@@ -53,6 +53,8 @@ typedef enum S390Opcode {
RIL_LLILF = 0xc00f,
RIL_NIHF = 0xc00a,
RIL_NILF = 0xc00b,
+ RIL_OIHF = 0xc00c,
+ RIL_OILF = 0xc00d,
RI_AGHI = 0xa70b,
RI_AHI = 0xa70a,
@@ -70,6 +72,10 @@ typedef enum S390Opcode {
RI_NIHL = 0xa505,
RI_NILH = 0xa506,
RI_NILL = 0xa507,
+ RI_OIHH = 0xa508,
+ RI_OIHL = 0xa509,
+ RI_OILH = 0xa50a,
+ RI_OILL = 0xa50b,
RRE_AGR = 0xb908,
RRE_CGR = 0xb920,
@@ -668,6 +674,47 @@ static void tgen64_andi(TCGContext *s, TCGReg dest, tcg_target_ulong val)
tgen64_andi(s, dest, val | 0x00000000ffffffffull);
}
+static void tgen64_ori(TCGContext *s, TCGReg dest, tcg_target_ulong val)
+{
+ static const S390Opcode oi_insns[4] = {
+ RI_OILL, RI_OILH, RI_OIHL, RI_OIHH
+ };
+ static const S390Opcode nif_insns[2] = {
+ RIL_OILF, RIL_OIHF
+ };
+
+ int i;
+
+ /* Zero-th, look for no-op. */
+ if (val == 0) {
+ return;
+ }
+
+ /* First, try all 32-bit insns that can perform it in one go. */
+ for (i = 0; i < 4; i++) {
+ tcg_target_ulong mask = (0xffffull << i*16);
+ if ((val & mask) != 0 && (val & ~mask) == 0) {
+ tcg_out_insn_RI(s, oi_insns[i], dest, val >> i*16);
+ return;
+ }
+ }
+
+ /* Second, try all 48-bit insns that can perform it in one go. */
+ for (i = 0; i < 2; i++) {
+ tcg_target_ulong mask = (0xffffffffull << i*32);
+ if ((val & mask) != 0 && (val & ~mask) == 0) {
+ tcg_out_insn_RIL(s, nif_insns[i], dest, val >> i*32);
+ return;
+ }
+ }
+
+ /* Last, perform the OR via sequential modifications to the
+ high and low parts. Do this via recursion to handle 16-bit
+ vs 32-bit masks in each half. */
+ tgen64_ori(s, dest, val & 0x00000000ffffffffull);
+ tgen64_ori(s, dest, val & 0xffffffff00000000ull);
+}
+
static void tgen32_cmp(TCGContext *s, TCGCond c, TCGReg r1, TCGReg r2)
{
if (c > TCG_COND_GT) {
@@ -1144,7 +1191,11 @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
case INDEX_op_or_i32:
- tcg_out_insn(s, RR, OR, args[0], args[2]);
+ if (const_args[2]) {
+ tgen64_ori(s, args[0], args[2] & 0xffffffff);
+ } else {
+ tcg_out_insn(s, RR, OR, args[0], args[2]);
+ }
break;
case INDEX_op_xor_i32:
tcg_out_insn(s, RR, XR, args[0], args[2]);
@@ -1158,7 +1209,11 @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode opc,
}
break;
case INDEX_op_or_i64:
- tcg_out_insn(s, RRE, OGR, args[0], args[2]);
+ if (const_args[2]) {
+ tgen64_ori(s, args[0], args[2]);
+ } else {
+ tcg_out_insn(s, RRE, OGR, args[0], args[2]);
+ }
break;
case INDEX_op_xor_i64:
tcg_out_insn(s, RRE, XGR, args[0], args[2]);
@@ -1414,7 +1469,7 @@ static const TCGTargetOpDef s390_op_defs[] = {
{ INDEX_op_divu2_i32, { "b", "a", "0", "1", "r" } },
{ INDEX_op_and_i32, { "r", "0", "ri" } },
- { INDEX_op_or_i32, { "r", "0", "r" } },
+ { INDEX_op_or_i32, { "r", "0", "ri" } },
{ INDEX_op_xor_i32, { "r", "0", "r" } },
{ INDEX_op_neg_i32, { "r", "r" } },
@@ -1475,7 +1530,7 @@ static const TCGTargetOpDef s390_op_defs[] = {
{ INDEX_op_divu2_i64, { "b", "a", "0", "1", "r" } },
{ INDEX_op_and_i64, { "r", "0", "ri" } },
- { INDEX_op_or_i64, { "r", "0", "r" } },
+ { INDEX_op_or_i64, { "r", "0", "ri" } },
{ INDEX_op_xor_i64, { "r", "0", "r" } },
{ INDEX_op_neg_i64, { "r", "r" } },
Signed-off-by: Richard Henderson <rth@twiddle.net> --- tcg/s390/tcg-target.c | 63 +++++++++++++++++++++++++++++++++++++++++++++--- 1 files changed, 59 insertions(+), 4 deletions(-)