From: Phoebe Chen <phoebe.chen@sifive.com>
Date: Wed, 4 Oct 2023 08:42:10 +0000 (-0700)
Subject: riscv: Further optimization for single block aes-zvkned encryption.
X-Git-Tag: openssl-3.3.0-alpha1~747
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=42f112284862bb0744d37c2f3301c7163179cf15;p=thirdparty%2Fopenssl.git

riscv: Further optimization for single block aes-zvkned encryption.

Interleave key loading and aes encrypt computing for single block aes.

Signed-off-by: Phoebe Chen <phoebe.chen@sifive.com>

Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21923)
---

diff --git a/crypto/aes/asm/aes-riscv64-zvkned.pl b/crypto/aes/asm/aes-riscv64-zvkned.pl
index e0f5f19ff5c..1cf5fddb5cf 100644
--- a/crypto/aes/asm/aes-riscv64-zvkned.pl
+++ b/crypto/aes/asm/aes-riscv64-zvkned.pl
@@ -273,42 +273,41 @@ ___
 $code .= <<___;
 .p2align 3
 L_enc_128:
-    @{[vsetivli__x0_4_e32_m1_tu_mu]}
+    @{[vsetivli "zero", 4, "e32", "m1", "ta", "ma"]}
+
+    @{[vle32_v $v1, ($INP)]}
 
     @{[vle32_v $v10, ($KEYP)]}
+    @{[vaesz_vs $v1, $v10]}    # with round key w[ 0, 3]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v11, ($KEYP)]}
+    @{[vaesem_vs $v1, $v11]}   # with round key w[ 4, 7]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v12, ($KEYP)]}
+    @{[vaesem_vs $v1, $v12]}   # with round key w[ 8,11]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v13, ($KEYP)]}
+    @{[vaesem_vs $v1, $v13]}   # with round key w[12,15]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v14, ($KEYP)]}
+    @{[vaesem_vs $v1, $v14]}   # with round key w[16,19]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v15, ($KEYP)]}
+    @{[vaesem_vs $v1, $v15]}   # with round key w[20,23]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v16, ($KEYP)]}
+    @{[vaesem_vs $v1, $v16]}   # with round key w[24,27]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v17, ($KEYP)]}
+    @{[vaesem_vs $v1, $v17]}   # with round key w[28,31]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v18, ($KEYP)]}
+    @{[vaesem_vs $v1, $v18]}   # with round key w[32,35]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v19, ($KEYP)]}
+    @{[vaesem_vs $v1, $v19]}   # with round key w[36,39]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v20, ($KEYP)]}
-
-    @{[vle32_v $v1, ($INP)]}
-
-    @{[vaesz_vs $v1, $v10]}    # with round key w[ 0, 3]
-    @{[vaesem_vs $v1, $v11]}   # with round key w[ 4, 7]
-    @{[vaesem_vs $v1, $v12]}   # with round key w[ 8,11]
-    @{[vaesem_vs $v1, $v13]}   # with round key w[12,15]
-    @{[vaesem_vs $v1, $v14]}   # with round key w[16,19]
-    @{[vaesem_vs $v1, $v15]}   # with round key w[20,23]
-    @{[vaesem_vs $v1, $v16]}   # with round key w[24,27]
-    @{[vaesem_vs $v1, $v17]}   # with round key w[28,31]
-    @{[vaesem_vs $v1, $v18]}   # with round key w[32,35]
-    @{[vaesem_vs $v1, $v19]}   # with round key w[36,39]
     @{[vaesef_vs $v1, $v20]}   # with round key w[40,43]
 
     @{[vse32_v $v1, ($OUTP)]}
@@ -320,48 +319,47 @@ ___
 $code .= <<___;
 .p2align 3
 L_enc_192:
-    @{[vsetivli__x0_4_e32_m1_tu_mu]}
+    @{[vsetivli "zero", 4, "e32", "m1", "ta", "ma"]}
+
+    @{[vle32_v $v1, ($INP)]}
 
