Skip to content

Commit

Permalink
add aesenc{128,256}kl, aesencwide{128,256}kl
Browse files Browse the repository at this point in the history
  • Loading branch information
herumi committed Dec 19, 2023
1 parent d9c7c99 commit 48551f5
Show file tree
Hide file tree
Showing 3 changed files with 31 additions and 0 deletions.
4 changes: 4 additions & 0 deletions gen/gen_code.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -2026,6 +2026,10 @@ void put64()
{ "aesdec256kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xDF, 8 },
{ "aesdecwide128kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xD8, 1 },
{ "aesdecwide256kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xD8, 3 },
{ "aesenc128kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xDC, 8 },
{ "aesenc256kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xDE, 8 },
{ "aesencwide128kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xD8, 0 },
{ "aesencwide256kl", T_F3|T_0F38, T_MUST_EVEX|T_F3, 0xD8, 2 },
};
for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
const Tbl *p = &tbl[i];
Expand Down
23 changes: 23 additions & 0 deletions test/apx.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -1792,6 +1792,17 @@ CYBOZU_TEST_AUTO(aeskl)
aesdecwide256kl(ptr[rax+rcx*4+0x12]);
aesdecwide256kl(ptr[r30+r29*8+0x34]);

aesenc128kl(xmm15, ptr[rax+rcx*4+0x12]);
aesenc128kl(xmm15, ptr[r30+r29*8+0x34]);

aesenc256kl(xmm15, ptr[rax+rcx*4+0x12]);
aesenc256kl(xmm15, ptr[r30+r29*8+0x34]);

aesencwide128kl(ptr[rax+rcx*4+0x12]);
aesencwide128kl(ptr[r30+r29*8+0x34]);

aesencwide256kl(ptr[rax+rcx*4+0x12]);
aesencwide256kl(ptr[r30+r29*8+0x34]);
}
} c;
const uint8_t tbl[] = {
Expand All @@ -1807,6 +1818,18 @@ CYBOZU_TEST_AUTO(aeskl)
// aesdecwide256kl
0x0f, 0x38, 0xd8, 0x5c, 0x88, 0x12,
0x62, 0x9c, 0x7a, 0x08, 0xd8, 0x5c, 0xee, 0x34,
// aesenc128kl
0xf3, 0x44, 0x0f, 0x38, 0xdc, 0x7c, 0x88, 0x12,
0x62, 0x1c, 0x7a, 0x08, 0xdc, 0x7c, 0xee, 0x34,
// aesenc256kl
0xf3, 0x44, 0x0f, 0x38, 0xde, 0x7c, 0x88, 0x12,
0x62, 0x1c, 0x7a, 0x08, 0xde, 0x7c, 0xee, 0x34,
// aesencwide128kl
0xf3, 0x0f, 0x38, 0xd8, 0x44, 0x88, 0x12,
0x62, 0x9c, 0x7a, 0x08, 0xd8, 0x44, 0xee, 0x34,
// aesencwide256kl
0xf3, 0x0f, 0x38, 0xd8, 0x54, 0x88, 0x12,
0x62, 0x9c, 0x7a, 0x08, 0xd8, 0x54, 0xee, 0x34,
};
const size_t n = sizeof(tbl);
CYBOZU_TEST_EQUAL(c.getSize(), n);
Expand Down
4 changes: 4 additions & 0 deletions xbyak/xbyak_mnemonic.h
Original file line number Diff line number Diff line change
Expand Up @@ -1930,6 +1930,10 @@ void aesdec128kl(const Xmm& x, const Address& addr) { opAESKL(&x, addr, T_F3|T_0
void aesdec256kl(const Xmm& x, const Address& addr) { opAESKL(&x, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xDF); }
void aesdecwide128kl(const Address& addr) { opAESKL(&xmm1, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xD8); }
void aesdecwide256kl(const Address& addr) { opAESKL(&xmm3, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xD8); }
void aesenc128kl(const Xmm& x, const Address& addr) { opAESKL(&x, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xDC); }
void aesenc256kl(const Xmm& x, const Address& addr) { opAESKL(&x, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xDE); }
void aesencwide128kl(const Address& addr) { opAESKL(&xmm0, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xD8); }
void aesencwide256kl(const Address& addr) { opAESKL(&xmm2, addr, T_F3|T_0F38, T_F3|T_MUST_EVEX, 0xD8); }
void ldtilecfg(const Address& addr) { if (opROO(Reg(), addr, tmm0, T_APX|T_0F38|T_W0, 0x49)) return; opVex(tmm0, &tmm0, addr, T_0F38|T_W0, 0x49); }
void sttilecfg(const Address& addr) { if (opROO(Reg(), addr, tmm0, T_APX|T_66|T_0F38|T_W0, 0x49)) return; opVex(tmm0, &tmm0, addr, T_66|T_0F38 | T_W0, 0x49); }
void tileloadd(const Tmm& tm, const Address& addr) { opAMX(tm, addr, T_F2|T_0F38|T_W0, 0x4B); }
Expand Down

0 comments on commit 48551f5

Please sign in to comment.