tcg: Reduce copies for plugin_gen_mem_callbacks

We only need to make copies for loads, when the destination
overlaps the address.  For now, only eliminate the copy for
stores and 128-bit loads.

Rename plugin_prep_mem_callbacks to plugin_maybe_preserve_addr,
returning NULL if no copy is made.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/tcg-op-ldst.c

@@ -114,7 +114,8 @@ static void tcg_gen_req_mo(TCGBar type)
     }
 }
 
-static inline TCGv plugin_prep_mem_callbacks(TCGv vaddr)
+/* Only required for loads, where value might overlap addr. */
+static TCGv plugin_maybe_preserve_addr(TCGv vaddr)
 {
 #ifdef CONFIG_PLUGIN
     if (tcg_ctx->plugin_insn != NULL) {
@@ -124,17 +125,20 @@ static inline TCGv plugin_prep_mem_callbacks(TCGv vaddr)
         return temp;
     }
 #endif
-    return vaddr;
+    return NULL;
 }
 
-static void plugin_gen_mem_callbacks(TCGv vaddr, MemOpIdx oi,
-                                     enum qemu_plugin_mem_rw rw)
+static void
+plugin_gen_mem_callbacks(TCGv copy_addr, TCGv orig_addr, MemOpIdx oi,
+                         enum qemu_plugin_mem_rw rw)
 {
 #ifdef CONFIG_PLUGIN
     if (tcg_ctx->plugin_insn != NULL) {
         qemu_plugin_meminfo_t info = make_plugin_meminfo(oi, rw);
-        plugin_gen_empty_mem_callback(vaddr, info);
-        tcg_temp_free(vaddr);
+        plugin_gen_empty_mem_callback(copy_addr ? : orig_addr, info);
+        if (copy_addr) {
+            tcg_temp_free(copy_addr);
+        }
     }
 #endif
 }
@@ -143,6 +147,7 @@ void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop)
 {
     MemOp orig_memop;
     MemOpIdx oi;
+    TCGv copy_addr;
 
     tcg_gen_req_mo(TCG_MO_LD_LD | TCG_MO_ST_LD);
     memop = tcg_canonicalize_memop(memop, 0, 0);
@@ -157,9 +162,9 @@ void tcg_gen_qemu_ld_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop)
         }
     }
 
-    addr = plugin_prep_mem_callbacks(addr);
+    copy_addr = plugin_maybe_preserve_addr(addr);
     gen_ldst_i32(INDEX_op_qemu_ld_i32, val, addr, memop, idx);
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_R);
+    plugin_gen_mem_callbacks(copy_addr, addr, oi, QEMU_PLUGIN_MEM_R);
 
     if ((orig_memop ^ memop) & MO_BSWAP) {
         switch (orig_memop & MO_SIZE) {
@@ -202,13 +207,12 @@ void tcg_gen_qemu_st_i32(TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop)
         memop &= ~MO_BSWAP;
     }
 
-    addr = plugin_prep_mem_callbacks(addr);
     if (TCG_TARGET_HAS_qemu_st8_i32 && (memop & MO_SIZE) == MO_8) {
         gen_ldst_i32(INDEX_op_qemu_st8_i32, val, addr, memop, idx);
     } else {
         gen_ldst_i32(INDEX_op_qemu_st_i32, val, addr, memop, idx);
     }
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_W);
+    plugin_gen_mem_callbacks(NULL, addr, oi, QEMU_PLUGIN_MEM_W);
 
     if (swap) {
         tcg_temp_free_i32(swap);
@@ -219,6 +223,7 @@ void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop)
 {
     MemOp orig_memop;
     MemOpIdx oi;
+    TCGv copy_addr;
 
     if (TCG_TARGET_REG_BITS == 32 && (memop & MO_SIZE) < MO_64) {
         tcg_gen_qemu_ld_i32(TCGV_LOW(val), addr, idx, memop);
@@ -243,9 +248,9 @@ void tcg_gen_qemu_ld_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop)
         }
     }
 
-    addr = plugin_prep_mem_callbacks(addr);
+    copy_addr = plugin_maybe_preserve_addr(addr);
     gen_ldst_i64(INDEX_op_qemu_ld_i64, val, addr, memop, idx);
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_R);
+    plugin_gen_mem_callbacks(copy_addr, addr, oi, QEMU_PLUGIN_MEM_R);
 
     if ((orig_memop ^ memop) & MO_BSWAP) {
         int flags = (orig_memop & MO_SIGN
@@ -300,9 +305,8 @@ void tcg_gen_qemu_st_i64(TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop)
         memop &= ~MO_BSWAP;
     }
 
-    addr = plugin_prep_mem_callbacks(addr);
     gen_ldst_i64(INDEX_op_qemu_st_i64, val, addr, memop, idx);
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_W);
+    plugin_gen_mem_callbacks(NULL, addr, oi, QEMU_PLUGIN_MEM_W);
 
     if (swap) {
         tcg_temp_free_i64(swap);
@@ -419,7 +423,6 @@ void tcg_gen_qemu_ld_i128(TCGv_i128 val, TCGv addr, TCGArg idx, MemOp memop)
     tcg_debug_assert((memop & MO_SIGN) == 0);
 
     tcg_gen_req_mo(TCG_MO_LD_LD | TCG_MO_ST_LD);
-    addr = plugin_prep_mem_callbacks(addr);
 
     /* TODO: For now, force 32-bit hosts to use the helper. */
     if (TCG_TARGET_HAS_qemu_ldst_i128 && TCG_TARGET_REG_BITS == 64) {
@@ -490,7 +493,7 @@ void tcg_gen_qemu_ld_i128(TCGv_i128 val, TCGv addr, TCGArg idx, MemOp memop)
         maybe_free_addr64(a64);
     }
 
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_R);
+    plugin_gen_mem_callbacks(NULL, addr, oi, QEMU_PLUGIN_MEM_R);
 }
 
 void tcg_gen_qemu_st_i128(TCGv_i128 val, TCGv addr, TCGArg idx, MemOp memop)
@@ -501,7 +504,6 @@ void tcg_gen_qemu_st_i128(TCGv_i128 val, TCGv addr, TCGArg idx, MemOp memop)
     tcg_debug_assert((memop & MO_SIGN) == 0);
 
     tcg_gen_req_mo(TCG_MO_ST_LD | TCG_MO_ST_ST);
-    addr = plugin_prep_mem_callbacks(addr);
 
     /* TODO: For now, force 32-bit hosts to use the helper. */
 
@@ -572,7 +574,7 @@ void tcg_gen_qemu_st_i128(TCGv_i128 val, TCGv addr, TCGArg idx, MemOp memop)
         maybe_free_addr64(a64);
     }
 
-    plugin_gen_mem_callbacks(addr, oi, QEMU_PLUGIN_MEM_W);
+    plugin_gen_mem_callbacks(NULL, addr, oi, QEMU_PLUGIN_MEM_W);
 }
 
 static void tcg_gen_ext_i32(TCGv_i32 ret, TCGv_i32 val, MemOp opc)