Skip to contentSkip to conversation

crypto: x86/aria - add missing vzeroupper in AVX2 code

PATCH RESEND Superseded Eric Biggers 2 participants 4 messages Aug 31 – Sep 4

Subsystems (3)
Stable branch x86 Crypto API
lore
Share & tools
lore

Eric Biggers

author
#0 lore raw

PATCH RESEND 1/1 crypto: x86/aria - add missing vzeroupper in AVX2 code

Since the AVX2 optimized ARIA code uses YMM registers, execute vzeroupper before returning from it. This is needed to avoid degrading the performance of any later SSE code that may happen to be executed.

Fixes 37d8d3ae7a58 ("crypto: x86/aria - implement aria-avx2")
Cc [email protected]
Cc Taehee Yoo <[email protected]>
Signed-off-by Eric Biggers <[email protected]>

---

This didn't get picked up by the x86 maintainers yet. Herbert, can you take this through the crypto tree?

 arch/x86/crypto/aria-aesni-avx2-asm_64.S | 6 ++++++
 1 file changed, 6 insertions(+)
arch/x86/crypto/aria-aesni-avx2-asm_64.S +6 -0
index ed53d4f46bd7..fda8cb8a99a8 100644
@@ -982,6 +982,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_encrypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %rax);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_encrypt_32way)

@@ -1007,6 +1008,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_decrypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %rax);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_decrypt_32way)

@@ -1209,6 +1211,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_ctr_crypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %r10);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_ctr_crypt_32way)

@@ -1359,6 +1362,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_encrypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %rax);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_gfni_encrypt_32way)

@@ -1384,6 +1388,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_decrypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %rax);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_gfni_decrypt_32way)

@@ -1428,6 +1433,7 @@ SYM_TYPED_FUNC_START(aria_aesni_avx2_gfni_ctr_crypt_32way)
 		     %ymm8, %ymm9, %ymm10, %ymm11, %ymm12, %ymm13, %ymm14,
 		     %ymm15, %r10);
 
+	vzeroupper;
 	FRAME_END
 	RET;
 SYM_FUNC_END(aria_aesni_avx2_gfni_ctr_crypt_32way)

base-commit: cee9395acd8043be0644b25c34bfa86623f2b935

Signature · 1 line
2.55.0
maintainer
#1 lore raw
Quoted 4 lines · Eric Biggers

Since the AVX2 optimized ARIA code uses YMM registers, execute vzeroupper before returning from it. This is needed to avoid degrading the performance of any later SSE code that may happen to be executed.

Quoted 8 lines
Fixes 37d8d3ae7a58 ("crypto: x86/aria - implement aria-avx2")
Cc [email protected]
Cc Taehee Yoo <[email protected]>
Signed-off-by Eric Biggers <[email protected]>

---

This didn't get picked up by the x86 maintainers yet. Herbert, can you take this through the crypto tree?

OK I could do that. I presume you want this to go into 7.3, right?

Cheers,

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
author
#2 lore raw
Quoted 11 lines · Herbert Xu
Quoted 10 lines · Eric Biggers

Since the AVX2 optimized ARIA code uses YMM registers, execute vzeroupper before returning from it. This is needed to avoid degrading the performance of any later SSE code that may happen to be executed.

Fixes 37d8d3ae7a58 ("crypto: x86/aria - implement aria-avx2")
Cc [email protected]
Cc Taehee Yoo <[email protected]>
Signed-off-by Eric Biggers <[email protected]>

---

Quoted 5 lines
Quoted 2 lines

This didn't get picked up by the x86 maintainers yet. Herbert, can you take this through the crypto tree?

OK I could do that. I presume you want this to go into 7.3, right?

It is a bug fix, so yes it should go into 7.3.

- Eric

maintainer
#3 lore raw
In reply to Eric Biggers
Quoted 8 lines · Eric Biggers

Since the AVX2 optimized ARIA code uses YMM registers, execute vzeroupper before returning from it. This is needed to avoid degrading the performance of any later SSE code that may happen to be executed.

Fixes 37d8d3ae7a58 ("crypto: x86/aria - implement aria-avx2")
Cc [email protected]
Cc Taehee Yoo <[email protected]>
Signed-off-by Eric Biggers <[email protected]>
Quoted 7 lines

---

This didn't get picked up by the x86 maintainers yet. Herbert, can you take this through the crypto tree?

 arch/x86/crypto/aria-aesni-avx2-asm_64.S | 6 ++++++
 1 file changed, 6 insertions(+)

Patch applied. Thanks.

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt