[Qemu-devel] [PATCH 22/25] tcg: Add helpers for clrsb

[Richard Henderson <rth@twiddle.net>]

The number of actual invocations does not warrent an opcode,
and the backends generating it.  But at least we can eliminate
redundant helpers.

Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 tcg-runtime.c     | 10 ++++++++++
 tcg/tcg-op.c      | 28 ++++++++++++++++++++++++++++
 tcg/tcg-op.h      |  4 ++++
 tcg/tcg-runtime.h |  2 ++
 4 files changed, 44 insertions(+)

diff --git a/tcg-runtime.c b/tcg-runtime.c
index eb3bade..c8b98df 100644
--- a/tcg-runtime.c
+++ b/tcg-runtime.c
@@ -121,6 +121,16 @@ uint64_t HELPER(ctz_i64)(uint64_t arg, uint64_t zero_val)
     return arg ? ctz64(arg) : zero_val;
 }
 
+uint32_t HELPER(clrsb_i32)(uint32_t arg)
+{
+    return clrsb32(arg);
+}
+
+uint64_t HELPER(clrsb_i64)(uint64_t arg)
+{
+    return clrsb64(arg);
+}
+
 void HELPER(exit_atomic)(CPUArchState *env)
 {
     cpu_loop_exit_atomic(ENV_GET_CPU(env), GETPC());
diff --git a/tcg/tcg-op.c b/tcg/tcg-op.c
index b45095c..728c4b3 100644
--- a/tcg/tcg-op.c
+++ b/tcg/tcg-op.c
@@ -489,6 +489,20 @@ void tcg_gen_ctzi_i32(TCGv_i32 ret, TCGv_i32 arg1, uint32_t arg2)
     tcg_temp_free_i32(t);
 }
 
+void tcg_gen_clrsb_i32(TCGv_i32 ret, TCGv_i32 arg)
+{
+    if (TCG_TARGET_HAS_clz_i32) {
+        TCGv_i32 t = tcg_temp_new_i32();
+        tcg_gen_sari_i32(t, arg, 31);
+        tcg_gen_xor_i32(t, t, arg);
+        tcg_gen_clzi_i32(t, t, 32);
+        tcg_gen_subi_i32(ret, t, 1);
+        tcg_temp_free_i32(t);
+    } else {
+        gen_helper_clrsb_i32(ret, arg);
+    }
+}
+
 void tcg_gen_rotl_i32(TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2)
 {
     if (TCG_TARGET_HAS_rot_i32) {
@@ -1789,6 +1803,20 @@ void tcg_gen_ctzi_i64(TCGv_i64 ret, TCGv_i64 arg1, uint64_t arg2)
     }
 }
 
+void tcg_gen_clrsb_i64(TCGv_i64 ret, TCGv_i64 arg)
+{
+    if (TCG_TARGET_HAS_clz_i64 || TCG_TARGET_HAS_clz_i32) {
+        TCGv_i64 t = tcg_temp_new_i64();
+        tcg_gen_sari_i64(t, arg, 63);
+        tcg_gen_xor_i64(t, t, arg);
+        tcg_gen_clzi_i64(t, t, 64);
+        tcg_gen_subi_i64(ret, t, 1);
+        tcg_temp_free_i64(t);
+    } else {
+        gen_helper_clrsb_i64(ret, arg);
+    }
+}
+
 void tcg_gen_rotl_i64(TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2)
 {
     if (TCG_TARGET_HAS_rot_i64) {
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index 7a24e84..c2f3db9 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -290,6 +290,7 @@ void tcg_gen_clz_i32(TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2);
 void tcg_gen_ctz_i32(TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2);
 void tcg_gen_clzi_i32(TCGv_i32 ret, TCGv_i32 arg1, uint32_t arg2);
 void tcg_gen_ctzi_i32(TCGv_i32 ret, TCGv_i32 arg1, uint32_t arg2);
+void tcg_gen_clrsb_i32(TCGv_i32 ret, TCGv_i32 arg);
 void tcg_gen_rotl_i32(TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2);
 void tcg_gen_rotli_i32(TCGv_i32 ret, TCGv_i32 arg1, unsigned arg2);
 void tcg_gen_rotr_i32(TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2);
@@ -477,6 +478,7 @@ void tcg_gen_clz_i64(TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2);
 void tcg_gen_ctz_i64(TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2);
 void tcg_gen_clzi_i64(TCGv_i64 ret, TCGv_i64 arg1, uint64_t arg2);
 void tcg_gen_ctzi_i64(TCGv_i64 ret, TCGv_i64 arg1, uint64_t arg2);
+void tcg_gen_clrsb_i64(TCGv_i64 ret, TCGv_i64 arg);
 void tcg_gen_rotl_i64(TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2);
 void tcg_gen_rotli_i64(TCGv_i64 ret, TCGv_i64 arg1, unsigned arg2);
 void tcg_gen_rotr_i64(TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2);
@@ -970,6 +972,7 @@ void tcg_gen_atomic_xor_fetch_i64(TCGv_i64, TCGv, TCGv_i64, TCGArg, TCGMemOp);
 #define tcg_gen_ctz_tl tcg_gen_ctz_i64
 #define tcg_gen_clzi_tl tcg_gen_clzi_i64
 #define tcg_gen_ctzi_tl tcg_gen_ctzi_i64
+#define tcg_gen_clrsb_tl tcg_gen_clrsb_i64
 #define tcg_gen_rotl_tl tcg_gen_rotl_i64
 #define tcg_gen_rotli_tl tcg_gen_rotli_i64
 #define tcg_gen_rotr_tl tcg_gen_rotr_i64
@@ -1065,6 +1068,7 @@ void tcg_gen_atomic_xor_fetch_i64(TCGv_i64, TCGv, TCGv_i64, TCGArg, TCGMemOp);
 #define tcg_gen_ctz_tl tcg_gen_ctz_i32
 #define tcg_gen_clzi_tl tcg_gen_clzi_i32
 #define tcg_gen_ctzi_tl tcg_gen_ctzi_i32
+#define tcg_gen_clrsb_tl tcg_gen_clrsb_i32
 #define tcg_gen_rotl_tl tcg_gen_rotl_i32
 #define tcg_gen_rotli_tl tcg_gen_rotli_i32
 #define tcg_gen_rotr_tl tcg_gen_rotr_i32
diff --git a/tcg/tcg-runtime.h b/tcg/tcg-runtime.h
index eb1cd76..0d30f1a 100644
--- a/tcg/tcg-runtime.h
+++ b/tcg/tcg-runtime.h
@@ -19,6 +19,8 @@ DEF_HELPER_FLAGS_2(clz_i32, TCG_CALL_NO_RWG_SE, i32, i32, i32)
 DEF_HELPER_FLAGS_2(ctz_i32, TCG_CALL_NO_RWG_SE, i32, i32, i32)
 DEF_HELPER_FLAGS_2(clz_i64, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(ctz_i64, TCG_CALL_NO_RWG_SE, i64, i64, i64)
+DEF_HELPER_FLAGS_1(clrsb_i32, TCG_CALL_NO_RWG_SE, i32, i32)
+DEF_HELPER_FLAGS_1(clrsb_i64, TCG_CALL_NO_RWG_SE, i64, i64)
 
 DEF_HELPER_FLAGS_1(exit_atomic, TCG_CALL_NO_WG, noreturn, env)
 
-- 
2.7.4