6, %ymm3, %ymm15;
+ vpsubq %ymm5, %ymm3, %ymm3; /* +32 */
+ vpshufb %xmm6, %xmm3, %xmm3;
+ vmovdqu %xmm3, (%r8);
+ vmovdqu (0 * 32)(%rcx), %ymm0;
+ vmovdqu (1 * 32)(%rcx), %ymm1;
+ vmovdqu (2 * 32)(%rcx), %ymm2;
+ vmovdqu (3 * 32)(%rcx), %ymm3;
+ vmovdqu (4 * 32)(%rcx), %ymm4;
+ vmovdqu (5 * 32)(%rcx), %ymm5;
+ vmovdqu (6 * 32)(%rcx), %ymm6;
+ vmovdqu (7 * 32)(%rcx), %ymm7;
+ jmp .Lctr_carry_done;
+
+ .Lhandle_ctr_carry:
+ /* construct IVs */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm9; /* +3 ; +2 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm10; /* +5 ; +4 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm11; /* +7 ; +6 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm12; /* +9 ; +8 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm13; /* +11 ; +10 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm14; /* +13 ; +12 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm15; /* +15 ; +14 */
+ vmovdqu %ymm8, (0 * 32)(%rcx);
+ vmovdqu %ymm9, (1 * 32)(%rcx);
+ vmovdqu %ymm10, (2 * 32)(%rcx);
+ vmovdqu %ymm11, (3 * 32)(%rcx);
+ vmovdqu %ymm12, (4 * 32)(%rcx);
+ vmovdqu %ymm13, (5 * 32)(%rcx);
+ vmovdqu %ymm14, (6 * 32)(%rcx);
+ vmovdqu %ymm15, (7 * 32)(%rcx);
+
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm8; /* +17 ; +16 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm9; /* +19 ; +18 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm10; /* +21 ; +20 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm11; /* +23 ; +22 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm12; /* +25 ; +24 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm13; /* +27 ; +26 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm14; /* +29 ; +28 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vpshufb %ymm6, %ymm3, %ymm15; /* +31 ; +30 */
+ inc_le128(%ymm3, %ymm0, %ymm4);
+ vextracti128 $1, %ymm3, %xmm3;
+ vpshufb %xmm6, %xmm3, %xmm3; /* +32 */
+ vmovdqu %xmm3, (%r8);
+ vmovdqu (0 * 32)(%rcx), %ymm0;
+ vmovdqu (1 * 32)(%rcx), %ymm1;
+ vmovdqu (2 * 32)(%rcx), %ymm2;
+ vmovdqu (3 * 32)(%rcx), %ymm3;
+ vmovdqu (4 * 32)(%rcx), %ymm4;
+ vmovdqu (5 * 32)(%rcx), %ymm5;
+ vmovdqu (6 * 32)(%rcx), %ymm6;
+ vmovdqu (7 * 32)(%rcx), %ymm7;
+
+ .Lctr_carry_done:
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(__aria_aesni_avx2_ctr_gen_keystream_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_ctr_crypt_32way)
+ /* input:
+ * %rdi: ctx
+ * %rsi: dst
+ * %rdx: src
+ * %rcx: keystream
+ * %r8: iv (big endian, 128bit)
+ */
+ FRAME_BEGIN
+
+ call __aria_aesni_avx2_ctr_gen_keystream_32way;
+
+ leaq (%rsi), %r10;
+ leaq (%rdx), %r11;
+ leaq (%rcx), %rsi;
+ leaq (%rcx), %rdx;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ call __aria_aesni_avx2_crypt_32way;
+
+ vpxor (0 * 32)(%r11), %ymm1, %ymm1;
+ vpxor (1 * 32)(%r11), %ymm0, %ymm0;
+ vpxor (2 * 32)(%r11), %ymm3, %ymm3;
+ vpxor (3 * 32)(%r11), %ymm2, %ymm2;
+ vpxor (4 * 32)(%r11), %ymm4, %ymm4;
+ vpxor (5 * 32)(%r11), %ymm5, %ymm5;
+ vpxor (6 * 32)(%r11), %ymm6, %ymm6;
+ vpxor (7 * 32)(%r11), %ymm7, %ymm7;
+ vpxor (8 * 32)(%r11), %ymm8, %ymm8;
+ vpxor (9 * 32)(%r11), %ymm9, %ymm9;
+ vpxor (10 * 32)(%r11), %ymm10, %ymm10;
+ vpxor (11 * 32)(%r11), %ymm11, %ymm11;
+ vpxor (12 * 32)(%r11), %ymm12, %ymm12;
+ vpxor (13 * 32)(%r11), %ymm13, %ymm13;
+ vpxor (14 * 32)(%r11), %ymm14, %ymm14;
+ vpxor (15 * 32)(%r11), %ymm15, %ymm15;
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %r10);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_ctr_crypt_32way)
+
+SYM_FUNC_START_LOCAL(__aria_aesni_avx2_gfni_crypt_32way)
+ /* input:
+ * %r9: rk
+ * %rsi: dst
+ * %rdx: src
+ * %ymm0..%ymm15: 16 byte-sliced blocks
+ */
+
+ FRAME_BEGIN
+
+ movq %rsi, %rax;
+ leaq 8 * 32(%rax), %r8;
+
+ inpack16_post(%ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r8);
+ aria_fo_gfni(%ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 0);
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 1);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 2);
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 3);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 4);
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 5);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 6);
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 7);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 8);
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 9);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 10);
+ cmpl $12, ARIA_CTX_rounds(CTX);
+ jne .Laria_gfni_192;
+ aria_ff_gfni(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 11, 12);
+ jmp .Laria_gfni_end;
+.Laria_gfni_192:
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 11);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 12);
+ cmpl $14, ARIA_CTX_rounds(CTX);
+ jne .Laria_gfni_256;
+ aria_ff_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 13, 14);
+ jmp .Laria_gfni_end;
+.Laria_gfni_256:
+ aria_fe_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 13);
+ aria_fo_gfni(%ymm9, %ymm8, %ymm11, %ymm10,
+ %ymm12, %ymm13, %ymm14, %ymm15,
+ %ymm0, %ymm1, %ymm2, %ymm3,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %rax, %r9, 14);
+ aria_ff_gfni(%ymm1, %ymm0, %ymm3, %ymm2,
+ %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11,
+ %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax, %r9, 15, 16);
+.Laria_gfni_end:
+ debyteslice_16x16b(%ymm8, %ymm12, %ymm1, %ymm4,
+ %ymm9, %ymm13, %ymm0, %ymm5,
+ %ymm10, %ymm14, %ymm3, %ymm6,
+ %ymm11, %ymm15, %ymm2, %ymm7,
+ (%rax), (%r8));
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(__aria_aesni_avx2_gfni_crypt_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_encrypt_32way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ inpack16_pre(%ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rdx);
+
+ call __aria_aesni_avx2_gfni_crypt_32way;
+
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_gfni_encrypt_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_decrypt_32way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_dec_key(CTX), %r9;
+
+ inpack16_pre(%ymm0, %ymm1, %ymm2, %ymm3, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rdx);
+
+ call __aria_aesni_avx2_gfni_crypt_32way;
+
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_gfni_decrypt_32way)
+
+SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_ctr_crypt_32way)
+ /* input:
+ * %rdi: ctx
+ * %rsi: dst
+ * %rdx: src
+ * %rcx: keystream
+ * %r8: iv (big endian, 128bit)
+ */
+ FRAME_BEGIN
+
+ call __aria_aesni_avx2_ctr_gen_keystream_32way
+
+ leaq (%rsi), %r10;
+ leaq (%rdx), %r11;
+ leaq (%rcx), %rsi;
+ leaq (%rcx), %rdx;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ call __aria_aesni_avx2_gfni_crypt_32way;
+
+ vpxor (0 * 32)(%r11), %ymm1, %ymm1;
+ vpxor (1 * 32)(%r11), %ymm0, %ymm0;
+ vpxor (2 * 32)(%r11), %ymm3, %ymm3;
+ vpxor (3 * 32)(%r11), %ymm2, %ymm2;
+ vpxor (4 * 32)(%r11), %ymm4, %ymm4;
+ vpxor (5 * 32)(%r11), %ymm5, %ymm5;
+ vpxor (6 * 32)(%r11), %ymm6, %ymm6;
+ vpxor (7 * 32)(%r11), %ymm7, %ymm7;
+ vpxor (8 * 32)(%r11), %ymm8, %ymm8;
+ vpxor (9 * 32)(%r11), %ymm9, %ymm9;
+ vpxor (10 * 32)(%r11), %ymm10, %ymm10;
+ vpxor (11 * 32)(%r11), %ymm11, %ymm11;
+ vpxor (12 * 32)(%r11), %ymm12, %ymm12;
+ vpxor (13 * 32)(%r11), %ymm13, %ymm13;
+ vpxor (14 * 32)(%r11), %ymm14, %ymm14;
+ vpxor (15 * 32)(%r11), %ymm15, %ymm15;
+ write_output(%ymm1, %ymm0, %ymm3, %ymm2, %ymm4, %ymm5, %ymm6, %ymm7,
+ %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
+ %ymm15, %r10);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_aesni_avx2_gfni_ctr_crypt_32way)
diff --git a/arch/x86/crypto/aria-avx.h b/arch/x86/crypto/aria-avx.h
index 01e9a01dc157..b997c4888fb7 100644
--- a/arch/x86/crypto/aria-avx.h
+++ b/arch/x86/crypto/aria-avx.h
@@ -7,10 +7,48 @@
#define ARIA_AESNI_PARALLEL_BLOCKS 16
#define ARIA_AESNI_PARALLEL_BLOCK_SIZE (ARIA_BLOCK_SIZE * 16)
+#define ARIA_AESNI_AVX2_PARALLEL_BLOCKS 32
+#define ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE (ARIA_BLOCK_SIZE * 32)
+
+asmlinkage void aria_aesni_avx_encrypt_16way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx_decrypt_16way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx_ctr_crypt_16way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+asmlinkage void aria_aesni_avx_gfni_encrypt_16way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx_gfni_decrypt_16way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx_gfni_ctr_crypt_16way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+
+asmlinkage void aria_aesni_avx2_encrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx2_decrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx2_ctr_crypt_32way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+asmlinkage void aria_aesni_avx2_gfni_encrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx2_gfni_decrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_aesni_avx2_gfni_ctr_crypt_32way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+
struct aria_avx_ops {
void (*aria_encrypt_16way)(const void *ctx, u8 *dst, const u8 *src);
void (*aria_decrypt_16way)(const void *ctx, u8 *dst, const u8 *src);
void (*aria_ctr_crypt_16way)(const void *ctx, u8 *dst, const u8 *src,
u8 *keystream, u8 *iv);
+ void (*aria_encrypt_32way)(const void *ctx, u8 *dst, const u8 *src);
+ void (*aria_decrypt_32way)(const void *ctx, u8 *dst, const u8 *src);
+ void (*aria_ctr_crypt_32way)(const void *ctx, u8 *dst, const u8 *src,
