[{"id":3684552,"web_url":"http://patchwork.ozlabs.org/comment/3684552/","msgid":"<2ab6ffc3-816a-4449-a997-2e7098729eac@linux.ibm.com>","list_archive_url":null,"date":"2026-04-30T09:08:40","subject":"Re: [PATCH v3] rs6000: Builtins for AES acceleration instructions\n [RFC02657]","submitter":{"id":88474,"url":"http://patchwork.ozlabs.org/api/people/88474/","name":"Surya Kumari Jangala","email":"jskumari@linux.ibm.com"},"content":"Hi,\n\nOn 30/04/26 11:49 AM, Avinash Jayakar wrote:\n> Hi,\n> \n> Updated the patch with extend.texi documentation. All other files\n> remains the same. Kindly review the documentation.\n> \n> Changes from v2:\n> * Added extend.texi documentation for the new builtins.\n> \n> Changes from v1:\n> * Update commit message.\n> * Add comments for iterators/attributes in crypto.md.\n> * Corrected formatting in crypto.md.\n> * Removed mma attribute for aes builtins.\n> * Updated function names in aes-builtin-1.c.\n> * aes-builtin-1.c now has 1 builtin per function.\n> * Separated out the combined builtin test case in aes-builtin-3.c.\n> \n> Following patch depends on these 2 patches in the following order:\n> 1. mcpu=future: https://gcc.gnu.org/pipermail/gcc-patches/2025-December/703739.html\n> 2. future builtin infra: https://gcc.gnu.org/pipermail/gcc-patches/2026-March/709782.html\n> \n> Bootstrapped and regtested on powerpc64le-linux-gnu with no regressions.\n> \n> Thanks and regards,\n> Avinash Jayakar\n> \n> This patch adds new builtins for AES acceleration instructions which\n> may or may not be supported in a future processor. Note, the names of\n> the builtins may change in future.\n> \n> The following new builtins for AES acceleration can be used with\n> -mcpu=future option:\n> __vector_pair __builtin_aes_encrypt_paired (__vector_pair,\n> \t\t\t\t\t __vector_pair, uint2);\n> __vector_pair __builtin_aes128_encrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes192_encrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes256_encrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes_decrypt_paired (__vector_pair,\n> \t\t\t\t\t __vector_pair, uint2);\n> __vector_pair __builtin_aes128_decrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes192_decrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes256_decrypt_paired (__vector_pair,\n> \t\t\t\t\t\t__vector_pair);\n> __vector_pair __builtin_aes_genlastkey_paired (__vector_pair, uint2);\n> __vector_pair __builtin_aes128_genlastkey_paired (__vector_pair);\n> __vector_pair __builtin_aes192_genlastkey_paired (__vector_pair);\n> __vector_pair __builtin_aes256_genlastkey_paired (__vector_pair);\n> vec_t __builtin_galois_field_mult (vec_t, vec_t, uint1);\n> vec_t __builtin_galois_field_mult_gcm (vec_t, vec_t);\n> vec_t __builtin_galois_field_mult_xts (vec_t, vec_t);\n> \n> 2026-04-30 Avinash Jayakar \n> \n> gcc/ChangeLog:\n> \t* config/rs6000/crypto.md (UNSPEC_XXAESENCP): New unspec entry.\n> \t(UNSPEC_XXAES128ENCP): Likewise.\n> \t(UNSPEC_XXAES192ENCP): Likewise.\n> \t(UNSPEC_XXAES256ENCP): Likewise.\n> \t(UNSPEC_XXAESDECP): Likewise.\n> \t(UNSPEC_XXAES128DECP): Likewise.\n> \t(UNSPEC_XXAES192DECP): Likewise.\n> \t(UNSPEC_XXAES256DECP): Likewise.\n> \t(UNSPEC_XXAESGENLKP): Likewise.\n> \t(UNSPEC_XXAES128GENLKP): Likewise.\n> \t(UNSPEC_XXAES192GENLKP): Likewise.\n> \t(UNSPEC_XXAES256GENLKP): Likewise.\n> \t(UNSPEC_XXGFMUL128): Likewise.\n> \t(UNSPEC_XXGFMUL128GCM): Likewise.\n> \t(UNSPEC_XXGFMUL128XTS): Likewise.\n> \t(AESACC_base_code): New iterator for xxaesencp and xxaesdecp base\n> \tmnemonics.\n> \t(AESACC_code): New iterator for xxaesencp and xxaesdecp extended\n> \tmnemonics.\n> \t(AESGENLKP_code): New iterator for xxaesgenlkp extended mnemonics.\n> \t(AESGF_code): New iterator for xxgfmul128 extended mnemonics.