     @{[vle32_v $v10, ($KEYP)]}
+    @{[vaesz_vs $v1, $v10]}     # with round key w[ 0, 3]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v11, ($KEYP)]}
+    @{[vaesem_vs $v1, $v11]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v12, ($KEYP)]}
+    @{[vaesem_vs $v1, $v12]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v13, ($KEYP)]}
+    @{[vaesem_vs $v1, $v13]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v14, ($KEYP)]}
+    @{[vaesem_vs $v1, $v14]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v15, ($KEYP)]}
+    @{[vaesem_vs $v1, $v15]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v16, ($KEYP)]}
+    @{[vaesem_vs $v1, $v16]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v17, ($KEYP)]}
+    @{[vaesem_vs $v1, $v17]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v18, ($KEYP)]}
+    @{[vaesem_vs $v1, $v18]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v19, ($KEYP)]}
+    @{[vaesem_vs $v1, $v19]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v20, ($KEYP)]}
+    @{[vaesem_vs $v1, $v20]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v21, ($KEYP)]}
+    @{[vaesem_vs $v1, $v21]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v22, ($KEYP)]}
-
-    @{[vle32_v $v1, ($INP)]}
-
-    @{[vaesz_vs $v1, $v10]}     # with round key w[ 0, 3]
-    @{[vaesem_vs $v1, $v11]}
-    @{[vaesem_vs $v1, $v12]}
-    @{[vaesem_vs $v1, $v13]}
-    @{[vaesem_vs $v1, $v14]}
-    @{[vaesem_vs $v1, $v15]}
-    @{[vaesem_vs $v1, $v16]}
-    @{[vaesem_vs $v1, $v17]}
-    @{[vaesem_vs $v1, $v18]}
-    @{[vaesem_vs $v1, $v19]}
-    @{[vaesem_vs $v1, $v20]}
-    @{[vaesem_vs $v1, $v21]}
     @{[vaesef_vs $v1, $v22]}
 
     @{[vse32_v $v1, ($OUTP)]}
@@ -372,54 +370,53 @@ ___
 $code .= <<___;
 .p2align 3
 L_enc_256:
-    @{[vsetivli__x0_4_e32_m1_tu_mu]}
+    @{[vsetivli "zero", 4, "e32", "m1", "ta", "ma"]}
+
+    @{[vle32_v $v1, ($INP)]}
 
     @{[vle32_v $v10, ($KEYP)]}
+    @{[vaesz_vs $v1, $v10]}     # with round key w[ 0, 3]
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v11, ($KEYP)]}
+    @{[vaesem_vs $v1, $v11]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v12, ($KEYP)]}
+    @{[vaesem_vs $v1, $v12]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v13, ($KEYP)]}
+    @{[vaesem_vs $v1, $v13]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v14, ($KEYP)]}
+    @{[vaesem_vs $v1, $v14]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v15, ($KEYP)]}
+    @{[vaesem_vs $v1, $v15]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v16, ($KEYP)]}
+    @{[vaesem_vs $v1, $v16]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v17, ($KEYP)]}
+    @{[vaesem_vs $v1, $v17]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v18, ($KEYP)]}
+    @{[vaesem_vs $v1, $v18]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v19, ($KEYP)]}
+    @{[vaesem_vs $v1, $v19]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v20, ($KEYP)]}
+    @{[vaesem_vs $v1, $v20]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v21, ($KEYP)]}
+    @{[vaesem_vs $v1, $v21]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v22, ($KEYP)]}
+    @{[vaesem_vs $v1, $v22]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v23, ($KEYP)]}
+    @{[vaesem_vs $v1, $v23]}
     addi $KEYP, $KEYP, 16
     @{[vle32_v $v24, ($KEYP)]}
-
-    @{[vle32_v $v1, ($INP)]}
-
-    @{[vaesz_vs $v1, $v10]}     # with round key w[ 0, 3]
-    @{[vaesem_vs $v1, $v11]}
-    @{[vaesem_vs $v1, $v12]}
-    @{[vaesem_vs $v1, $v13]}
-    @{[vaesem_vs $v1, $v14]}
-    @{[vaesem_vs $v1, $v15]}
-    @{[vaesem_vs $v1, $v16]}
-    @{[vaesem_vs $v1, $v17]}
-    @{[vaesem_vs $v1, $v18]}
-    @{[vaesem_vs $v1, $v19]}
-    @{[vaesem_vs $v1, $v20]}
-    @{[vaesem_vs $v1, $v21]}
-    @{[vaesem_vs $v1, $v22]}
-    @{[vaesem_vs $v1, $v23]}
     @{[vaesef_vs $v1, $v24]}
 
     @{[vse32_v $v1, ($OUTP)]}