+ u8 *keystream, u8 *iv);
+
};
#endif
diff --git a/arch/x86/crypto/aria_aesni_avx2_glue.c b/arch/x86/crypto/aria_aesni_avx2_glue.c
new file mode 100644
index 000000000000..95fccc6dc420
--- /dev/null
+++ b/arch/x86/crypto/aria_aesni_avx2_glue.c
@@ -0,0 +1,252 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Glue Code for the AVX2/AES-NI/GFNI assembler implementation of the ARIA Cipher
+ *
+ * Copyright (c) 2022 Taehee Yoo <ap420073@gmail.com>
+ */
+
+#include <crypto/algapi.h>
+#include <crypto/internal/simd.h>
+#include <crypto/aria.h>
+#include <linux/crypto.h>
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/types.h>
+
+#include "ecb_cbc_helpers.h"
+#include "aria-avx.h"
+
+asmlinkage void aria_aesni_avx2_encrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_encrypt_32way);
+asmlinkage void aria_aesni_avx2_decrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_decrypt_32way);
+asmlinkage void aria_aesni_avx2_ctr_crypt_32way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_ctr_crypt_32way);
+asmlinkage void aria_aesni_avx2_gfni_encrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_gfni_encrypt_32way);
+asmlinkage void aria_aesni_avx2_gfni_decrypt_32way(const void *ctx, u8 *dst,
+ const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_gfni_decrypt_32way);
+asmlinkage void aria_aesni_avx2_gfni_ctr_crypt_32way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+EXPORT_SYMBOL_GPL(aria_aesni_avx2_gfni_ctr_crypt_32way);
+
+static struct aria_avx_ops aria_ops;
+
+struct aria_avx2_request_ctx {
+ u8 keystream[ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE];
+};
+
+static int ecb_do_encrypt(struct skcipher_request *req, const u32 *rkey)
+{
+ ECB_WALK_START(req, ARIA_BLOCK_SIZE, ARIA_AESNI_PARALLEL_BLOCKS);
+ ECB_BLOCK(ARIA_AESNI_AVX2_PARALLEL_BLOCKS, aria_ops.aria_encrypt_32way);
+ ECB_BLOCK(ARIA_AESNI_PARALLEL_BLOCKS, aria_ops.aria_encrypt_16way);
+ ECB_BLOCK(1, aria_encrypt);
+ ECB_WALK_END();
+}
+
+static int ecb_do_decrypt(struct skcipher_request *req, const u32 *rkey)
+{
+ ECB_WALK_START(req, ARIA_BLOCK_SIZE, ARIA_AESNI_PARALLEL_BLOCKS);
+ ECB_BLOCK(ARIA_AESNI_AVX2_PARALLEL_BLOCKS, aria_ops.aria_decrypt_32way);
+ ECB_BLOCK(ARIA_AESNI_PARALLEL_BLOCKS, aria_ops.aria_decrypt_16way);
+ ECB_BLOCK(1, aria_decrypt);
+ ECB_WALK_END();
+}
+
+static int aria_avx2_ecb_encrypt(struct skcipher_request *req)
+{
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+ return ecb_do_encrypt(req, ctx->enc_key[0]);
+}
+
+static int aria_avx2_ecb_decrypt(struct skcipher_request *req)
+{
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+ return ecb_do_decrypt(req, ctx->dec_key[0]);
+}
+
+static int aria_avx2_set_key(struct crypto_skcipher *tfm, const u8 *key,
+ unsigned int keylen)
+{
+ return aria_set_key(&tfm->base, key, keylen);
+}
+
+static int aria_avx2_ctr_encrypt(struct skcipher_request *req)
+{
+ struct aria_avx2_request_ctx *req_ctx = skcipher_request_ctx(req);
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+ struct skcipher_walk walk;
+ unsigned int nbytes;
+ int err;
+
+ err = skcipher_walk_virt(&walk, req, false);
+
+ while ((nbytes = walk.nbytes) > 0) {
+ const u8 *src = walk.src.virt.addr;
+ u8 *dst = walk.dst.virt.addr;
+
+ while (nbytes >= ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE) {
+ kernel_fpu_begin();
+ aria_ops.aria_ctr_crypt_32way(ctx, dst, src,
+ &req_ctx->keystream[0],
+ walk.iv);
+ kernel_fpu_end();
+ dst += ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ src += ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ nbytes -= ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ }
+
+ while (nbytes >= ARIA_AESNI_PARALLEL_BLOCK_SIZE) {
+ kernel_fpu_begin();
+ aria_ops.aria_ctr_crypt_16way(ctx, dst, src,
+ &req_ctx->keystream[0],
+ walk.iv);
+ kernel_fpu_end();
+ dst += ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ src += ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ nbytes -= ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ }
+
+ while (nbytes >= ARIA_BLOCK_SIZE) {
+ memcpy(&req_ctx->keystream[0], walk.iv, ARIA_BLOCK_SIZE);
+ crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
+
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
+
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ ARIA_BLOCK_SIZE);
+ dst += ARIA_BLOCK_SIZE;
+ src += ARIA_BLOCK_SIZE;
+ nbytes -= ARIA_BLOCK_SIZE;
+ }
+
+ if (walk.nbytes == walk.total && nbytes > 0) {
+ memcpy(&req_ctx->keystream[0], walk.iv,
+ ARIA_BLOCK_SIZE);
+ crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
+
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
+
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ nbytes);
+ dst += nbytes;
+ src += nbytes;
+ nbytes = 0;
+ }
+ err = skcipher_walk_done(&walk, nbytes);
+ }
+
+ return err;
+}
+
+static int aria_avx2_init_tfm(struct crypto_skcipher *tfm)
+{
+ crypto_skcipher_set_reqsize(tfm, sizeof(struct aria_avx2_request_ctx));
+
+ return 0;
+}
+
+static struct skcipher_alg aria_algs[] = {
+ {
+ .base.cra_name = "__ecb(aria)",
+ .base.cra_driver_name = "__ecb-aria-avx2",
+ .base.cra_priority = 500,
+ .base.cra_flags = CRYPTO_ALG_INTERNAL,
+ .base.cra_blocksize = ARIA_BLOCK_SIZE,
+ .base.cra_ctxsize = sizeof(struct aria_ctx),
+ .base.cra_module = THIS_MODULE,
+ .min_keysize = ARIA_MIN_KEY_SIZE,
+ .max_keysize = ARIA_MAX_KEY_SIZE,
+ .setkey = aria_avx2_set_key,
+ .encrypt = aria_avx2_ecb_encrypt,
+ .decrypt = aria_avx2_ecb_decrypt,
+ }, {
+ .base.cra_name = "__ctr(aria)",
+ .base.cra_driver_name = "__ctr-aria-avx2",
+ .base.cra_priority = 500,
+ .base.cra_flags = CRYPTO_ALG_INTERNAL |
+ CRYPTO_ALG_SKCIPHER_REQSIZE_LARGE,
+ .base.cra_blocksize = 1,
+ .base.cra_ctxsize = sizeof(struct aria_ctx),
+ .base.cra_module = THIS_MODULE,
+ .min_keysize = ARIA_MIN_KEY_SIZE,
+ .max_keysize = ARIA_MAX_KEY_SIZE,
+ .ivsize = ARIA_BLOCK_SIZE,
+ .chunksize = ARIA_BLOCK_SIZE,
+ .setkey = aria_avx2_set_key,
+ .encrypt = aria_avx2_ctr_encrypt,
+ .decrypt = aria_avx2_ctr_encrypt,
+ .init = aria_avx2_init_tfm,
+ }
+};
+
+static struct simd_skcipher_alg *aria_simd_algs[ARRAY_SIZE(aria_algs)];
+
+static int __init aria_avx2_init(void)
+{
+ const char *feature_name;
+
+ if (!boot_cpu_has(X86_FEATURE_AVX) ||
+ !boot_cpu_has(X86_FEATURE_AVX2) ||
+ !boot_cpu_has(X86_FEATURE_AES) ||
+ !boot_cpu_has(X86_FEATURE_OSXSAVE)) {
+ pr_info("AVX2 or AES-NI instructions are not detected.\n");
+ return -ENODEV;
+ }
+
+ if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM,
+ &feature_name)) {
+ pr_info("CPU feature '%s' is not supported.\n", feature_name);
+ return -ENODEV;
+ }
+
+ if (boot_cpu_has(X86_FEATURE_GFNI)) {
+ aria_ops.aria_encrypt_16way = aria_aesni_avx_gfni_encrypt_16way;
+ aria_ops.aria_decrypt_16way = aria_aesni_avx_gfni_decrypt_16way;
+ aria_ops.aria_ctr_crypt_16way = aria_aesni_avx_gfni_ctr_crypt_16way;
+ aria_ops.aria_encrypt_32way = aria_aesni_avx2_gfni_encrypt_32way;
+ aria_ops.aria_decrypt_32way = aria_aesni_avx2_gfni_decrypt_32way;
+ aria_ops.aria_ctr_crypt_32way = aria_aesni_avx2_gfni_ctr_crypt_32way;
+ } else {
+ aria_ops.aria_encrypt_16way = aria_aesni_avx_encrypt_16way;
+ aria_ops.aria_decrypt_16way = aria_aesni_avx_decrypt_16way;
+ aria_ops.aria_ctr_crypt_16way = aria_aesni_avx_ctr_crypt_16way;
+ aria_ops.aria_encrypt_32way = aria_aesni_avx2_encrypt_32way;
+ aria_ops.aria_decrypt_32way = aria_aesni_avx2_decrypt_32way;
+ aria_ops.aria_ctr_crypt_32way = aria_aesni_avx2_ctr_crypt_32way;
+ }
+
+ return simd_register_skciphers_compat(aria_algs,
+ ARRAY_SIZE(aria_algs),
+ aria_simd_algs);
+}
+
+static void __exit aria_avx2_exit(void)
+{
+ simd_unregister_skciphers(aria_algs, ARRAY_SIZE(aria_algs),
+ aria_simd_algs);
+}
+
+module_init(aria_avx2_init);
+module_exit(aria_avx2_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Taehee Yoo <ap420073@gmail.com>");
+MODULE_DESCRIPTION("ARIA Cipher Algorithm, AVX2/AES-NI/GFNI optimized");
+MODULE_ALIAS_CRYPTO("aria");
+MODULE_ALIAS_CRYPTO("aria-aesni-avx2");
diff --git a/arch/x86/crypto/aria_aesni_avx_glue.c b/arch/x86/crypto/aria_aesni_avx_glue.c
index 5f97e442349f..487094d64863 100644
--- a/arch/x86/crypto/aria_aesni_avx_glue.c
+++ b/arch/x86/crypto/aria_aesni_avx_glue.c
@@ -18,18 +18,24 @@
asmlinkage void aria_aesni_avx_encrypt_16way(const void *ctx, u8 *dst,
const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_encrypt_16way);
asmlinkage void aria_aesni_avx_decrypt_16way(const void *ctx, u8 *dst,
const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_decrypt_16way);
asmlinkage void aria_aesni_avx_ctr_crypt_16way(const void *ctx, u8 *dst,
const u8 *src,
u8 *keystream, u8 *iv);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_ctr_crypt_16way);
asmlinkage void aria_aesni_avx_gfni_encrypt_16way(const void *ctx, u8 *dst,
const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_gfni_encrypt_16way);
asmlinkage void aria_aesni_avx_gfni_decrypt_16way(const void *ctx, u8 *dst,
const u8 *src);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_gfni_decrypt_16way);
asmlinkage void aria_aesni_avx_gfni_ctr_crypt_16way(const void *ctx, u8 *dst,
const u8 *src,
u8 *keystream, u8 *iv);
+EXPORT_SYMBOL_GPL(aria_aesni_avx_gfni_ctr_crypt_16way);
static struct aria_avx_ops aria_ops;
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH v6 4/4] crypto: aria: implement aria-avx512
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
` (2 preceding siblings ...)
2022-11-21 0:39 ` [PATCH v6 3/4] crypto: aria: implement aria-avx2 Taehee Yoo
@ 2022-11-21 0:39 ` Taehee Yoo
2022-11-21 9:47 ` [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Borislav Petkov
4 siblings, 0 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 0:39 UTC (permalink / raw)
To: linux-crypto, herbert, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Cc: ap420073
aria-avx512 implementation uses AVX512 and GFNI.
It supports 64way parallel processing.
So, byteslicing code is changed to support 64way parallel.
And it exports some aria-avx2 functions such as encrypt() and decrypt().
AVX and AVX2 have 16 registers.
They should use memory to store/load state because of lack of registers.
But AVX512 supports 32 registers.
So, it doesn't require store/load in the s-box layer.
It means that it can reduce overhead of store/load in the s-box layer.
Also code become much simpler.
Benchmark with modprobe tcrypt mode=610 num_mb=8192, i3-12100:
ARIA-AVX512(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx512) encryption
tcrypt: 1 operation in 1504 cycles (1024 bytes)
tcrypt: 1 operation in 4595 cycles (4096 bytes)
tcrypt: 1 operation in 1763 cycles (1024 bytes)
tcrypt: 1 operation in 5540 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx512) decryption
tcrypt: 1 operation in 1502 cycles (1024 bytes)
tcrypt: 1 operation in 4615 cycles (4096 bytes)
tcrypt: 1 operation in 1759 cycles (1024 bytes)
tcrypt: 1 operation in 5554 cycles (4096 bytes)
ARIA-AVX2 with GFNI(128bit and 256bit)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) encryption
tcrypt: 1 operation in 2003 cycles (1024 bytes)
tcrypt: 1 operation in 5867 cycles (4096 bytes)
tcrypt: 1 operation in 2358 cycles (1024 bytes)
tcrypt: 1 operation in 7295 cycles (4096 bytes)
testing speed of multibuffer ecb(aria) (ecb-aria-avx2) decryption
tcrypt: 1 operation in 2004 cycles (1024 bytes)
tcrypt: 1 operation in 5956 cycles (4096 bytes)
tcrypt: 1 operation in 2409 cycles (1024 bytes)
tcrypt: 1 operation in 7564 cycles (4096 bytes)
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
---
v6:
- Use SYM_TYPED_FUNC_START instead of SYM_FUNC_START.
v5:
- Set CRYPTO_ALG_SKCIPHER_REQSIZE_LARGE flag.
v4:
- Add aria_avx512_request_ctx for keystream array.
v3:
- Use ARIA_CTX_enc_key, ARIA_CTX_dec_key, and ARIA_CTX_rounds defines.
v2:
- Add new "add keystream array into struct aria_ctx" patch.
- Use keystream array in the aria_ctx instead of stack memory
arch/x86/crypto/Kconfig | 19 +
arch/x86/crypto/Makefile | 3 +
arch/x86/crypto/aria-avx.h | 8 +
arch/x86/crypto/aria-gfni-avx512-asm_64.S | 1020 +++++++++++++++++++++
arch/x86/crypto/aria_gfni_avx512_glue.c | 250 +++++
5 files changed, 1300 insertions(+)
create mode 100644 arch/x86/crypto/aria-gfni-avx512-asm_64.S
create mode 100644 arch/x86/crypto/aria_gfni_avx512_glue.c
diff --git a/arch/x86/crypto/Kconfig b/arch/x86/crypto/Kconfig
index 3837ba8b78c5..688e848f740d 100644
--- a/arch/x86/crypto/Kconfig
+++ b/arch/x86/crypto/Kconfig
@@ -323,6 +323,25 @@ config CRYPTO_ARIA_AESNI_AVX2_X86_64
Processes 32 blocks in parallel.
+config CRYPTO_ARIA_GFNI_AVX512_X86_64
+ tristate "Ciphers: ARIA with modes: ECB, CTR (AVX512/GFNI)"
+ depends on X86 && 64BIT
+ select CRYPTO_SKCIPHER
+ select CRYPTO_SIMD
+ select CRYPTO_ALGAPI
+ select CRYPTO_ARIA
+ select CRYPTO_ARIA_AESNI_AVX_X86_64
+ select CRYPTO_ARIA_AESNI_AVX2_X86_64
+ help
+ Length-preserving cipher: ARIA cipher algorithms
+ (RFC 5794) with ECB and CTR modes
+
+ Architecture: x86_64 using:
+ - AVX512 (Advanced Vector Extensions)
+ - GFNI (Galois Field New Instructions)
+
+ Processes 64 blocks in parallel.
+
config CRYPTO_CHACHA20_X86_64
tristate "Ciphers: ChaCha20, XChaCha20, XChaCha12 (SSSE3/AVX2/AVX-512VL)"
depends on X86 && 64BIT
diff --git a/arch/x86/crypto/Makefile b/arch/x86/crypto/Makefile
index 5a349c7a8127..1844925f1f4a 100644
--- a/arch/x86/crypto/Makefile
+++ b/arch/x86/crypto/Makefile
@@ -106,6 +106,9 @@ aria-aesni-avx-x86_64-y := aria-aesni-avx-asm_64.o aria_aesni_avx_glue.o
obj-$(CONFIG_CRYPTO_ARIA_AESNI_AVX2_X86_64) += aria-aesni-avx2-x86_64.o
aria-aesni-avx2-x86_64-y := aria-aesni-avx2-asm_64.o aria_aesni_avx2_glue.o
+obj-$(CONFIG_CRYPTO_ARIA_GFNI_AVX512_X86_64) += aria-gfni-avx512-x86_64.o
+aria-gfni-avx512-x86_64-y := aria-gfni-avx512-asm_64.o aria_gfni_avx512_glue.o
+
quiet_cmd_perlasm = PERLASM $@
cmd_perlasm = $(PERL) $< > $@
$(obj)/%.S: $(src)/%.pl FORCE
diff --git a/arch/x86/crypto/aria-avx.h b/arch/x86/crypto/aria-avx.h
index b997c4888fb7..3eeb14d1d197 100644
--- a/arch/x86/crypto/aria-avx.h
+++ b/arch/x86/crypto/aria-avx.h
@@ -10,6 +10,9 @@
#define ARIA_AESNI_AVX2_PARALLEL_BLOCKS 32
#define ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE (ARIA_BLOCK_SIZE * 32)
+#define ARIA_GFNI_AVX512_PARALLEL_BLOCKS 64
+#define ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE (ARIA_BLOCK_SIZE * 64)
+
asmlinkage void aria_aesni_avx_encrypt_16way(const void *ctx, u8 *dst,
const u8 *src);
asmlinkage void aria_aesni_avx_decrypt_16way(const void *ctx, u8 *dst,
@@ -49,6 +52,11 @@ struct aria_avx_ops {
void (*aria_decrypt_32way)(const void *ctx, u8 *dst, const u8 *src);
void (*aria_ctr_crypt_32way)(const void *ctx, u8 *dst, const u8 *src,
u8 *keystream, u8 *iv);
+ void (*aria_encrypt_64way)(const void *ctx, u8 *dst, const u8 *src);
+ void (*aria_decrypt_64way)(const void *ctx, u8 *dst, const u8 *src);
+ void (*aria_ctr_crypt_64way)(const void *ctx, u8 *dst, const u8 *src,
+ u8 *keystream, u8 *iv);
+
};
#endif
diff --git a/arch/x86/crypto/aria-gfni-avx512-asm_64.S b/arch/x86/crypto/aria-gfni-avx512-asm_64.S
new file mode 100644
index 000000000000..32222cd3c7ee
--- /dev/null
+++ b/arch/x86/crypto/aria-gfni-avx512-asm_64.S
@@ -0,0 +1,1020 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * ARIA Cipher 64-way parallel algorithm (AVX512)
+ *
+ * Copyright (c) 2022 Taehee Yoo <ap420073@gmail.com>
+ *
+ */
+
+#include <linux/linkage.h>
+#include <asm/frame.h>
+#include <asm/asm-offsets.h>
+#include <linux/cfi_types.h>
+
+/* register macros */
+#define CTX %rdi
+
+
+#define BV8(a0, a1, a2, a3, a4, a5, a6, a7) \
+ ( (((a0) & 1) << 0) | \
+ (((a1) & 1) << 1) | \
+ (((a2) & 1) << 2) | \
+ (((a3) & 1) << 3) | \
+ (((a4) & 1) << 4) | \
+ (((a5) & 1) << 5) | \
+ (((a6) & 1) << 6) | \
+ (((a7) & 1) << 7) )
+
+#define BM8X8(l0, l1, l2, l3, l4, l5, l6, l7) \
+ ( ((l7) << (0 * 8)) | \
+ ((l6) << (1 * 8)) | \
+ ((l5) << (2 * 8)) | \
+ ((l4) << (3 * 8)) | \
+ ((l3) << (4 * 8)) | \
+ ((l2) << (5 * 8)) | \
+ ((l1) << (6 * 8)) | \
+ ((l0) << (7 * 8)) )
+
+#define add_le128(out, in, lo_counter, hi_counter1) \
+ vpaddq lo_counter, in, out; \
+ vpcmpuq $1, lo_counter, out, %k1; \
+ kaddb %k1, %k1, %k1; \
+ vpaddq hi_counter1, out, out{%k1};
+
+#define filter_8bit(x, lo_t, hi_t, mask4bit, tmp0) \
+ vpandq x, mask4bit, tmp0; \
+ vpandqn x, mask4bit, x; \
+ vpsrld $4, x, x; \
+ \
+ vpshufb tmp0, lo_t, tmp0; \
+ vpshufb x, hi_t, x; \
+ vpxorq tmp0, x, x;
+
+#define transpose_4x4(x0, x1, x2, x3, t1, t2) \
+ vpunpckhdq x1, x0, t2; \
+ vpunpckldq x1, x0, x0; \
+ \
+ vpunpckldq x3, x2, t1; \
+ vpunpckhdq x3, x2, x2; \
+ \
+ vpunpckhqdq t1, x0, x1; \
+ vpunpcklqdq t1, x0, x0; \
+ \
+ vpunpckhqdq x2, t2, x3; \
+ vpunpcklqdq x2, t2, x2;
+
+#define byteslice_16x16b(a0, b0, c0, d0, \
+ a1, b1, c1, d1, \
+ a2, b2, c2, d2, \
+ a3, b3, c3, d3, \
+ st0, st1) \
+ vmovdqu64 d2, st0; \
+ vmovdqu64 d3, st1; \
+ transpose_4x4(a0, a1, a2, a3, d2, d3); \
+ transpose_4x4(b0, b1, b2, b3, d2, d3); \
+ vmovdqu64 st0, d2; \
+ vmovdqu64 st1, d3; \
+ \
+ vmovdqu64 a0, st0; \
+ vmovdqu64 a1, st1; \
+ transpose_4x4(c0, c1, c2, c3, a0, a1); \
+ transpose_4x4(d0, d1, d2, d3, a0, a1); \
+ \
+ vbroadcasti64x2 .Lshufb_16x16b, a0; \
+ vmovdqu64 st1, a1; \
+ vpshufb a0, a2, a2; \
+ vpshufb a0, a3, a3; \
+ vpshufb a0, b0, b0; \
+ vpshufb a0, b1, b1; \
+ vpshufb a0, b2, b2; \
+ vpshufb a0, b3, b3; \
+ vpshufb a0, a1, a1; \
+ vpshufb a0, c0, c0; \
+ vpshufb a0, c1, c1; \
+ vpshufb a0, c2, c2; \
+ vpshufb a0, c3, c3; \
+ vpshufb a0, d0, d0; \
+ vpshufb a0, d1, d1; \
+ vpshufb a0, d2, d2; \
+ vpshufb a0, d3, d3; \
+ vmovdqu64 d3, st1; \
+ vmovdqu64 st0, d3; \
+ vpshufb a0, d3, a0; \
+ vmovdqu64 d2, st0; \
+ \
+ transpose_4x4(a0, b0, c0, d0, d2, d3); \
+ transpose_4x4(a1, b1, c1, d1, d2, d3); \
+ vmovdqu64 st0, d2; \
+ vmovdqu64 st1, d3; \
+ \
+ vmovdqu64 b0, st0; \
+ vmovdqu64 b1, st1; \
+ transpose_4x4(a2, b2, c2, d2, b0, b1); \
+ transpose_4x4(a3, b3, c3, d3, b0, b1); \
+ vmovdqu64 st0, b0; \
+ vmovdqu64 st1, b1; \
+ /* does not adjust output bytes inside vectors */
+
+#define debyteslice_16x16b(a0, b0, c0, d0, \
+ a1, b1, c1, d1, \
+ a2, b2, c2, d2, \
+ a3, b3, c3, d3, \
+ st0, st1) \
+ vmovdqu64 d2, st0; \
+ vmovdqu64 d3, st1; \
+ transpose_4x4(a0, a1, a2, a3, d2, d3); \
+ transpose_4x4(b0, b1, b2, b3, d2, d3); \
+ vmovdqu64 st0, d2; \
+ vmovdqu64 st1, d3; \
+ \
+ vmovdqu64 a0, st0; \
+ vmovdqu64 a1, st1; \
+ transpose_4x4(c0, c1, c2, c3, a0, a1); \
+ transpose_4x4(d0, d1, d2, d3, a0, a1); \
+ \
+ vbroadcasti64x2 .Lshufb_16x16b, a0; \
+ vmovdqu64 st1, a1; \
+ vpshufb a0, a2, a2; \
+ vpshufb a0, a3, a3; \
+ vpshufb a0, b0, b0; \
+ vpshufb a0, b1, b1; \
+ vpshufb a0, b2, b2; \
+ vpshufb a0, b3, b3; \
+ vpshufb a0, a1, a1; \
+ vpshufb a0, c0, c0; \
+ vpshufb a0, c1, c1; \
+ vpshufb a0, c2, c2; \
+ vpshufb a0, c3, c3; \
+ vpshufb a0, d0, d0; \
+ vpshufb a0, d1, d1; \
+ vpshufb a0, d2, d2; \
+ vpshufb a0, d3, d3; \
+ vmovdqu64 d3, st1; \
+ vmovdqu64 st0, d3; \
+ vpshufb a0, d3, a0; \
+ vmovdqu64 d2, st0; \
+ \
+ transpose_4x4(c0, d0, a0, b0, d2, d3); \
+ transpose_4x4(c1, d1, a1, b1, d2, d3); \
+ vmovdqu64 st0, d2; \
+ vmovdqu64 st1, d3; \
+ \
+ vmovdqu64 b0, st0; \
+ vmovdqu64 b1, st1; \
+ transpose_4x4(c2, d2, a2, b2, b0, b1); \
+ transpose_4x4(c3, d3, a3, b3, b0, b1); \
+ vmovdqu64 st0, b0; \
+ vmovdqu64 st1, b1; \
+ /* does not adjust output bytes inside vectors */
+
+/* load blocks to registers and apply pre-whitening */
+#define inpack16_pre(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ rio) \
+ vmovdqu64 (0 * 64)(rio), x0; \
+ vmovdqu64 (1 * 64)(rio), x1; \
+ vmovdqu64 (2 * 64)(rio), x2; \
+ vmovdqu64 (3 * 64)(rio), x3; \
+ vmovdqu64 (4 * 64)(rio), x4; \
+ vmovdqu64 (5 * 64)(rio), x5; \
+ vmovdqu64 (6 * 64)(rio), x6; \
+ vmovdqu64 (7 * 64)(rio), x7; \
+ vmovdqu64 (8 * 64)(rio), y0; \
+ vmovdqu64 (9 * 64)(rio), y1; \
+ vmovdqu64 (10 * 64)(rio), y2; \
+ vmovdqu64 (11 * 64)(rio), y3; \
+ vmovdqu64 (12 * 64)(rio), y4; \
+ vmovdqu64 (13 * 64)(rio), y5; \
+ vmovdqu64 (14 * 64)(rio), y6; \
+ vmovdqu64 (15 * 64)(rio), y7;
+
+/* byteslice pre-whitened blocks and store to temporary memory */
+#define inpack16_post(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem_ab, mem_cd) \
+ byteslice_16x16b(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ (mem_ab), (mem_cd)); \
+ \
+ vmovdqu64 x0, 0 * 64(mem_ab); \
+ vmovdqu64 x1, 1 * 64(mem_ab); \
+ vmovdqu64 x2, 2 * 64(mem_ab); \
+ vmovdqu64 x3, 3 * 64(mem_ab); \
+ vmovdqu64 x4, 4 * 64(mem_ab); \
+ vmovdqu64 x5, 5 * 64(mem_ab); \
+ vmovdqu64 x6, 6 * 64(mem_ab); \
+ vmovdqu64 x7, 7 * 64(mem_ab); \
+ vmovdqu64 y0, 0 * 64(mem_cd); \
+ vmovdqu64 y1, 1 * 64(mem_cd); \
+ vmovdqu64 y2, 2 * 64(mem_cd); \
+ vmovdqu64 y3, 3 * 64(mem_cd); \
+ vmovdqu64 y4, 4 * 64(mem_cd); \
+ vmovdqu64 y5, 5 * 64(mem_cd); \
+ vmovdqu64 y6, 6 * 64(mem_cd); \
+ vmovdqu64 y7, 7 * 64(mem_cd);
+
+#define write_output(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ mem) \
+ vmovdqu64 x0, 0 * 64(mem); \
+ vmovdqu64 x1, 1 * 64(mem); \
+ vmovdqu64 x2, 2 * 64(mem); \
+ vmovdqu64 x3, 3 * 64(mem); \
+ vmovdqu64 x4, 4 * 64(mem); \
+ vmovdqu64 x5, 5 * 64(mem); \
+ vmovdqu64 x6, 6 * 64(mem); \
+ vmovdqu64 x7, 7 * 64(mem); \
+ vmovdqu64 y0, 8 * 64(mem); \
+ vmovdqu64 y1, 9 * 64(mem); \
+ vmovdqu64 y2, 10 * 64(mem); \
+ vmovdqu64 y3, 11 * 64(mem); \
+ vmovdqu64 y4, 12 * 64(mem); \
+ vmovdqu64 y5, 13 * 64(mem); \
+ vmovdqu64 y6, 14 * 64(mem); \
+ vmovdqu64 y7, 15 * 64(mem); \
+
+#define aria_store_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, idx) \
+ vmovdqu64 x0, ((idx + 0) * 64)(mem_tmp); \
+ vmovdqu64 x1, ((idx + 1) * 64)(mem_tmp); \
+ vmovdqu64 x2, ((idx + 2) * 64)(mem_tmp); \
+ vmovdqu64 x3, ((idx + 3) * 64)(mem_tmp); \
+ vmovdqu64 x4, ((idx + 4) * 64)(mem_tmp); \
+ vmovdqu64 x5, ((idx + 5) * 64)(mem_tmp); \
+ vmovdqu64 x6, ((idx + 6) * 64)(mem_tmp); \
+ vmovdqu64 x7, ((idx + 7) * 64)(mem_tmp);
+
+#define aria_load_state_8way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ mem_tmp, idx) \
+ vmovdqu64 ((idx + 0) * 64)(mem_tmp), x0; \
+ vmovdqu64 ((idx + 1) * 64)(mem_tmp), x1; \
+ vmovdqu64 ((idx + 2) * 64)(mem_tmp), x2; \
+ vmovdqu64 ((idx + 3) * 64)(mem_tmp), x3; \
+ vmovdqu64 ((idx + 4) * 64)(mem_tmp), x4; \
+ vmovdqu64 ((idx + 5) * 64)(mem_tmp), x5; \
+ vmovdqu64 ((idx + 6) * 64)(mem_tmp), x6; \
+ vmovdqu64 ((idx + 7) * 64)(mem_tmp), x7;
+
+#define aria_ark_16way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ t0, rk, round) \
+ /* AddRoundKey */ \
+ vpbroadcastb ((round * 16) + 3)(rk), t0; \
+ vpxorq t0, x0, x0; \
+ vpbroadcastb ((round * 16) + 2)(rk), t0; \
+ vpxorq t0, x1, x1; \
+ vpbroadcastb ((round * 16) + 1)(rk), t0; \
+ vpxorq t0, x2, x2; \
+ vpbroadcastb ((round * 16) + 0)(rk), t0; \
+ vpxorq t0, x3, x3; \
+ vpbroadcastb ((round * 16) + 7)(rk), t0; \
+ vpxorq t0, x4, x4; \
+ vpbroadcastb ((round * 16) + 6)(rk), t0; \
+ vpxorq t0, x5, x5; \
+ vpbroadcastb ((round * 16) + 5)(rk), t0; \
+ vpxorq t0, x6, x6; \
+ vpbroadcastb ((round * 16) + 4)(rk), t0; \
+ vpxorq t0, x7, x7; \
+ vpbroadcastb ((round * 16) + 11)(rk), t0; \
+ vpxorq t0, y0, y0; \
+ vpbroadcastb ((round * 16) + 10)(rk), t0; \
+ vpxorq t0, y1, y1; \
+ vpbroadcastb ((round * 16) + 9)(rk), t0; \
+ vpxorq t0, y2, y2; \
+ vpbroadcastb ((round * 16) + 8)(rk), t0; \
+ vpxorq t0, y3, y3; \
+ vpbroadcastb ((round * 16) + 15)(rk), t0; \
+ vpxorq t0, y4, y4; \
+ vpbroadcastb ((round * 16) + 14)(rk), t0; \
+ vpxorq t0, y5, y5; \
+ vpbroadcastb ((round * 16) + 13)(rk), t0; \
+ vpxorq t0, y6, y6; \
+ vpbroadcastb ((round * 16) + 12)(rk), t0; \
+ vpxorq t0, y7, y7;
+
+#define aria_sbox_8way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ t0, t1, t2, t3, \
+ t4, t5, t6, t7) \
+ vpbroadcastq .Ltf_s2_bitmatrix, t0; \
+ vpbroadcastq .Ltf_inv_bitmatrix, t1; \
+ vpbroadcastq .Ltf_id_bitmatrix, t2; \
+ vpbroadcastq .Ltf_aff_bitmatrix, t3; \
+ vpbroadcastq .Ltf_x2_bitmatrix, t4; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x1, x1; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x5, x5; \
+ vgf2p8affineqb $(tf_inv_const), t1, x2, x2; \
+ vgf2p8affineqb $(tf_inv_const), t1, x6, x6; \
+ vgf2p8affineinvqb $0, t2, x2, x2; \
+ vgf2p8affineinvqb $0, t2, x6, x6; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x0, x0; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x4, x4; \
+ vgf2p8affineqb $(tf_x2_const), t4, x3, x3; \
+ vgf2p8affineqb $(tf_x2_const), t4, x7, x7; \
+ vgf2p8affineinvqb $0, t2, x3, x3; \
+ vgf2p8affineinvqb $0, t2, x7, x7;
+
+#define aria_sbox_16way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ t0, t1, t2, t3, \
+ t4, t5, t6, t7) \
+ vpbroadcastq .Ltf_s2_bitmatrix, t0; \
+ vpbroadcastq .Ltf_inv_bitmatrix, t1; \
+ vpbroadcastq .Ltf_id_bitmatrix, t2; \
+ vpbroadcastq .Ltf_aff_bitmatrix, t3; \
+ vpbroadcastq .Ltf_x2_bitmatrix, t4; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x1, x1; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, x5, x5; \
+ vgf2p8affineqb $(tf_inv_const), t1, x2, x2; \
+ vgf2p8affineqb $(tf_inv_const), t1, x6, x6; \
+ vgf2p8affineinvqb $0, t2, x2, x2; \
+ vgf2p8affineinvqb $0, t2, x6, x6; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x0, x0; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, x4, x4; \
+ vgf2p8affineqb $(tf_x2_const), t4, x3, x3; \
+ vgf2p8affineqb $(tf_x2_const), t4, x7, x7; \
+ vgf2p8affineinvqb $0, t2, x3, x3; \
+ vgf2p8affineinvqb $0, t2, x7, x7; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, y1, y1; \
+ vgf2p8affineinvqb $(tf_s2_const), t0, y5, y5; \
+ vgf2p8affineqb $(tf_inv_const), t1, y2, y2; \
+ vgf2p8affineqb $(tf_inv_const), t1, y6, y6; \
+ vgf2p8affineinvqb $0, t2, y2, y2; \
+ vgf2p8affineinvqb $0, t2, y6, y6; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, y0, y0; \
+ vgf2p8affineinvqb $(tf_aff_const), t3, y4, y4; \
+ vgf2p8affineqb $(tf_x2_const), t4, y3, y3; \
+ vgf2p8affineqb $(tf_x2_const), t4, y7, y7; \
+ vgf2p8affineinvqb $0, t2, y3, y3; \
+ vgf2p8affineinvqb $0, t2, y7, y7;
+
+
+#define aria_diff_m(x0, x1, x2, x3, \
+ t0, t1, t2, t3) \
+ /* T = rotr32(X, 8); */ \
+ /* X ^= T */ \
+ vpxorq x0, x3, t0; \
+ vpxorq x1, x0, t1; \
+ vpxorq x2, x1, t2; \
+ vpxorq x3, x2, t3; \
+ /* X = T ^ rotr(X, 16); */ \
+ vpxorq t2, x0, x0; \
+ vpxorq x1, t3, t3; \
+ vpxorq t0, x2, x2; \
+ vpxorq t1, x3, x1; \
+ vmovdqu64 t3, x3;
+
+#define aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7) \
+ /* t1 ^= t2; */ \
+ vpxorq y0, x4, x4; \
+ vpxorq y1, x5, x5; \
+ vpxorq y2, x6, x6; \
+ vpxorq y3, x7, x7; \
+ \
+ /* t2 ^= t3; */ \
+ vpxorq y4, y0, y0; \
+ vpxorq y5, y1, y1; \
+ vpxorq y6, y2, y2; \
+ vpxorq y7, y3, y3; \
+ \
+ /* t0 ^= t1; */ \
+ vpxorq x4, x0, x0; \
+ vpxorq x5, x1, x1; \
+ vpxorq x6, x2, x2; \
+ vpxorq x7, x3, x3; \
+ \
+ /* t3 ^= t1; */ \
+ vpxorq x4, y4, y4; \
+ vpxorq x5, y5, y5; \
+ vpxorq x6, y6, y6; \
+ vpxorq x7, y7, y7; \
+ \
+ /* t2 ^= t0; */ \
+ vpxorq x0, y0, y0; \
+ vpxorq x1, y1, y1; \
+ vpxorq x2, y2, y2; \
+ vpxorq x3, y3, y3; \
+ \
+ /* t1 ^= t2; */ \
+ vpxorq y0, x4, x4; \
+ vpxorq y1, x5, x5; \
+ vpxorq y2, x6, x6; \
+ vpxorq y3, x7, x7;
+
+#define aria_fe_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7, \
+ mem_tmp, rk, round) \
+ aria_ark_16way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, y1, y2, y3, y4, y5, y6, y7, \
+ z0, rk, round); \
+ \
+ aria_sbox_16way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y2, y3, y0, y1, \
+ y6, y7, y4, y5, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, z0, z1, z2, z3); \
+ aria_diff_m(x4, x5, x6, x7, z0, z1, z2, z3); \
+ aria_diff_m(y0, y1, y2, y3, z0, z1, z2, z3); \
+ aria_diff_m(y4, y5, y6, y7, z0, z1, z2, z3); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T3 = ABCD -> BADC \
+ * T3 = y4, y5, y6, y7 -> y5, y4, y7, y6 \
+ * T0 = ABCD -> CDAB \
+ * T0 = x0, x1, x2, x3 -> x2, x3, x0, x1 \
+ * T1 = ABCD -> DCBA \
+ * T1 = x4, x5, x6, x7 -> x7, x6, x5, x4 \
+ */ \
+ aria_diff_word(x2, x3, x0, x1, \
+ x7, x6, x5, x4, \
+ y0, y1, y2, y3, \
+ y5, y4, y7, y6); \
+
+
+#define aria_fo_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7, \
+ mem_tmp, rk, round) \
+ aria_ark_16way(x0, x1, x2, x3, x4, x5, x6, x7, \
+ y0, y1, y2, y3, y4, y5, y6, y7, \
+ z0, rk, round); \
+ \
+ aria_sbox_16way_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7); \
+ \
+ aria_diff_m(x0, x1, x2, x3, z0, z1, z2, z3); \
+ aria_diff_m(x4, x5, x6, x7, z0, z1, z2, z3); \
+ aria_diff_m(y0, y1, y2, y3, z0, z1, z2, z3); \
+ aria_diff_m(y4, y5, y6, y7, z0, z1, z2, z3); \
+ aria_diff_word(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7); \
+ /* aria_diff_byte() \
+ * T1 = ABCD -> BADC \
+ * T1 = x4, x5, x6, x7 -> x5, x4, x7, x6 \
+ * T2 = ABCD -> CDAB \
+ * T2 = y0, y1, y2, y3, -> y2, y3, y0, y1 \
+ * T3 = ABCD -> DCBA \
+ * T3 = y4, y5, y6, y7 -> y7, y6, y5, y4 \
+ */ \
+ aria_diff_word(x0, x1, x2, x3, \
+ x5, x4, x7, x6, \
+ y2, y3, y0, y1, \
+ y7, y6, y5, y4);
+
+#define aria_ff_gfni(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7, \
+ mem_tmp, rk, round, last_round) \
+ aria_ark_16way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, rk, round); \
+ aria_sbox_16way_gfni(x2, x3, x0, x1, \
+ x6, x7, x4, x5, \
+ y2, y3, y0, y1, \
+ y6, y7, y4, y5, \
+ z0, z1, z2, z3, \
+ z4, z5, z6, z7); \
+ aria_ark_16way(x0, x1, x2, x3, \
+ x4, x5, x6, x7, \
+ y0, y1, y2, y3, \
+ y4, y5, y6, y7, \
+ z0, rk, last_round);
+
+
+.section .rodata.cst64, "aM", @progbits, 64
+.align 64
+.Lpack_bswap:
+ .long 0x00010203, 0x04050607, 0x80808080, 0x80808080
+ .long 0x00010203, 0x04050607, 0x80808080, 0x80808080
+ .long 0x00010203, 0x04050607, 0x80808080, 0x80808080
+ .long 0x00010203, 0x04050607, 0x80808080, 0x80808080
+
+.Lcounter0123_lo:
+ .quad 0, 0
+ .quad 1, 0
+ .quad 2, 0
+ .quad 3, 0
+
+.section .rodata.cst32.shufb_16x16b, "aM", @progbits, 32
+.align 32
+#define SHUFB_BYTES(idx) \
+ 0 + (idx), 4 + (idx), 8 + (idx), 12 + (idx)
+.Lshufb_16x16b:
+ .byte SHUFB_BYTES(0), SHUFB_BYTES(1), SHUFB_BYTES(2), SHUFB_BYTES(3)
+ .byte SHUFB_BYTES(0), SHUFB_BYTES(1), SHUFB_BYTES(2), SHUFB_BYTES(3)
+
+.section .rodata.cst16, "aM", @progbits, 16
+.align 16
+
+.Lcounter4444_lo:
+ .quad 4, 0
+.Lcounter8888_lo:
+ .quad 8, 0
+.Lcounter16161616_lo:
+ .quad 16, 0
+.Lcounter1111_hi:
+ .quad 0, 1
+
+/* For isolating SubBytes from AESENCLAST, inverse shift row */
+.Linv_shift_row:
+ .byte 0x00, 0x0d, 0x0a, 0x07, 0x04, 0x01, 0x0e, 0x0b
+ .byte 0x08, 0x05, 0x02, 0x0f, 0x0c, 0x09, 0x06, 0x03
+.Lshift_row:
+ .byte 0x00, 0x05, 0x0a, 0x0f, 0x04, 0x09, 0x0e, 0x03
+ .byte 0x08, 0x0d, 0x02, 0x07, 0x0c, 0x01, 0x06, 0x0b
+/* For CTR-mode IV byteswap */
+.Lbswap128_mask:
+ .byte 0x0f, 0x0e, 0x0d, 0x0c, 0x0b, 0x0a, 0x09, 0x08
+ .byte 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00
+
+/* AES inverse affine and S2 combined:
+ * 1 1 0 0 0 0 0 1 x0 0
+ * 0 1 0 0 1 0 0 0 x1 0
+ * 1 1 0 0 1 1 1 1 x2 0
+ * 0 1 1 0 1 0 0 1 x3 1
+ * 0 1 0 0 1 1 0 0 * x4 + 0
+ * 0 1 0 1 1 0 0 0 x5 0
+ * 0 0 0 0 0 1 0 1 x6 0
+ * 1 1 1 0 0 1 1 1 x7 1
+ */
+.Ltf_lo__inv_aff__and__s2:
+ .octa 0x92172DA81A9FA520B2370D883ABF8500
+.Ltf_hi__inv_aff__and__s2:
+ .octa 0x2B15FFC1AF917B45E6D8320C625CB688
+
+/* X2 and AES forward affine combined:
+ * 1 0 1 1 0 0 0 1 x0 0
+ * 0 1 1 1 1 0 1 1 x1 0
+ * 0 0 0 1 1 0 1 0 x2 1
+ * 0 1 0 0 0 1 0 0 x3 0
+ * 0 0 1 1 1 0 1 1 * x4 + 0
+ * 0 1 0 0 1 0 0 0 x5 0
+ * 1 1 0 1 0 0 1 1 x6 0
+ * 0 1 0 0 1 0 1 0 x7 0
+ */
+.Ltf_lo__x2__and__fwd_aff:
+ .octa 0xEFAE0544FCBD1657B8F95213ABEA4100
+.Ltf_hi__x2__and__fwd_aff:
+ .octa 0x3F893781E95FE1576CDA64D2BA0CB204
+
+.section .rodata.cst8, "aM", @progbits, 8
+.align 8
+/* AES affine: */
+#define tf_aff_const BV8(1, 1, 0, 0, 0, 1, 1, 0)
+.Ltf_aff_bitmatrix:
+ .quad BM8X8(BV8(1, 0, 0, 0, 1, 1, 1, 1),
+ BV8(1, 1, 0, 0, 0, 1, 1, 1),
+ BV8(1, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 1, 0, 0, 0, 1),
+ BV8(1, 1, 1, 1, 1, 0, 0, 0),
+ BV8(0, 1, 1, 1, 1, 1, 0, 0),
+ BV8(0, 0, 1, 1, 1, 1, 1, 0),
+ BV8(0, 0, 0, 1, 1, 1, 1, 1))
+
+/* AES inverse affine: */
+#define tf_inv_const BV8(1, 0, 1, 0, 0, 0, 0, 0)
+.Ltf_inv_bitmatrix:
+ .quad BM8X8(BV8(0, 0, 1, 0, 0, 1, 0, 1),
+ BV8(1, 0, 0, 1, 0, 0, 1, 0),
+ BV8(0, 1, 0, 0, 1, 0, 0, 1),
+ BV8(1, 0, 1, 0, 0, 1, 0, 0),
+ BV8(0, 1, 0, 1, 0, 0, 1, 0),
+ BV8(0, 0, 1, 0, 1, 0, 0, 1),
+ BV8(1, 0, 0, 1, 0, 1, 0, 0),
+ BV8(0, 1, 0, 0, 1, 0, 1, 0))
+
+/* S2: */
+#define tf_s2_const BV8(0, 1, 0, 0, 0, 1, 1, 1)
+.Ltf_s2_bitmatrix:
+ .quad BM8X8(BV8(0, 1, 0, 1, 0, 1, 1, 1),
+ BV8(0, 0, 1, 1, 1, 1, 1, 1),
+ BV8(1, 1, 1, 0, 1, 1, 0, 1),
+ BV8(1, 1, 0, 0, 0, 0, 1, 1),
+ BV8(0, 1, 0, 0, 0, 0, 1, 1),
+ BV8(1, 1, 0, 0, 1, 1, 1, 0),
+ BV8(0, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 1, 0, 1, 1, 0))
+
+/* X2: */
+#define tf_x2_const BV8(0, 0, 1, 1, 0, 1, 0, 0)
+.Ltf_x2_bitmatrix:
+ .quad BM8X8(BV8(0, 0, 0, 1, 1, 0, 0, 0),
+ BV8(0, 0, 1, 0, 0, 1, 1, 0),
+ BV8(0, 0, 0, 0, 1, 0, 1, 0),
+ BV8(1, 1, 1, 0, 0, 0, 1, 1),
+ BV8(1, 1, 1, 0, 1, 1, 0, 0),
+ BV8(0, 1, 1, 0, 1, 0, 1, 1),
+ BV8(1, 0, 1, 1, 1, 1, 0, 1),
+ BV8(1, 0, 0, 1, 0, 0, 1, 1))
+
+/* Identity matrix: */
+.Ltf_id_bitmatrix:
+ .quad BM8X8(BV8(1, 0, 0, 0, 0, 0, 0, 0),
+ BV8(0, 1, 0, 0, 0, 0, 0, 0),
+ BV8(0, 0, 1, 0, 0, 0, 0, 0),
+ BV8(0, 0, 0, 1, 0, 0, 0, 0),
+ BV8(0, 0, 0, 0, 1, 0, 0, 0),
+ BV8(0, 0, 0, 0, 0, 1, 0, 0),
+ BV8(0, 0, 0, 0, 0, 0, 1, 0),
+ BV8(0, 0, 0, 0, 0, 0, 0, 1))
+
+/* 4-bit mask */
+.section .rodata.cst4.L0f0f0f0f, "aM", @progbits, 4
+.align 4
+.L0f0f0f0f:
+ .long 0x0f0f0f0f
+
+.text
+SYM_FUNC_START_LOCAL(__aria_gfni_avx512_crypt_64way)
+ /* input:
+ * %r9: rk
+ * %rsi: dst
+ * %rdx: src
+ * %zmm0..%zmm15: byte-sliced blocks
+ */
+
+ FRAME_BEGIN
+
+ movq %rsi, %rax;
+ leaq 8 * 64(%rax), %r8;
+
+ inpack16_post(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14,
+ %zmm15, %rax, %r8);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 0);
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 1);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 2);
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 3);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 4);
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 5);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 6);
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 7);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 8);
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 9);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 10);
+ cmpl $12, ARIA_CTX_rounds(CTX);
+ jne .Laria_gfni_192;
+ aria_ff_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 11, 12);
+ jmp .Laria_gfni_end;
+.Laria_gfni_192:
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 11);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 12);
+ cmpl $14, ARIA_CTX_rounds(CTX);
+ jne .Laria_gfni_256;
+ aria_ff_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 13, 14);
+ jmp .Laria_gfni_end;
+.Laria_gfni_256:
+ aria_fe_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 13);
+ aria_fo_gfni(%zmm0, %zmm1, %zmm2, %zmm3,
+ %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 14);
+ aria_ff_gfni(%zmm3, %zmm2, %zmm1, %zmm0,
+ %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10,
+ %zmm12, %zmm13, %zmm14, %zmm15,
+ %zmm24, %zmm25, %zmm26, %zmm27,
+ %zmm28, %zmm29, %zmm30, %zmm31,
+ %rax, %r9, 15, 16);
+.Laria_gfni_end:
+ debyteslice_16x16b(%zmm9, %zmm12, %zmm3, %zmm6,
+ %zmm8, %zmm13, %zmm2, %zmm7,
+ %zmm11, %zmm14, %zmm1, %zmm4,
+ %zmm10, %zmm15, %zmm0, %zmm5,
+ (%rax), (%r8));
+ FRAME_END
+ RET;
+SYM_FUNC_END(__aria_gfni_avx512_crypt_64way)
+
+SYM_TYPED_FUNC_START(aria_gfni_avx512_encrypt_64way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ inpack16_pre(%zmm0, %zmm1, %zmm2, %zmm3, %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11, %zmm12, %zmm13, %zmm14,
+ %zmm15, %rdx);
+
+ call __aria_gfni_avx512_crypt_64way;
+
+ write_output(%zmm3, %zmm2, %zmm1, %zmm0, %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10, %zmm12, %zmm13, %zmm14,
+ %zmm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_gfni_avx512_encrypt_64way)
+
+SYM_TYPED_FUNC_START(aria_gfni_avx512_decrypt_64way)
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: dst
+ * %rdx: src
+ */
+
+ FRAME_BEGIN
+
+ leaq ARIA_CTX_dec_key(CTX), %r9;
+
+ inpack16_pre(%zmm0, %zmm1, %zmm2, %zmm3, %zmm4, %zmm5, %zmm6, %zmm7,
+ %zmm8, %zmm9, %zmm10, %zmm11, %zmm12, %zmm13, %zmm14,
+ %zmm15, %rdx);
+
+ call __aria_gfni_avx512_crypt_64way;
+
+ write_output(%zmm3, %zmm2, %zmm1, %zmm0, %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10, %zmm12, %zmm13, %zmm14,
+ %zmm15, %rax);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_gfni_avx512_decrypt_64way)
+
+SYM_FUNC_START_LOCAL(__aria_gfni_avx512_ctr_gen_keystream_64way)
+ /* input:
+ * %rdi: ctx
+ * %rsi: dst
+ * %rdx: src
+ * %rcx: keystream
+ * %r8: iv (big endian, 128bit)
+ */
+
+ FRAME_BEGIN
+
+ vbroadcasti64x2 .Lbswap128_mask (%rip), %zmm19;
+ vmovdqa64 .Lcounter0123_lo (%rip), %zmm21;
+ vbroadcasti64x2 .Lcounter4444_lo (%rip), %zmm22;
+ vbroadcasti64x2 .Lcounter8888_lo (%rip), %zmm23;
+ vbroadcasti64x2 .Lcounter16161616_lo (%rip), %zmm24;
+ vbroadcasti64x2 .Lcounter1111_hi (%rip), %zmm25;
+
+ /* load IV and byteswap */
+ movq 8(%r8), %r11;
+ movq (%r8), %r10;
+ bswapq %r11;
+ bswapq %r10;
+ vbroadcasti64x2 (%r8), %zmm20;
+ vpshufb %zmm19, %zmm20, %zmm20;
+
+ /* check need for handling 64-bit overflow and carry */
+ cmpq $(0xffffffffffffffff - 64), %r11;
+ ja .Lload_ctr_carry;
+
+ /* construct IVs */
+ vpaddq %zmm21, %zmm20, %zmm0; /* +0:+1:+2:+3 */
+ vpaddq %zmm22, %zmm0, %zmm1; /* +4:+5:+6:+7 */
+ vpaddq %zmm23, %zmm0, %zmm2; /* +8:+9:+10:+11 */
+ vpaddq %zmm23, %zmm1, %zmm3; /* +12:+13:+14:+15 */
+ vpaddq %zmm24, %zmm0, %zmm4; /* +16... */
+ vpaddq %zmm24, %zmm1, %zmm5; /* +20... */
+ vpaddq %zmm24, %zmm2, %zmm6; /* +24... */
+ vpaddq %zmm24, %zmm3, %zmm7; /* +28... */
+ vpaddq %zmm24, %zmm4, %zmm8; /* +32... */
+ vpaddq %zmm24, %zmm5, %zmm9; /* +36... */
+ vpaddq %zmm24, %zmm6, %zmm10; /* +40... */
+ vpaddq %zmm24, %zmm7, %zmm11; /* +44... */
+ vpaddq %zmm24, %zmm8, %zmm12; /* +48... */
+ vpaddq %zmm24, %zmm9, %zmm13; /* +52... */
+ vpaddq %zmm24, %zmm10, %zmm14; /* +56... */
+ vpaddq %zmm24, %zmm11, %zmm15; /* +60... */
+ jmp .Lload_ctr_done;
+
+.Lload_ctr_carry:
+ /* construct IVs */
+ add_le128(%zmm0, %zmm20, %zmm21, %zmm25); /* +0:+1:+2:+3 */
+ add_le128(%zmm1, %zmm0, %zmm22, %zmm25); /* +4:+5:+6:+7 */
+ add_le128(%zmm2, %zmm0, %zmm23, %zmm25); /* +8:+9:+10:+11 */
+ add_le128(%zmm3, %zmm1, %zmm23, %zmm25); /* +12:+13:+14:+15 */
+ add_le128(%zmm4, %zmm0, %zmm24, %zmm25); /* +16... */
+ add_le128(%zmm5, %zmm1, %zmm24, %zmm25); /* +20... */
+ add_le128(%zmm6, %zmm2, %zmm24, %zmm25); /* +24... */
+ add_le128(%zmm7, %zmm3, %zmm24, %zmm25); /* +28... */
+ add_le128(%zmm8, %zmm4, %zmm24, %zmm25); /* +32... */
+ add_le128(%zmm9, %zmm5, %zmm24, %zmm25); /* +36... */
+ add_le128(%zmm10, %zmm6, %zmm24, %zmm25); /* +40... */
+ add_le128(%zmm11, %zmm7, %zmm24, %zmm25); /* +44... */
+ add_le128(%zmm12, %zmm8, %zmm24, %zmm25); /* +48... */
+ add_le128(%zmm13, %zmm9, %zmm24, %zmm25); /* +52... */
+ add_le128(%zmm14, %zmm10, %zmm24, %zmm25); /* +56... */
+ add_le128(%zmm15, %zmm11, %zmm24, %zmm25); /* +60... */
+
+.Lload_ctr_done:
+ /* Byte-swap IVs and update counter. */
+ addq $64, %r11;
+ adcq $0, %r10;
+ vpshufb %zmm19, %zmm15, %zmm15;
+ vpshufb %zmm19, %zmm14, %zmm14;
+ vpshufb %zmm19, %zmm13, %zmm13;
+ vpshufb %zmm19, %zmm12, %zmm12;
+ vpshufb %zmm19, %zmm11, %zmm11;
+ vpshufb %zmm19, %zmm10, %zmm10;
+ vpshufb %zmm19, %zmm9, %zmm9;
+ vpshufb %zmm19, %zmm8, %zmm8;
+ bswapq %r11;
+ bswapq %r10;
+ vpshufb %zmm19, %zmm7, %zmm7;
+ vpshufb %zmm19, %zmm6, %zmm6;
+ vpshufb %zmm19, %zmm5, %zmm5;
+ vpshufb %zmm19, %zmm4, %zmm4;
+ vpshufb %zmm19, %zmm3, %zmm3;
+ vpshufb %zmm19, %zmm2, %zmm2;
+ vpshufb %zmm19, %zmm1, %zmm1;
+ vpshufb %zmm19, %zmm0, %zmm0;
+ movq %r11, 8(%r8);
+ movq %r10, (%r8);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(__aria_gfni_avx512_ctr_gen_keystream_64way)
+
+SYM_TYPED_FUNC_START(aria_gfni_avx512_ctr_crypt_64way)
+ /* input:
+ * %rdi: ctx
+ * %rsi: dst
+ * %rdx: src
+ * %rcx: keystream
+ * %r8: iv (big endian, 128bit)
+ */
+ FRAME_BEGIN
+
+ call __aria_gfni_avx512_ctr_gen_keystream_64way
+
+ leaq (%rsi), %r10;
+ leaq (%rdx), %r11;
+ leaq (%rcx), %rsi;
+ leaq (%rcx), %rdx;
+ leaq ARIA_CTX_enc_key(CTX), %r9;
+
+ call __aria_gfni_avx512_crypt_64way;
+
+ vpxorq (0 * 64)(%r11), %zmm3, %zmm3;
+ vpxorq (1 * 64)(%r11), %zmm2, %zmm2;
+ vpxorq (2 * 64)(%r11), %zmm1, %zmm1;
+ vpxorq (3 * 64)(%r11), %zmm0, %zmm0;
+ vpxorq (4 * 64)(%r11), %zmm6, %zmm6;
+ vpxorq (5 * 64)(%r11), %zmm7, %zmm7;
+ vpxorq (6 * 64)(%r11), %zmm4, %zmm4;
+ vpxorq (7 * 64)(%r11), %zmm5, %zmm5;
+ vpxorq (8 * 64)(%r11), %zmm9, %zmm9;
+ vpxorq (9 * 64)(%r11), %zmm8, %zmm8;
+ vpxorq (10 * 64)(%r11), %zmm11, %zmm11;
+ vpxorq (11 * 64)(%r11), %zmm10, %zmm10;
+ vpxorq (12 * 64)(%r11), %zmm12, %zmm12;
+ vpxorq (13 * 64)(%r11), %zmm13, %zmm13;
+ vpxorq (14 * 64)(%r11), %zmm14, %zmm14;
+ vpxorq (15 * 64)(%r11), %zmm15, %zmm15;
+ write_output(%zmm3, %zmm2, %zmm1, %zmm0, %zmm6, %zmm7, %zmm4, %zmm5,
+ %zmm9, %zmm8, %zmm11, %zmm10, %zmm12, %zmm13, %zmm14,
+ %zmm15, %r10);
+
+ FRAME_END
+ RET;
+SYM_FUNC_END(aria_gfni_avx512_ctr_crypt_64way)
diff --git a/arch/x86/crypto/aria_gfni_avx512_glue.c b/arch/x86/crypto/aria_gfni_avx512_glue.c
new file mode 100644
index 000000000000..f4a2208d2638
--- /dev/null
+++ b/arch/x86/crypto/aria_gfni_avx512_glue.c
@@ -0,0 +1,250 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Glue Code for the AVX512/GFNI assembler implementation of the ARIA Cipher
+ *
+ * Copyright (c) 2022 Taehee Yoo <ap420073@gmail.com>
+ */
+
+#include <crypto/algapi.h>
+#include <crypto/internal/simd.h>
+#include <crypto/aria.h>
+#include <linux/crypto.h>
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/types.h>
+
+#include "ecb_cbc_helpers.h"
+#include "aria-avx.h"
+
+asmlinkage void aria_gfni_avx512_encrypt_64way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_gfni_avx512_decrypt_64way(const void *ctx, u8 *dst,
+ const u8 *src);
+asmlinkage void aria_gfni_avx512_ctr_crypt_64way(const void *ctx, u8 *dst,
+ const u8 *src,
+ u8 *keystream, u8 *iv);
+
+static struct aria_avx_ops aria_ops;
+
+struct aria_avx512_request_ctx {
+ u8 keystream[ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE];
+};
+
+static int ecb_do_encrypt(struct skcipher_request *req, const u32 *rkey)
+{
+ ECB_WALK_START(req, ARIA_BLOCK_SIZE, ARIA_AESNI_PARALLEL_BLOCKS);
+ ECB_BLOCK(ARIA_GFNI_AVX512_PARALLEL_BLOCKS, aria_ops.aria_encrypt_64way);
+ ECB_BLOCK(ARIA_AESNI_AVX2_PARALLEL_BLOCKS, aria_ops.aria_encrypt_32way);
+ ECB_BLOCK(ARIA_AESNI_PARALLEL_BLOCKS, aria_ops.aria_encrypt_16way);
+ ECB_BLOCK(1, aria_encrypt);
+ ECB_WALK_END();
+}
+
+static int ecb_do_decrypt(struct skcipher_request *req, const u32 *rkey)
+{
+ ECB_WALK_START(req, ARIA_BLOCK_SIZE, ARIA_AESNI_PARALLEL_BLOCKS);
+ ECB_BLOCK(ARIA_GFNI_AVX512_PARALLEL_BLOCKS, aria_ops.aria_decrypt_64way);
+ ECB_BLOCK(ARIA_AESNI_AVX2_PARALLEL_BLOCKS, aria_ops.aria_decrypt_32way);
+ ECB_BLOCK(ARIA_AESNI_PARALLEL_BLOCKS, aria_ops.aria_decrypt_16way);
+ ECB_BLOCK(1, aria_decrypt);
+ ECB_WALK_END();
+}
+
+static int aria_avx512_ecb_encrypt(struct skcipher_request *req)
+{
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+ return ecb_do_encrypt(req, ctx->enc_key[0]);
+}
+
+static int aria_avx512_ecb_decrypt(struct skcipher_request *req)
+{
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+ return ecb_do_decrypt(req, ctx->dec_key[0]);
+}
+
+static int aria_avx512_set_key(struct crypto_skcipher *tfm, const u8 *key,
+ unsigned int keylen)
+{
+ return aria_set_key(&tfm->base, key, keylen);
+}
+
+static int aria_avx512_ctr_encrypt(struct skcipher_request *req)
+{
+ struct aria_avx512_request_ctx *req_ctx = skcipher_request_ctx(req);
+ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+ struct aria_ctx *ctx = crypto_skcipher_ctx(tfm);
+ struct skcipher_walk walk;
+ unsigned int nbytes;
+ int err;
+
+ err = skcipher_walk_virt(&walk, req, false);
+
+ while ((nbytes = walk.nbytes) > 0) {
+ const u8 *src = walk.src.virt.addr;
+ u8 *dst = walk.dst.virt.addr;
+
+ while (nbytes >= ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE) {
+ kernel_fpu_begin();
+ aria_ops.aria_ctr_crypt_64way(ctx, dst, src,
+ &req_ctx->keystream[0],
+ walk.iv);
+ kernel_fpu_end();
+ dst += ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE;
+ src += ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE;
+ nbytes -= ARIA_GFNI_AVX512_PARALLEL_BLOCK_SIZE;
+ }
+
+ while (nbytes >= ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE) {
+ kernel_fpu_begin();
+ aria_ops.aria_ctr_crypt_32way(ctx, dst, src,
+ &req_ctx->keystream[0],
+ walk.iv);
+ kernel_fpu_end();
+ dst += ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ src += ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ nbytes -= ARIA_AESNI_AVX2_PARALLEL_BLOCK_SIZE;
+ }
+
+ while (nbytes >= ARIA_AESNI_PARALLEL_BLOCK_SIZE) {
+ kernel_fpu_begin();
+ aria_ops.aria_ctr_crypt_16way(ctx, dst, src,
+ &req_ctx->keystream[0],
+ walk.iv);
+ kernel_fpu_end();
+ dst += ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ src += ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ nbytes -= ARIA_AESNI_PARALLEL_BLOCK_SIZE;
+ }
+
+ while (nbytes >= ARIA_BLOCK_SIZE) {
+ memcpy(&req_ctx->keystream[0], walk.iv,
+ ARIA_BLOCK_SIZE);
+ crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
+
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
+
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ ARIA_BLOCK_SIZE);
+ dst += ARIA_BLOCK_SIZE;
+ src += ARIA_BLOCK_SIZE;
+ nbytes -= ARIA_BLOCK_SIZE;
+ }
+
+ if (walk.nbytes == walk.total && nbytes > 0) {
+ memcpy(&req_ctx->keystream[0], walk.iv,
+ ARIA_BLOCK_SIZE);
+ crypto_inc(walk.iv, ARIA_BLOCK_SIZE);
+
+ aria_encrypt(ctx, &req_ctx->keystream[0],
+ &req_ctx->keystream[0]);
+
+ crypto_xor_cpy(dst, src, &req_ctx->keystream[0],
+ nbytes);
+ dst += nbytes;
+ src += nbytes;
+ nbytes = 0;
+ }
+ err = skcipher_walk_done(&walk, nbytes);
+ }
+
+ return err;
+}
+
+static int aria_avx512_init_tfm(struct crypto_skcipher *tfm)
+{
+ crypto_skcipher_set_reqsize(tfm,
+ sizeof(struct aria_avx512_request_ctx));
+
+ return 0;
+}
+
+static struct skcipher_alg aria_algs[] = {
+ {
+ .base.cra_name = "__ecb(aria)",
+ .base.cra_driver_name = "__ecb-aria-avx512",
+ .base.cra_priority = 600,
+ .base.cra_flags = CRYPTO_ALG_INTERNAL,
+ .base.cra_blocksize = ARIA_BLOCK_SIZE,
+ .base.cra_ctxsize = sizeof(struct aria_ctx),
+ .base.cra_module = THIS_MODULE,
+ .min_keysize = ARIA_MIN_KEY_SIZE,
+ .max_keysize = ARIA_MAX_KEY_SIZE,
+ .setkey = aria_avx512_set_key,
+ .encrypt = aria_avx512_ecb_encrypt,
+ .decrypt = aria_avx512_ecb_decrypt,
+ }, {
+ .base.cra_name = "__ctr(aria)",
+ .base.cra_driver_name = "__ctr-aria-avx512",
+ .base.cra_priority = 600,
+ .base.cra_flags = CRYPTO_ALG_INTERNAL |
+ CRYPTO_ALG_SKCIPHER_REQSIZE_LARGE,
+ .base.cra_blocksize = 1,
+ .base.cra_ctxsize = sizeof(struct aria_ctx),
+ .base.cra_module = THIS_MODULE,
+ .min_keysize = ARIA_MIN_KEY_SIZE,
+ .max_keysize = ARIA_MAX_KEY_SIZE,
+ .ivsize = ARIA_BLOCK_SIZE,
+ .chunksize = ARIA_BLOCK_SIZE,
+ .setkey = aria_avx512_set_key,
+ .encrypt = aria_avx512_ctr_encrypt,
+ .decrypt = aria_avx512_ctr_encrypt,
+ .init = aria_avx512_init_tfm,
+ }
+};
+
+static struct simd_skcipher_alg *aria_simd_algs[ARRAY_SIZE(aria_algs)];
+
+static int __init aria_avx512_init(void)
+{
+ const char *feature_name;
+
+ if (!boot_cpu_has(X86_FEATURE_AVX) ||
+ !boot_cpu_has(X86_FEATURE_AVX2) ||
+ !boot_cpu_has(X86_FEATURE_AVX512F) ||
+ !boot_cpu_has(X86_FEATURE_AVX512VL) ||