\n> \t(AESACC_base_insn): New attribute iterator for xxaesencp and xxaesdecp\n> \tbase mnemonics.\n> \t(AESACC_insn): New attribute iterator for xxaesencp and xxaesdecp\n> \textended mnemonics.\n> \t(AESGENLKP_insn): New attribute iterator for xxaesgenlkp extended\n> \tmnemonics.\n> \t(AESGF_insn): New attribute iterator for xxgfmul128 extended mnemonics.\n> \t(): New define_insn for xxaesencp and xxaesdecp base\n> \tmnemonics.\n> \t(): New define_insn for xxaesencp and xxaesdecp extended\n> \tmnemonics.\n> \t(): New define_insn for xxaesgenlkp extended mnemonics.\n> \t(xxaesgenlkp): New define_insn for genlkp base mnemonic.\n> \t(): New define_insn for xxgfmul128 extended mnemonics.\n> \t(xxgfmul128): New define_insn for xxgfmul128 base mnemonic.\n> \t* config/rs6000/rs6000-builtins.def: Added new builtin definitions for\n> \tAES acceleration.\n> \t* doc/extend.texi: Add new section for AES acceleration builtins\n> \tfor ISA 3.2.\n\nThe next ISA version number has not yet been finalized. So pls use\n'future ISA' instead of 3.2\n\n> \n> gcc/testsuite/ChangeLog:\n> \t* gcc.target/powerpc/aes-builtin-1.c: New test.\n> \t* gcc.target/powerpc/aes-builtin-2.c: New test.\n> \t* gcc.target/powerpc/aes-builtin-3.c: New test.\n> ---\n> gcc/config/rs6000/crypto.md | 102 +++++++++++-\n> gcc/config/rs6000/rs6000-builtins.def | 46 ++++++\n> gcc/doc/extend.texi | 51 ++++++\n> .../gcc.target/powerpc/aes-builtin-1.c | 155 ++++++++++++++++++\n> .../gcc.target/powerpc/aes-builtin-2.c | 34 ++++\n> .../gcc.target/powerpc/aes-builtin-3.c | 52 ++++++\n> 6 files changed, 439 insertions(+), 1 deletion(-)\n> create mode 100644 gcc/testsuite/gcc.target/powerpc/aes-builtin-1.c\n> create mode 100644 gcc/testsuite/gcc.target/powerpc/aes-builtin-2.c\n> create mode 100644 gcc/testsuite/gcc.target/powerpc/aes-builtin-3.c\n> \n> diff --git a/gcc/config/rs6000/crypto.md b/gcc/config/rs6000/crypto.md\n> index f91791673c9..a25d9f75439 100644\n> --- a/gcc/config/rs6000/crypto.md\n> +++ b/gcc/config/rs6000/crypto.md\n> @@ -35,7 +35,22 @@ (define_c_enum \"unspec\"\n> UNSPEC_VSBOX\n> UNSPEC_VSHASIGMA\n> UNSPEC_VPERMXOR\n> - UNSPEC_VPMSUM])\n> + UNSPEC_VPMSUM\n> + UNSPEC_XXAESENCP\n> + UNSPEC_XXAES128ENCP\n> + UNSPEC_XXAES192ENCP\n> + UNSPEC_XXAES256ENCP\n> + UNSPEC_XXAESDECP\n> + UNSPEC_XXAES128DECP\n> + UNSPEC_XXAES192DECP\n> + UNSPEC_XXAES256DECP\n> + UNSPEC_XXAESGENLKP\n> + UNSPEC_XXAES128GENLKP\n> + UNSPEC_XXAES192GENLKP\n> + UNSPEC_XXAES256GENLKP\n> + UNSPEC_XXGFMUL128\n> + UNSPEC_XXGFMUL128GCM\n> + UNSPEC_XXGFMUL128XTS])\n> \n> ;; Iterator for VPMSUM/VPERMXOR\n> (define_mode_iterator CR_mode [V16QI V8HI V4SI V2DI])\n> @@ -62,6 +77,40 @@ (define_int_attr CR_insn [(UNSPEC_VCIPHER \"vcipher\")\n> \t\t\t (UNSPEC_VCIPHERLAST \"vcipherlast\")\n> \t\t\t (UNSPEC_VNCIPHERLAST \"vncipherlast\")])\n> \n> +;; Iterator and attribute for AES encrypt/decrypt\n> +(define_int_iterator AESACC_base_code [UNSPEC_XXAESENCP\n> +\t\t\t\t UNSPEC_XXAESDECP])\n> +(define_int_attr AESACC_base_insn [(UNSPEC_XXAESENCP \"xxaesencp\")\n> +\t\t\t\t (UNSPEC_XXAESDECP \"xxaesdecp\")])\n> +\n> +;; Iterator and attribute for AES encrypt/decrypt extended mnemonics\n> +(define_int_iterator AESACC_code [UNSPEC_XXAES128ENCP\n> +\t\t\t\t UNSPEC_XXAES192ENCP\n> +\t\t\t\t UNSPEC_XXAES256ENCP\n> +\t\t\t\t UNSPEC_XXAES128DECP\n> +\t\t\t\t UNSPEC_XXAES192DECP\n> +\t\t\t\t UNSPEC_XXAES256DECP])\n> +(define_int_attr AESACC_insn [(UNSPEC_XXAES128ENCP \"xxaes128encp\")\n> +\t\t\t (UNSPEC_XXAES192ENCP \"xxaes192encp\")\n> +\t\t\t (UNSPEC_XXAES256ENCP \"xxaes256encp\")\n> +\t\t\t (UNSPEC_XXAES128DECP \"xxaes128decp\")\n> +\t\t\t (UNSPEC_XXAES192DECP \"xxaes192decp\")\n> +\t\t\t (UNSPEC_XXAES256DECP \"xxaes256decp\")])\n> +\n> +;; Iterator and attribute for AES gen last key extended mnemonics\n> +(define_int_iterator AESGENLKP_code [UNSPEC_XXAES128GENLKP\n> +\t\t\t\t UNSPEC_XXAES192GENLKP\n> +\t\t\t\t UNSPEC_XXAES256GENLKP])\n> +(define_int_attr AESGENLKP_insn [(UNSPEC_XXAES128GENLKP \"xxaes128genlkp\")\n> +\t\t\t\t (UNSPEC_XXAES192GENLKP \"xxaes192genlkp\")\n> +\t\t\t\t (UNSPEC_XXAES256GENLKP \"xxaes256genlkp\")])\n> +\n> +;; Iterator and attribute for AES galois field mult extended mnemonics\n> +(define_int_iterator AESGF_code [UNSPEC_XXGFMUL128GCM\n> +\t\t\t\t UNSPEC_XXGFMUL128XTS])\n> +(define_int_attr AESGF_insn [(UNSPEC_XXGFMUL128GCM \"xxgfmul128gcm\")\n> +\t\t\t (UNSPEC_XXGFMUL128XTS \"xxgfmul128xts\")])\n> +\n> ;; 2 operand crypto instructions\n> (define_insn \"crypto__\"\n> [(set (match_operand:CR_vqdi 0 \"register_operand\" \"=v\")\n> @@ -111,3 +160,54 @@ (define_insn \"crypto_vshasigma\"\n> \"TARGET_CRYPTO\"\n> \"vshasigma %0,%1,%2,%3\"\n> [(set_attr \"type\" \"vecsimple\")])\n> +\n> +;; AES acceleration instructions\n> +\n> +(define_insn \"\"\n> + [(set (match_operand:OO 0 \"vsx_register_operand\" \"=wa\")\n> +\t(unspec:OO [(match_operand:OO 1 \"vsx_register_operand\" \"wa\")\n> +\t\t (match_operand:OO 2 \"vsx_register_operand\" \"wa\")\n> +\t\t (match_operand:SI 3 \"const_0_to_3_operand\" \"n\")]\n> +\t\t AESACC_base_code))]\n> + \"TARGET_FUTURE\"\n> + \" %x0,%x1,%x2,%3\")\n> +\n> +(define_insn \"\"\n> + [(set (match_operand:OO 0 \"vsx_register_operand\" \"=wa\")\n> + (unspec:OO [(match_operand:OO 1 \"vsx_register_operand\" \"wa\")\n> +\t\t(match_operand:OO 2 \"vsx_register_operand\" \"wa\")]\n> +\t AESACC_code))]\n> + \"TARGET_FUTURE\"\n> + \" %x0,%x1,%x2\")\n> +\n> +(define_insn \"xxaesgenlkp\"\n> + [(set (match_operand:OO 0 \"vsx_register_operand\" \"=wa\")\n> + (unspec:OO [(match_operand:OO 1 \"vsx_register_operand\" \"wa\")\n> +\t\t(match_operand:SI 2 \"const_0_to_3_operand\" \"n\")]\n> +\t UNSPEC_XXAESGENLKP))]\n> + \"TARGET_FUTURE\"\n> + \"xxaesgenlkp %x0,%x1,%2\")\n> +\n> +(define_insn \"\"\n> + [(set (match_operand:OO 0 \"vsx_register_operand\" \"=wa\")\n> + (unspec:OO [(match_operand:OO 1 \"vsx_register_operand\" \"wa\")]\n> +\t AESGENLKP_code))]\n> + \"TARGET_FUTURE\"\n> + \" %x0,%x1\")\n> +\n> +(define_insn \"xxgfmul128\"\n> + [(set (match_operand:V16QI 0 \"vsx_register_operand\" \"=wa\")\n> + (unspec:V16QI [(match_operand:V16QI 1 \"vsx_register_operand\" \"wa\")\n> +\t\t (match_operand:V16QI 2 \"vsx_register_operand\" \"wa\")\n> +\t\t (match_operand:SI 3 \"const_0_to_1_operand\" \"n\")]\n> +\t\t UNSPEC_XXGFMUL128))]\n> + \"TARGET_FUTURE\"\n> + \"xxgfmul128 %x0,%x1,%x2,%3\")\n> +\n> +(define_insn \"\"\n> + [(set (match_operand:V16QI 0 \"vsx_register_operand\" \"=wa\")\n> + (unspec:V16QI [(match_operand:V16QI 1 \"vsx_register_operand\" \"wa\")\n> +\t\t (match_operand:V16QI 2 \"vsx_register_operand\" \"wa\")]\n> +\t\t AESGF_code))]\n> + \"TARGET_FUTURE\"\n> + \" %x0,%x1,%x2\")\n> diff --git a/gcc/config/rs6000/rs6000-builtins.def b/gcc/config/rs6000/rs6000-builtins.def\n> index 7e5a4fb96e7..0d1529b71d4 100644\n> --- a/gcc/config/rs6000/rs6000-builtins.def\n> +++ b/gcc/config/rs6000/rs6000-builtins.def\n> @@ -3924,3 +3924,49 @@\n> \n> void __builtin_vsx_stxvp (v256, unsigned long, const v256 *);\n> STXVP nothing {mma,pair}\n> +\n> +[future]\n> + const v256 __builtin_aes_encrypt_paired (v256, v256, const int<2>);\n> + XXAESENCP xxaesencp {}\n> +\n> + const v256 __builtin_aes128_encrypt_paired (v256, v256);\n> + XXAES128ENCP xxaes128encp {}\n> +\n> + const v256 __builtin_aes192_encrypt_paired (v256, v256);\n> + XXAES192ENCP xxaes192encp {}\n> +\n> + const v256 __builtin_aes256_encrypt_paired (v256, v256);\n> + XXAES256ENCP xxaes256encp {}\n> +\n> + const