There is an older form that produces per-byte results,
and a newer form that produces per-register results,
and a vector form that produces per-element results.
Signed-off-by: Richard Henderson <richard.hender...@linaro.org>
---
tcg/s390x/tcg-target.h | 5 ++--
tcg/s390x/tcg-target.c.inc | 51 ++++++++++++++++++++++++++++++++++++++
2 files changed, 54 insertions(+), 2 deletions(-)
diff --git a/tcg/s390x/tcg-target.h b/tcg/s390x/tcg-target.h
index 191c6a073e..5d184d8e14 100644
--- a/tcg/s390x/tcg-target.h
+++ b/tcg/s390x/tcg-target.h
@@ -62,6 +62,7 @@ typedef enum TCGReg {
#define FACILITY_LOAD_ON_COND 45
#define FACILITY_FAST_BCR_SER FACILITY_LOAD_ON_COND
#define FACILITY_DISTINCT_OPS FACILITY_LOAD_ON_COND
+#define FACILITY_POPCOUNT FACILITY_LOAD_ON_COND
#define FACILITY_LOAD_ON_COND2 53
#define FACILITY_MISC_INSN_EXT2 58
#define FACILITY_MISC_INSN_EXT3 61
@@ -91,7 +92,7 @@ extern uint64_t s390_facilities[3];
#define TCG_TARGET_HAS_nor_i32 HAVE_FACILITY(MISC_INSN_EXT3)
#define TCG_TARGET_HAS_clz_i32 0
#define TCG_TARGET_HAS_ctz_i32 0
-#define TCG_TARGET_HAS_ctpop_i32 0
+#define TCG_TARGET_HAS_ctpop_i32 HAVE_FACILITY(POPCOUNT)
#define TCG_TARGET_HAS_deposit_i32 HAVE_FACILITY(GEN_INST_EXT)
#define TCG_TARGET_HAS_extract_i32 HAVE_FACILITY(GEN_INST_EXT)
#define TCG_TARGET_HAS_sextract_i32 0
@@ -128,7 +129,7 @@ extern uint64_t s390_facilities[3];
#define TCG_TARGET_HAS_nor_i64 HAVE_FACILITY(MISC_INSN_EXT3)
#define TCG_TARGET_HAS_clz_i64 HAVE_FACILITY(EXT_IMM)
#define TCG_TARGET_HAS_ctz_i64 0
-#define TCG_TARGET_HAS_ctpop_i64 0
+#define TCG_TARGET_HAS_ctpop_i64 HAVE_FACILITY(POPCOUNT)
#define TCG_TARGET_HAS_deposit_i64 HAVE_FACILITY(GEN_INST_EXT)
#define TCG_TARGET_HAS_extract_i64 HAVE_FACILITY(GEN_INST_EXT)
#define TCG_TARGET_HAS_sextract_i64 0
diff --git a/tcg/s390x/tcg-target.c.inc b/tcg/s390x/tcg-target.c.inc
index 23cbb10168..7744c6ad54 100644
--- a/tcg/s390x/tcg-target.c.inc
+++ b/tcg/s390x/tcg-target.c.inc
@@ -208,6 +208,7 @@ typedef enum S390Opcode {
RRFc_LOCR = 0xb9f2,
RRFc_LOCGR = 0xb9e2,
+ RRFc_POPCNT = 0xb9e1,
RR_AR = 0x1a,
RR_ALR = 0x1e,
@@ -259,6 +260,7 @@ typedef enum S390Opcode {
RXY_LRVG = 0xe30f,
RXY_LRVH = 0xe31f,
RXY_LY = 0xe358,
+ RXY_MSG = 0xe30c,
RXY_NG = 0xe380,
RXY_OG = 0xe381,
RXY_STCY = 0xe372,
@@ -276,6 +278,7 @@ typedef enum S390Opcode {
RX_L = 0x58,
RX_LA = 0x41,
RX_LH = 0x48,
+ RX_MS = 0x71,
RX_ST = 0x50,
RX_STC = 0x42,
RX_STH = 0x40,
@@ -1568,6 +1571,45 @@ static void tgen_clz(TCGContext *s, TCGReg dest, TCGReg
a1,
tgen_movcond_int(s, TCG_TYPE_I64, dest, a2, a2const, TCG_REG_R0, 8, 2);
}
+static void tgen_ctpop(TCGContext *s, TCGType type, TCGReg dest, TCGReg src)
+{
+ /* With MIE3, and bit 0 of m4 set, we get the complete result. */
+ if (HAVE_FACILITY(MISC_INSN_EXT3)) {
+ if (type == TCG_TYPE_I32) {
+ tgen_ext32u(s, dest, src);
+ src = dest;
+ }
+ tcg_out_insn(s, RRFc, POPCNT, dest, src, 8);
+ return;
+ }
+
+ /* Without MIE3, each byte gets the count of bits for the byte. */
+ tcg_out_insn(s, RRFc, POPCNT, dest, src, 0);
+
+ /* Multiply to sum each byte at the top of the word. */
+ if (type == TCG_TYPE_I32 && HAVE_FACILITY(GEN_INST_EXT)) {
+ tcg_out_insn(s, RIL, MSFI, dest, 0x01010101);
+ } else {
+ /* No space to save: share the constant between TCG_TYPE_I32/I64. */
+ tcg_out_insn(s, RIL, LARL, TCG_TMP0, 0);
+ new_pool_label(s, 0x0101010101010101ull,
+ R_390_PC32DBL, s->code_ptr - 2, 2);
+
+ if (type == TCG_TYPE_I32) {
+ tcg_out_insn(s, RX, MS, dest, TCG_TMP0, TCG_REG_NONE, 0);
+ } else {
+ tcg_out_insn(s, RXY, MSG, dest, TCG_TMP0, TCG_REG_NONE, 0);
+ }
+ }
+
+ /* Shift result down from the top byte. */
+ if (type == TCG_TYPE_I32) {
+ tcg_out_sh32(s, RS_SRL, dest, TCG_REG_NONE, 24);
+ } else {
+ tcg_out_sh64(s, RSY_SRLG, dest, dest, TCG_REG_NONE, 56);
+ }
+}
+
static void tgen_deposit(TCGContext *s, TCGReg dest, TCGReg src,
int ofs, int len, int z)
{
@@ -2733,6 +2775,13 @@ static inline void tcg_out_op(TCGContext *s, TCGOpcode
opc,
tgen_clz(s, args[0], args[1], args[2], const_args[2]);
break;
+ case INDEX_op_ctpop_i32:
+ tgen_ctpop(s, TCG_TYPE_I32, args[0], args[1]);
+ break;
+ case INDEX_op_ctpop_i64:
+ tgen_ctpop(s, TCG_TYPE_I64, args[0], args[1]);
+ break;
+
case INDEX_op_mb:
/* The host memory model is quite strong, we simply need to
serialize the instruction stream. */
@@ -3302,6 +3351,8 @@ static TCGConstraintSetIndex tcg_target_op_def(TCGOpcode
op)
case INDEX_op_extu_i32_i64:
case INDEX_op_extract_i32:
case INDEX_op_extract_i64:
+ case INDEX_op_ctpop_i32:
+ case INDEX_op_ctpop_i64:
return C_O1_I1(r, r);
case INDEX_op_qemu_ld_i32:
--
2.34.1
