The Arm MVE VDUP implementation would like to be able to emit code to
duplicate a byte or halfword value into an i32. We have code to do
this already in tcg-op-gvec.c, so all we need to do is make the
functions global.
For consistency with other functions made available to the frontends:
* we rename to tcg_gen_dup_*
* we expose both the _i32 and _i64 forms
* we provide the #define for a _tl form
Suggested-by: Richard Henderson <richard.hender...@linaro.org>
Reviewed-by: Richard Henderson <richard.hender...@linaro.org>
Signed-off-by: Peter Maydell <peter.mayd...@linaro.org>
Message-id: 20210617121628.20116-10-peter.mayd...@linaro.org
---
include/tcg/tcg-op.h | 8 ++++++++
include/tcg/tcg.h | 1 -
tcg/tcg-op-gvec.c | 20 ++++++++++----------
3 files changed, 18 insertions(+), 11 deletions(-)
diff --git a/include/tcg/tcg-op.h b/include/tcg/tcg-op.h
index ef8a008ea74..1a2ae937583 100644
--- a/include/tcg/tcg-op.h
+++ b/include/tcg/tcg-op.h
@@ -338,6 +338,9 @@ void tcg_gen_umin_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32
arg2);
void tcg_gen_umax_i32(TCGv_i32, TCGv_i32 arg1, TCGv_i32 arg2);
void tcg_gen_abs_i32(TCGv_i32, TCGv_i32);
+/* Replicate a value of size @vece from @in to all the lanes in @out */
+void tcg_gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in);
+
static inline void tcg_gen_discard_i32(TCGv_i32 arg)
{
tcg_gen_op1_i32(INDEX_op_discard, arg);
@@ -534,6 +537,9 @@ void tcg_gen_umin_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64
arg2);
void tcg_gen_umax_i64(TCGv_i64, TCGv_i64 arg1, TCGv_i64 arg2);
void tcg_gen_abs_i64(TCGv_i64, TCGv_i64);
+/* Replicate a value of size @vece from @in to all the lanes in @out */
+void tcg_gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in);
+
#if TCG_TARGET_REG_BITS == 64
static inline void tcg_gen_discard_i64(TCGv_i64 arg)
{
@@ -1127,6 +1133,7 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg
offset, TCGType t);
#define tcg_gen_atomic_smax_fetch_tl tcg_gen_atomic_smax_fetch_i64
#define tcg_gen_atomic_umax_fetch_tl tcg_gen_atomic_umax_fetch_i64
#define tcg_gen_dup_tl_vec tcg_gen_dup_i64_vec
+#define tcg_gen_dup_tl tcg_gen_dup_i64
#else
#define tcg_gen_movi_tl tcg_gen_movi_i32
#define tcg_gen_mov_tl tcg_gen_mov_i32
@@ -1241,6 +1248,7 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg
offset, TCGType t);
#define tcg_gen_atomic_smax_fetch_tl tcg_gen_atomic_smax_fetch_i32
#define tcg_gen_atomic_umax_fetch_tl tcg_gen_atomic_umax_fetch_i32
#define tcg_gen_dup_tl_vec tcg_gen_dup_i32_vec
+#define tcg_gen_dup_tl tcg_gen_dup_i32
#endif
#if UINTPTR_MAX == UINT32_MAX
diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h
index 064dab383bc..483e1e1f24e 100644
--- a/include/tcg/tcg.h
+++ b/include/tcg/tcg.h
@@ -1331,7 +1331,6 @@ uint64_t dup_const(unsigned vece, uint64_t c);
: (qemu_build_not_reached_always(), 0)) \
: dup_const(VECE, C))
-
/*
* Memory helpers that will be used by TCG generated code.
*/
diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c
index 498a959839f..515db120cc6 100644
--- a/tcg/tcg-op-gvec.c
+++ b/tcg/tcg-op-gvec.c
@@ -386,7 +386,7 @@ uint64_t (dup_const)(unsigned vece, uint64_t c)
}
/* Duplicate IN into OUT as per VECE. */
-static void gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in)
+void tcg_gen_dup_i32(unsigned vece, TCGv_i32 out, TCGv_i32 in)
{
switch (vece) {
case MO_8:
@@ -404,7 +404,7 @@ static void gen_dup_i32(unsigned vece, TCGv_i32 out,
TCGv_i32 in)
}
}
-static void gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in)
+void tcg_gen_dup_i64(unsigned vece, TCGv_i64 out, TCGv_i64 in)
{
switch (vece) {
case MO_8:
@@ -578,15 +578,15 @@ static void do_dup(unsigned vece, uint32_t dofs, uint32_t
oprsz,
&& (vece != MO_32 || !check_size_impl(oprsz, 4))) {
t_64 = tcg_temp_new_i64();
tcg_gen_extu_i32_i64(t_64, in_32);
- gen_dup_i64(vece, t_64, t_64);
+ tcg_gen_dup_i64(vece, t_64, t_64);
} else {
t_32 = tcg_temp_new_i32();
- gen_dup_i32(vece, t_32, in_32);
+ tcg_gen_dup_i32(vece, t_32, in_32);
}
} else if (in_64) {
/* We are given a 64-bit variable input. */
t_64 = tcg_temp_new_i64();
- gen_dup_i64(vece, t_64, in_64);
+ tcg_gen_dup_i64(vece, t_64, in_64);
} else {
/* We are given a constant input. */
/* For 64-bit hosts, use 64-bit constants for "simple" constants
@@ -1311,14 +1311,14 @@ void tcg_gen_gvec_2s(uint32_t dofs, uint32_t aofs,
uint32_t oprsz,
} else if (g->fni8 && check_size_impl(oprsz, 8)) {
TCGv_i64 t64 = tcg_temp_new_i64();
- gen_dup_i64(g->vece, t64, c);
+ tcg_gen_dup_i64(g->vece, t64, c);
expand_2s_i64(dofs, aofs, oprsz, t64, g->scalar_first, g->fni8);
tcg_temp_free_i64(t64);
} else if (g->fni4 && check_size_impl(oprsz, 4)) {
TCGv_i32 t32 = tcg_temp_new_i32();
tcg_gen_extrl_i64_i32(t32, c);
- gen_dup_i32(g->vece, t32, t32);
+ tcg_gen_dup_i32(g->vece, t32, t32);
expand_2s_i32(dofs, aofs, oprsz, t32, g->scalar_first, g->fni4);
tcg_temp_free_i32(t32);
} else {
@@ -2538,7 +2538,7 @@ void tcg_gen_gvec_ands(unsigned vece, uint32_t dofs,
uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_ands);
tcg_temp_free_i64(tmp);
}
@@ -2562,7 +2562,7 @@ void tcg_gen_gvec_xors(unsigned vece, uint32_t dofs,
uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_xors);
tcg_temp_free_i64(tmp);
}
@@ -2586,7 +2586,7 @@ void tcg_gen_gvec_ors(unsigned vece, uint32_t dofs,
uint32_t aofs,
TCGv_i64 c, uint32_t oprsz, uint32_t maxsz)
{
TCGv_i64 tmp = tcg_temp_new_i64();
- gen_dup_i64(vece, tmp, c);
+ tcg_gen_dup_i64(vece, tmp, c);
tcg_gen_gvec_2s(dofs, aofs, oprsz, maxsz, tmp, &gop_ors);
tcg_temp_free_i64(tmp);
}
--
2.20.1