v256 __builtin_aes_decrypt_paired (v256, v256, const int<2>);\n> + XXAESDECP xxaesdecp {}\n> +\n> + const v256 __builtin_aes128_decrypt_paired (v256, v256);\n> + XXAES128DECP xxaes128decp {}\n> +\n> + const v256 __builtin_aes192_decrypt_paired (v256, v256);\n> + XXAES192DECP xxaes192decp {}\n> +\n> + const v256 __builtin_aes256_decrypt_paired (v256, v256);\n> + XXAES256DECP xxaes256decp {}\n> +\n> + const v256 __builtin_aes_genlastkey_paired (v256, const int<2>);\n> + XXAESGENLKP xxaesgenlkp {}\n> +\n> + const v256 __builtin_aes128_genlastkey_paired (v256);\n> + XXAES128GENLKP xxaes128genlkp {}\n> +\n> + const v256 __builtin_aes192_genlastkey_paired (v256);\n> + XXAES192GENLKP xxaes192genlkp {}\n> +\n> + const v256 __builtin_aes256_genlastkey_paired (v256);\n> + XXAES256GENLKP xxaes256genlkp {}\n> +\n> + const vuc __builtin_galois_field_mult (vuc, vuc, const int<1>);\n> + XXGFMUL128 xxgfmul128 {}\n> +\n> + const vuc __builtin_galois_field_mult_gcm (vuc, vuc);\n> + XXGFMUL128GCM xxgfmul128gcm {}\n> +\n> + const vuc __builtin_galois_field_mult_xts (vuc, vuc);\n> + XXGFMUL128XTS xxgfmul128xts {}\n> diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi\n> index 0faa5323ce1..4bf1167e1b3 100644\n> --- a/gcc/doc/extend.texi\n> +++ b/gcc/doc/extend.texi\n> @@ -24696,6 +24696,7 @@ The PVIPR documents the following overloaded functions:\n> * PowerPC AltiVec Built-in Functions Available on ISA 2.07::\n> * PowerPC AltiVec Built-in Functions Available on ISA 3.0::\n> * PowerPC AltiVec Built-in Functions Available on ISA 3.1::\n> +* PowerPC AltiVec Built-in Functions Available on ISA 3.2 (Future)::\n\nRemove 3.2 here and everywhere else it is mentioned.\n\n> @end menu\n> \n> @node PowerPC AltiVec Built-in Functions on ISA 2.05\n> @@ -26729,6 +26730,56 @@ vector unsigned char);\n> vector unsigned char);\n> @end smallexample\n> \n> +@node PowerPC AltiVec Built-in Functions Available on ISA 3.2 (Future)\n> +@subsubsection PowerPC AltiVec Built-in Functions Available on ISA 3.2 (Future)\n\nRename as: \nPowerPC AltiVec/VSX Built-in Functions Available on Future ISA\n\nThe term VSX is important because instructions generated by the new builtins are\nVSX instructions.\n\n> +\n> +The following additional built-in functions are also available for the\n\nRemove 'also'.\n\n> +PowerPC family of processors, starting with ISA 3.2 (@option{-mcpu=future}):\n\nDon't mention -mcpu-future.\n\n> +\n> +Following built-ins can be used for AES acceleration in hardware.\n\nCan you check \"PowerPC Matrix-Multiply Assist Built-in Functions\" and try to come up \nwith something similar?\n\n-Surya\n\n> +\n> +@smallexample\n> +__vector_pair __builtin_aes_encrypt_paired (__vector_pair, __vector_pair,\n> + int);\n> +\n> +__vector_pair __builtin_aes128_encrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes192_encrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes256_encrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes_decrypt_paired (__vector_pair, __vector_pair,\n> + int);\n> +\n> +__vector_pair __builtin_aes128_decrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes192_decrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes256_decrypt_paired (__vector_pair, __vector_pair);\n> +\n> +__vector_pair __builtin_aes_genlastkey_paired (__vector_pair, int);\n> +\n> +__vector_pair __builtin_aes128_genlastkey_paired (__vector_pair);\n> +\n> +__vector_pair __builtin_aes192_genlastkey_paired (__vector_pair);\n> +\n> +__vector_pair __builtin_aes256_genlastkey_paired (__vector_pair);\n> +\n> +vec_t __builtin_galois_field_mult (vec_t, vec_t, int);\n> +\n> +vec_t __builtin_galois_field_mult_gcm (vec_t, vec_t);\n> +\n> +vec_t __builtin_galois_field_mult_xts (vec_t, vec_t);\n> +@end smallexample\n> +\n> +The third argument to @var{__builtin_aes_encrypt_paired} and\n> +@var{__builtin_aes_decrypt_paired} must be a constant integer that is 0, 1 or\n> +2. The values correspond to 128, 192 and 256 bit AES encryption/decryption\n> +respectively.\n> +The third argument to @var{__builtin_galois_field_mult} must be a constant\n> +integer that is 0 or 1. The values correspond to gcm and xts variant\n> +respectively.\n> +\n> @node PowerPC Hardware Transactional Memory Built-in Functions\n> @subsection PowerPC Hardware Transactional Memory Built-in Functions\n> GCC provides two interfaces for accessing the Hardware Transactional\n> diff --git a/gcc/testsuite/gcc.target/powerpc/aes-builtin-1.c b/gcc/testsuite/gcc.target/powerpc/aes-builtin-1.c\n> new file mode 100644\n> index 00000000000..2777fcdc7ae\n> --- /dev/null\n> +++ b/gcc/testsuite/gcc.target/powerpc/aes-builtin-1.c\n> @@ -0,0 +1,155 @@\n> +/* { dg-do compile } */\n> +/* { dg-options \"-mdejagnu-cpu=future -O2\" } */\n> +\n> +void\n> +aes_enc_pair_0 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_encrypt_paired (t, k, 0);\n> + *res = c;\n> +}\n> +void\n> +aes_enc_pair_1 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_encrypt_paired (t, k, 1);\n> + *res = c;\n> +}\n> +void\n> +aes_enc_pair_2 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_encrypt_paired (t, k, 2);\n> + *res = c;\n> +}\n> +void\n> +aes_enc_pair_128 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes128_encrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_enc_pair_192 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes192_encrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_enc_pair_256 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes256_encrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_0 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_decrypt_paired (t, k, 0);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_1 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_decrypt_paired (t, k, 1);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_2 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_decrypt_paired (t, k, 2);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_128 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes128_decrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_192 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes192_decrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_256 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes256_decrypt_paired (t, k);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_0 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_genlastkey_paired (k, 0);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_1 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_genlastkey_paired (k, 1);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_2 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes_genlastkey_paired (k, 2);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_128 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes128_genlastkey_paired (k);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_192 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes192_genlastkey_paired (k);\n> + *res = c;\n> +}\n> +void\n> +aes_genlastkey_paired_256 (__vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair k = *key;\n> + __vector_pair c = __builtin_aes256_genlastkey_paired (k);\n> + *res = c;\n> +}\n> +\n> +\n> +/* { dg-final { scan-assembler-times {\\mxxaesencp\\M} 3 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes128encp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes192encp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes256encp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaesdecp\\M} 