+ !boot_cpu_has(X86_FEATURE_GFNI) ||
+ !boot_cpu_has(X86_FEATURE_OSXSAVE)) {
+ pr_info("AVX512/GFNI instructions are not detected.\n");
+ return -ENODEV;
+ }
+
+ if (!cpu_has_xfeatures(XFEATURE_MASK_SSE | XFEATURE_MASK_YMM |
+ XFEATURE_MASK_AVX512, &feature_name)) {
+ pr_info("CPU feature '%s' is not supported.\n", feature_name);
+ return -ENODEV;
+ }
+
+ aria_ops.aria_encrypt_16way = aria_aesni_avx_gfni_encrypt_16way;
+ aria_ops.aria_decrypt_16way = aria_aesni_avx_gfni_decrypt_16way;
+ aria_ops.aria_ctr_crypt_16way = aria_aesni_avx_gfni_ctr_crypt_16way;
+ aria_ops.aria_encrypt_32way = aria_aesni_avx2_gfni_encrypt_32way;
+ aria_ops.aria_decrypt_32way = aria_aesni_avx2_gfni_decrypt_32way;
+ aria_ops.aria_ctr_crypt_32way = aria_aesni_avx2_gfni_ctr_crypt_32way;
+ aria_ops.aria_encrypt_64way = aria_gfni_avx512_encrypt_64way;
+ aria_ops.aria_decrypt_64way = aria_gfni_avx512_decrypt_64way;
+ aria_ops.aria_ctr_crypt_64way = aria_gfni_avx512_ctr_crypt_64way;
+
+ return simd_register_skciphers_compat(aria_algs,
+ ARRAY_SIZE(aria_algs),
+ aria_simd_algs);
+}
+
+static void __exit aria_avx512_exit(void)
+{
+ simd_unregister_skciphers(aria_algs, ARRAY_SIZE(aria_algs),
+ aria_simd_algs);
+}
+
+module_init(aria_avx512_init);
+module_exit(aria_avx512_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Taehee Yoo <ap420073@gmail.com>");
+MODULE_DESCRIPTION("ARIA Cipher Algorithm, AVX512/GFNI optimized");
+MODULE_ALIAS_CRYPTO("aria");
+MODULE_ALIAS_CRYPTO("aria-gfni-avx512");
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
` (3 preceding siblings ...)
2022-11-21 0:39 ` [PATCH v6 4/4] crypto: aria: implement aria-avx512 Taehee Yoo
@ 2022-11-21 9:47 ` Borislav Petkov
2022-11-21 11:20 ` Taehee Yoo
4 siblings, 1 reply; 9+ messages in thread
From: Borislav Petkov @ 2022-11-21 9:47 UTC (permalink / raw)
To: Taehee Yoo
Cc: linux-crypto, herbert, davem, tglx, mingo, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
On Mon, Nov 21, 2022 at 12:39:51AM +0000, Taehee Yoo wrote:
> This patchset is to implement aria-avx2 and aria-avx512.
> There are some differences between aria-avx, aria-avx2, and aria-avx512,
> but they are not core logic(s-box, diffusion layer).
From: Documentation/process/submitting-patches.rst
"Don't get discouraged - or impatient
------------------------------------
After you have submitted your change, be patient and wait. Reviewers are
busy people and may not get to your patch right away.
Once upon a time, patches used to disappear into the void without comment,
but the development process works more smoothly than that now. You should
receive comments within a week or so; if that does not happen, make sure
that you have sent your patches to the right place. Wait for a minimum of
^^^^^^^^^^^^^^^^^^^^^
one week before resubmitting or pinging reviewers - possibly longer during
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
busy times like merge windows."
IOW, pls stop spamming with your patchset.
Thx.
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512
2022-11-21 9:47 ` [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Borislav Petkov
@ 2022-11-21 11:20 ` Taehee Yoo
0 siblings, 0 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-11-21 11:20 UTC (permalink / raw)
To: Borislav Petkov
Cc: linux-crypto, herbert, davem, tglx, mingo, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Hi Borislav,
Thank you so much for your comment!
On 11/21/22 18:47, Borislav Petkov wrote:
> On Mon, Nov 21, 2022 at 12:39:51AM +0000, Taehee Yoo wrote:
>> This patchset is to implement aria-avx2 and aria-avx512.
>> There are some differences between aria-avx, aria-avx2, and aria-avx512,
>> but they are not core logic(s-box, diffusion layer).
>
> From: Documentation/process/submitting-patches.rst
>
> "Don't get discouraged - or impatient
> ------------------------------------
>
> After you have submitted your change, be patient and wait. Reviewers are
> busy people and may not get to your patch right away.
>
> Once upon a time, patches used to disappear into the void without
comment,
> but the development process works more smoothly than that now. You
should
> receive comments within a week or so; if that does not happen, make sure
> that you have sent your patches to the right place. Wait for a
minimum of
> ^^^^^^^^^^^^^^^^^^^^^
> one week before resubmitting or pinging reviewers - possibly longer
during
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> busy times like merge windows."
>
> IOW, pls stop spamming with your patchset.
>
> Thx.
>
Sorry for that.
After now, I will wait at least one week for review.
Thanks a lot!
Taehee Yoo
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx
2022-11-21 0:39 ` [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx Taehee Yoo
@ 2022-12-02 10:11 ` Herbert Xu
2022-12-03 5:01 ` Taehee Yoo
0 siblings, 1 reply; 9+ messages in thread
From: Herbert Xu @ 2022-12-02 10:11 UTC (permalink / raw)
To: Taehee Yoo
Cc: linux-crypto, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
On Mon, Nov 21, 2022 at 12:39:53AM +0000, Taehee Yoo wrote:
>
> +#if defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64) || \
> + defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64_MODULE)
Why isn't this IS_ENABLED like the hunk below?
> +
> + /* Offset for fields in aria_ctx */
> + BLANK();
> + OFFSET(ARIA_CTX_enc_key, aria_ctx, enc_key);
> + OFFSET(ARIA_CTX_dec_key, aria_ctx, dec_key);
> + OFFSET(ARIA_CTX_rounds, aria_ctx, rounds);
> +#endif
> +
> if (IS_ENABLED(CONFIG_KVM_INTEL)) {
> BLANK();
> OFFSET(VMX_spec_ctrl, vcpu_vmx, spec_ctrl);
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
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx
2022-12-02 10:11 ` Herbert Xu
@ 2022-12-03 5:01 ` Taehee Yoo
0 siblings, 0 replies; 9+ messages in thread
From: Taehee Yoo @ 2022-12-03 5:01 UTC (permalink / raw)
To: Herbert Xu
Cc: linux-crypto, davem, tglx, mingo, bp, dave.hansen, hpa,
kirill.shutemov, richard, viro, sathyanarayanan.kuppuswamy,
jpoimboe, elliott, x86, jussi.kivilinna, ebiggers
Hi Herbert,
Thank you so much for your review!
On 12/2/22 19:11, Herbert Xu wrote:
> On Mon, Nov 21, 2022 at 12:39:53AM +0000, Taehee Yoo wrote:
>>
>> +#if defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64) || \
>> + defined(CONFIG_CRYPTO_ARIA_AESNI_AVX_X86_64_MODULE)
>
> Why isn't this IS_ENABLED like the hunk below?
>
>> +
>> + /* Offset for fields in aria_ctx */
>> + BLANK();
>> + OFFSET(ARIA_CTX_enc_key, aria_ctx, enc_key);
>> + OFFSET(ARIA_CTX_dec_key, aria_ctx, dec_key);
>> + OFFSET(ARIA_CTX_rounds, aria_ctx, rounds);
>> +#endif
>> +
>> if (IS_ENABLED(CONFIG_KVM_INTEL)) {
>> BLANK();
>> OFFSET(VMX_spec_ctrl, vcpu_vmx, spec_ctrl);
>
> Thanks,
I missed cleaning it up.
I will use IS_ENABLED() instead of defined() in the v7.
Thanks a lot!
Taehee Yoo
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2022-12-03 5:02 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-11-21 0:39 [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 1/4] crypto: aria: add keystream array into request ctx Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 2/4] crypto: aria: do not use magic number offsets of aria_ctx Taehee Yoo
2022-12-02 10:11 ` Herbert Xu
2022-12-03 5:01 ` Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 3/4] crypto: aria: implement aria-avx2 Taehee Yoo
2022-11-21 0:39 ` [PATCH v6 4/4] crypto: aria: implement aria-avx512 Taehee Yoo
2022-11-21 9:47 ` [PATCH v6 0/4] crypto: aria: implement aria-avx2 and aria-avx512 Borislav Petkov
2022-11-21 11:20 ` Taehee Yoo
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).