3 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes128decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes192decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes256decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaesgenlkp\\M} 3 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes128genlkp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes192genlkp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes256genlkp\\M} 1 } } */\n> diff --git a/gcc/testsuite/gcc.target/powerpc/aes-builtin-2.c b/gcc/testsuite/gcc.target/powerpc/aes-builtin-2.c\n> new file mode 100644\n> index 00000000000..98f16ad0fe8\n> --- /dev/null\n> +++ b/gcc/testsuite/gcc.target/powerpc/aes-builtin-2.c\n> @@ -0,0 +1,34 @@\n> +/* { dg-do compile } */\n> +/* { dg-options \"-mdejagnu-cpu=future -O2\" } */\n> +\n> +typedef unsigned char vec_t __attribute__((vector_size(16)));\n> +\n> +void\n> +gfmul (vec_t *a, vec_t *b, vec_t *res)\n> +{\n> + vec_t A = *a;\n> + vec_t B = *b;\n> + vec_t R = __builtin_galois_field_mult (A, B, 0);\n> + R = __builtin_galois_field_mult (R, B, 1);\n> + *res = R;\n> +}\n> +void\n> +gfmul_gcm (vec_t *a, vec_t *b, vec_t *res)\n> +{\n> + vec_t A = *a;\n> + vec_t B = *b;\n> + vec_t R = __builtin_galois_field_mult_gcm (A, B);\n> + *res = R;\n> +}\n> +void\n> +gfmul_xts (vec_t *a, vec_t *b, vec_t *res)\n> +{\n> + vec_t A = *a;\n> + vec_t B = *b;\n> + vec_t R = __builtin_galois_field_mult_xts (A, B);\n> + *res = R;\n> +}\n> +\n> +/* { dg-final { scan-assembler-times {\\mxxgfmul128\\M} 2 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxgfmul128gcm\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxgfmul128xts\\M} 1 } } */\n> diff --git a/gcc/testsuite/gcc.target/powerpc/aes-builtin-3.c b/gcc/testsuite/gcc.target/powerpc/aes-builtin-3.c\n> new file mode 100644\n> index 00000000000..007bf878afb\n> --- /dev/null\n> +++ b/gcc/testsuite/gcc.target/powerpc/aes-builtin-3.c\n> @@ -0,0 +1,52 @@\n> +/* { dg-do compile } */\n> +/* { dg-options \"-mdejagnu-cpu=future -O2\" } */\n> +\n> +void\n> +aes_dec_pair (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair lk = __builtin_aes_genlastkey_paired (k, 0);\n> + __vector_pair c = __builtin_aes_decrypt_paired (t, lk, 0);\n> + lk = __builtin_aes_genlastkey_paired (k, 1);\n> + c = __builtin_aes_decrypt_paired (c, lk, 1);\n> + lk = __builtin_aes_genlastkey_paired (k, 2);\n> + c = __builtin_aes_decrypt_paired (c, lk, 2);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_128 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair lk = __builtin_aes128_genlastkey_paired (k);\n> + __vector_pair c = __builtin_aes128_decrypt_paired (t, lk);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_192 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair lk = __builtin_aes192_genlastkey_paired (k);\n> + __vector_pair c = __builtin_aes192_decrypt_paired (t, lk);\n> + *res = c;\n> +}\n> +void\n> +aes_dec_pair_256 (__vector_pair *text, __vector_pair *key, __vector_pair *res)\n> +{\n> + __vector_pair t = *text;\n> + __vector_pair k = *key;\n> + __vector_pair lk = __builtin_aes256_genlastkey_paired (k);\n> + __vector_pair c = __builtin_aes256_decrypt_paired (t, lk);\n> + *res = c;\n> +}\n> +\n> +/* { dg-final { scan-assembler-times {\\mxxaesdecp\\M} 3 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes128decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes192decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes256decp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaesgenlkp\\M} 3 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes128genlkp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes192genlkp\\M} 1 } } */\n> +/* { dg-final { scan-assembler-times {\\mxxaes256genlkp\\M} 1 } } */\n> \\ No newline at end of file","headers":{"Return-Path":"","X-Original-To":["incoming@patchwork.ozlabs.org","gcc-patches@gcc.gnu.org"],"Delivered-To":["patchwork-incoming@legolas.ozlabs.org","gcc-patches@gcc.gnu.org"],"Authentication-Results":["legolas.ozlabs.org;\n\tdkim=pass (2048-bit key;\n unprotected) header.d=ibm.com header.i=@ibm.com header.a=rsa-sha256\n header.s=pp1 header.b=rlpUjmNL;\n\tdkim-atps=neutral","legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org\n (client-ip=38.145.34.32; helo=vm01.sourceware.org;\n envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org;\n receiver=patchwork.ozlabs.org)","sourceware.org;\n\tdkim=pass (2048-bit key,\n unprotected) header.d=ibm.com header.i=@ibm.com header.a=rsa-sha256\n header.s=pp1 header.b=rlpUjmNL","sourceware.org;\n dmarc=none (p=none dis=none) header.from=linux.ibm.com","sourceware.org; spf=pass smtp.mailfrom=linux.ibm.com","server2.sourceware.org;\n arc=none smtp.remote-ip=148.163.158.5"],"Received":["from vm01.sourceware.org (vm01.sourceware.org [38.145.34.32])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519 server-signature ECDSA (secp384r1) server-digest SHA384)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4g5pJx59v0z1yGq\n\tfor ; Thu, 30 Apr 2026 19:09:31 +1000 (AEST)","from vm01.sourceware.org (localhost [127.0.0.1])\n\tby sourceware.org (Postfix) with ESMTP id 44AE9436A044\n\tfor ; Thu, 30 Apr 2026 09:09:25 +0000 (GMT)","from mx0b-001b2d01.pphosted.com (mx0b-001b2d01.pphosted.com\n [148.163.158.5])\n by sourceware.org (Postfix) with ESMTPS id C0DFC436A064;\n Thu, 30 Apr 2026 09:08:50 +0000 (GMT)","from pps.filterd (m0353725.ppops.net [127.0.0.1])\n by mx0a-001b2d01.pphosted.com (8.18.1.11/8.18.1.11) with ESMTP id\n 63U0IYS43636208; Thu, 30 Apr 2026 09:08:50 GMT","from ppma22.wdc07v.mail.ibm.com\n (5c.69.3da9.ip4.static.sl-reverse.com [169.61.105.92])\n by mx0a-001b2d01.pphosted.com (PPS) with ESMTPS id 4drm1e56v4-1\n (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT);\n Thu, 30 Apr 2026 09:08:49 +0000 (GMT)","from pps.filterd (ppma22.wdc07v.mail.ibm.com [127.0.0.1])\n by ppma22.wdc07v.mail.ibm.com (8.18.1.7/8.18.1.7) with ESMTP id\n 63U98nV0017733;\n Thu, 30 Apr 2026 09:08:49 GMT","from smtprelay07.wdc07v.mail.ibm.com ([172.16.1.74])\n by ppma22.wdc07v.mail.ibm.com (PPS) with ESMTPS id 4ds8aw2a2d-1\n (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT);\n Thu, 30 Apr 2026 09:08:49 +0000 (GMT)","from smtpav01.dal12v.mail.ibm.com (smtpav01.dal12v.mail.ibm.com\n [10.241.53.100])\n by smtprelay07.wdc07v.mail.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id\n 63U98lMV28115610\n (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK);\n Thu, 30 Apr 2026 09:08:47 GMT","from smtpav01.dal12v.mail.ibm.com (unknown [127.0.0.1])\n by IMSVA (Postfix) with ESMTP id 6540F58058;\n Thu, 30 Apr 2026 09:08:47 +0000 (GMT)","from smtpav01.dal12v.mail.ibm.com (unknown [127.0.0.1])\n by IMSVA (Postfix) with ESMTP id 6F63658057;\n Thu, 30 Apr 2026 09:08:42 +0000 (GMT)","from [9.124.223.41] (unknown [9.124.223.41])\n by smtpav01.dal12v.mail.ibm.com (Postfix) with ESMTP;\n Thu, 30 Apr 2026 09:08:42 +0000 (GMT)"],"DKIM-Filter":["OpenDKIM Filter v2.11.0 sourceware.org 44AE9436A044","OpenDKIM Filter v2.11.0 sourceware.org C0DFC436A064"],"DMARC-Filter":"OpenDMARC Filter v1.4.2 sourceware.org C0DFC436A064","ARC-Filter":"OpenARC Filter v1.0.0 sourceware.org C0DFC436A064","ARC-Seal":"i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1777540130; cv=none;\n b=Fy70TdqrE3oKx/fGXW5IPCxBQiu0ttmt1IJR85aWJEp5rbnSXj7YB0fIGXtPK9qDaiph0yN1AciDWCraKKzsPUYJwRWW1fGJEzJRf5FsQH/WQV99vB/UzM5w8jV41JdUUiAV/1Gik7YOE25s7MhphIvVbobvDplroB7+lLkMXLk=","ARC-Message-Signature":"i=1; a=rsa-sha256; d=sourceware.org; s=key;\n t=1777540130; c=relaxed/simple;\n bh=7zMbjYGY0Fo/poCqhCoyFERHFtLmhxgH4UEkYtDYuCg=;\n h=DKIM-Signature:Message-ID:Date:MIME-Version:Subject:To:From;\n b=kExVmyVx915PK4vUBpw+7pexNpxOGStYznrGlShCbTLam0di3rZsBwZF78cHA6jHW75hdl3qhCzgpTt2lx2wD4cWWzsnyTJUv73RtRh790vj91j6SZXuzrP/spNwVPxPD7mzdLgdNXI1gnzYKs2g+6edjEoix+jxhTrxT2Tu3Ss=","ARC-Authentication-Results":"i=1; server2.sourceware.org","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed; d=ibm.com; h=cc\n :content-transfer-encoding:content-type:date:from:in-reply-to\n :message-id:mime-version:references:subject:to; s=pp1; bh=ZNmVOL\n UGFcP9OYNQYhlfw+ncev7q8F5pbtZ79cYrZUQ=; b=rlpUjmNLmqyddX2pS2To86\n O4ZLgsnezA9xEnaktI3gPD1I5/EHETsekxDgtpxYxdEqGhBGg2DoxLuklZf7WcTO\n 7/CBHVPrGROXNvuI6E+E8N53Pma4u4SNY+WrkjnPpF3DbenbDZ8Mlewf1d+teU/l\n x8m2Yr4NPKDMOIXT/4D4RDDjxi+viTdmiKpIz0xuMp5KF/C8BOfpCGEtsgBKPn9+\n w8uBek0rPctMsdV/EPfH68WzrXNmTSVQ1jMXMQH0FPjKI5GxVhTlYTdzrbzoV8sX\n +MGuRZXlTqqqh52LAq+hYmztOhj1YdLMDartEbjQpsgdsljf9tLXDRywsPc1qxYQ\n ==","Message-ID":"<2ab6ffc3-816a-4449-a997-2e7098729eac@linux.ibm.com>","Date":"Thu, 30 Apr 2026 14:38:40 +0530","MIME-Version":"1.0","User-Agent":"Mozilla Thunderbird","Subject":"Re: [PATCH v3] rs6000: Builtins for AES acceleration instructions\n [RFC02657]","To":"Avinash Jayakar , gcc-patches@gcc.gnu.org","Cc":"segher@kernel.crashing.org, meissner@linux.ibm.com, dje.gcc@gmail.com,\n linkw@gcc.gnu.org, jeevitha@linux.ibm.com, kishan@linux.ibm.com,\n mmatti@linux.ibm.com, vijay@linux.ibm.com","References":"<20260430061943.532637-1-avinashd@linux.ibm.com>","From":"Surya Kumari Jangala ","Content-Language":"en-US","In-Reply-To":"<20260430061943.532637-1-avinashd@linux.ibm.com>","Content-Type":"text/plain; charset=UTF-8","Content-Transfer-Encoding":"7bit","X-TM-AS-GCONF":"00","X-Proofpoint-Reinject":"loops=2 maxloops=12","X-Proofpoint-ORIG-GUID":"vkNqqEv8UxTOv7aLF2VPlJejAgMcZH9b","X-Proofpoint-Spam-Details-Enc":"AW1haW4tMjYwNDMwMDA4NyBTYWx0ZWRfX1nM7CLc4/Rju\n qIcmMlHrhPFtlvwWAYb4zcS1O/OSWbMGmqO99aID+Y1RE36RHt/OQ5mQuihSKterBDcf8f0JksF\n KAqUgOn5UAteZH1Lve1k1OIF8Yk55Gzn6BLWAjVBFWULV6epTBm63hv6NAK3Hl/JT82YzUtqal3\n st7CvkpbzXcy0JMNFG8YACl2Tn4PBaW9EH1p7ONjadd/Lll3lnSmamlHeA/hchBTOtKtC0v9tMu\n 95SBJXiTc81sMnrMu3OGtaEFQz9+zYLjOA54dLmlPemHvmmvNLwjsxqKIsg3gI36SH3nws+Gavg\n Q9oAFp2b0JwYSBXMzjKKnbeEc7KESVeiDlK+jrA73ac4J4Z6jJrYcytzAzl3oOo3GLsQIMj3UFl\n iD+N+9g7IGSkyy0zy2erus4e19I5nTKWR0ESuYbYPlzM4R28Jypcl/I/c52WG26xt1hrQnVy2UW\n 7sF8SvnMHC0ORsWGM4Q==","X-Authority-Analysis":"v=2.4 cv=VZLH+lp9 c=1 sm=1 tr=0 ts=69f31c22 cx=c_pps\n a=5BHTudwdYE3Te8bg5FgnPg==:117 a=5BHTudwdYE3Te8bg5FgnPg==:17\n a=IkcTkHD0fZMA:10 a=A5OVakUREuEA:10 a=VkNPw1HP01LnGYTKEx00:22\n a=RnoormkPH1_aCDwRdu11:22 a=V8glGbnc2Ofi9Qvn3v5h:22 a=mDV3o1hIAAAA:8\n a=VnNF1IyMAAAA:8 a=A9N1LdzQasVdY6276QEA:9 a=QEXdDO2ut3YA:10","X-Proofpoint-GUID":"l-J4_hI9sOune4leWWHJa-GuM01L1-5G","X-Proofpoint-Virus-Version":"vendor=baseguard\n engine=ICAP:2.0.293,Aquarius:18.0.1143,Hydra:6.1.51,FMLib:17.12.100.49\n definitions=2026-04-30_02,2026-04-28_01,2025-10-01_01","X-Proofpoint-Spam-Details":"rule=outbound_notspam policy=outbound score=0\n clxscore=1015 suspectscore=0 adultscore=0 lowpriorityscore=0 phishscore=0\n spamscore=0 malwarescore=0 bulkscore=0 priorityscore=1501 impostorscore=0\n classifier=typeunknown authscore=0 authtc= authcc= route=outbound adjust=0\n reason=mlx scancount=1 engine=8.22.0-2604200000 definitions=main-2604300087","X-BeenThere":"gcc-patches@gcc.gnu.org","X-Mailman-Version":"2.1.30","Precedence":"list","List-Id":"Gcc-patches mailing list ","List-Unsubscribe":",\n ","List-Archive":"","List-Post":"","List-Help":"","List-Subscribe":",\n ","Errors-To":"gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org"}}]