Patchwork [1/7] tcg: Define "deposit" as an optional operation.

login
register
mail settings
Submitter Richard Henderson
Date Jan. 11, 2011, 3:23 a.m.
Message ID <1294716228-9299-2-git-send-email-rth@twiddle.net>
Download mbox | patch
Permalink /patch/78272/
State New
Headers show

Comments

Richard Henderson - Jan. 11, 2011, 3:23 a.m.
Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 tcg/README    |   14 ++++++++++++
 tcg/tcg-op.h  |   64 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 tcg/tcg-opc.h |    6 +++++
 3 files changed, 84 insertions(+), 0 deletions(-)
Aurelien Jarno - Jan. 11, 2011, 11:45 p.m.
On Mon, Jan 10, 2011 at 07:23:42PM -0800, Richard Henderson wrote:
> Signed-off-by: Richard Henderson <rth@twiddle.net>
> ---
>  tcg/README    |   14 ++++++++++++
>  tcg/tcg-op.h  |   64 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tcg/tcg-opc.h |    6 +++++
>  3 files changed, 84 insertions(+), 0 deletions(-)

Acked-by: Aurelien Jarno <aurelien@aurel32.net>

> diff --git a/tcg/README b/tcg/README
> index a18a87f..a50ecc6 100644
> --- a/tcg/README
> +++ b/tcg/README
> @@ -285,6 +285,20 @@ the four high order bytes are set to zero.
>  Indicate that the value of t0 won't be used later. It is useful to
>  force dead code elimination.
>  
> +* deposit_i32/i64 dest, t1, t2, pos, loc
> +
> +Deposit T2 as a bitfield into T1, placing the result in DEST.
> +The bitfield is described by POS/LOC, which are immediate values:
> +
> +  LEN - the length of the bitfield
> +  POS - the position of the first bit, counting from the LSB
> +
> +For example, pos=8, len=4 indicates a 4-bit field at bit 8.
> +This operation would be equivalent to
> +
> +  dest = (t1 & ~0x0f00) | ((t2 << 8) & 0x0f00)
> +
> +
>  ********* Conditional moves
>  
>  * setcond_i32/i64 cond, dest, t1, t2
> diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
> index 3ee0a58..207a89f 100644
> --- a/tcg/tcg-op.h
> +++ b/tcg/tcg-op.h
> @@ -254,6 +254,30 @@ static inline void tcg_gen_op5i_i64(TCGOpcode opc, TCGv_i64 arg1, TCGv_i64 arg2,
>      *gen_opparam_ptr++ = arg5;
>  }
>  
> +static inline void tcg_gen_op5ii_i32(TCGOpcode opc, TCGv_i32 arg1,
> +                                     TCGv_i32 arg2, TCGv_i32 arg3,
> +                                     TCGArg arg4, TCGArg arg5)
> +{
> +    *gen_opc_ptr++ = opc;
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg1);
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg2);
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg3);
> +    *gen_opparam_ptr++ = arg4;
> +    *gen_opparam_ptr++ = arg5;
> +}
> +
> +static inline void tcg_gen_op5ii_i64(TCGOpcode opc, TCGv_i64 arg1,
> +                                     TCGv_i64 arg2, TCGv_i64 arg3,
> +                                     TCGArg arg4, TCGArg arg5)
> +{
> +    *gen_opc_ptr++ = opc;
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg1);
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg2);
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg3);
> +    *gen_opparam_ptr++ = arg4;
> +    *gen_opparam_ptr++ = arg5;
> +}
> +
>  static inline void tcg_gen_op6_i32(TCGOpcode opc, TCGv_i32 arg1, TCGv_i32 arg2,
>                                     TCGv_i32 arg3, TCGv_i32 arg4, TCGv_i32 arg5,
>                                     TCGv_i32 arg6)
> @@ -2071,6 +2095,44 @@ static inline void tcg_gen_rotri_i64(TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2)
>      }
>  }
>  
> +static inline void tcg_gen_deposit_i32(TCGv_i32 ret, TCGv_i32 arg1,
> +				       TCGv_i32 arg2, unsigned int ofs,
> +				       unsigned int len)
> +{
> +#ifdef TCG_TARGET_HAS_deposit_i32
> +  tcg_gen_op5ii_i32(INDEX_op_deposit_i32, ret, arg1, arg2, ofs, len);
> +#else
> +  uint32_t mask = (1u << len) - 1;
> +  TCGv_i32 t1 = tcg_temp_new_i32 ();
> +
> +  tcg_gen_andi_i32(t1, arg2, mask);
> +  tcg_gen_shli_i32(t1, t1, ofs);
> +  tcg_gen_andi_i32(ret, arg1, ~(mask << ofs));
> +  tcg_gen_or_i32(ret, ret, t1);
> +
> +  tcg_temp_free_i32(t1);
> +#endif
> +}
> +
> +static inline void tcg_gen_deposit_i64(TCGv_i64 ret, TCGv_i64 arg1,
> +				       TCGv_i64 arg2, unsigned int ofs,
> +				       unsigned int len)
> +{
> +#ifdef TCG_TARGET_HAS_deposit_i64
> +  tcg_gen_op5ii_i64(INDEX_op_deposit_i64, ret, arg1, arg2, ofs, len);
> +#else
> +  uint64_t mask = (1ull << len) - 1;
> +  TCGv_i64 t1 = tcg_temp_new_i64 ();
> +
> +  tcg_gen_andi_i64(t1, arg2, mask);
> +  tcg_gen_shli_i64(t1, t1, ofs);
> +  tcg_gen_andi_i64(ret, arg1, ~(mask << ofs));
> +  tcg_gen_or_i64(ret, ret, t1);
> +
> +  tcg_temp_free_i64(t1);
> +#endif
> +}
> +
>  /***************************************/
>  /* QEMU specific operations. Their type depend on the QEMU CPU
>     type. */
> @@ -2384,6 +2446,7 @@ static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
>  #define tcg_gen_rotli_tl tcg_gen_rotli_i64
>  #define tcg_gen_rotr_tl tcg_gen_rotr_i64
>  #define tcg_gen_rotri_tl tcg_gen_rotri_i64
> +#define tcg_gen_deposit_tl tcg_gen_deposit_i64
>  #define tcg_const_tl tcg_const_i64
>  #define tcg_const_local_tl tcg_const_local_i64
>  #else
> @@ -2454,6 +2517,7 @@ static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
>  #define tcg_gen_rotli_tl tcg_gen_rotli_i32
>  #define tcg_gen_rotr_tl tcg_gen_rotr_i32
>  #define tcg_gen_rotri_tl tcg_gen_rotri_i32
> +#define tcg_gen_deposit_tl tcg_gen_deposit_i32
>  #define tcg_const_tl tcg_const_i32
>  #define tcg_const_local_tl tcg_const_local_i32
>  #endif
> diff --git a/tcg/tcg-opc.h b/tcg/tcg-opc.h
> index 2a98fed..2c7ca1a 100644
> --- a/tcg/tcg-opc.h
> +++ b/tcg/tcg-opc.h
> @@ -78,6 +78,9 @@ DEF(sar_i32, 1, 2, 0, 0)
>  DEF(rotl_i32, 1, 2, 0, 0)
>  DEF(rotr_i32, 1, 2, 0, 0)
>  #endif
> +#ifdef TCG_TARGET_HAS_deposit_i32
> +DEF(deposit_i32, 1, 2, 2, 0)
> +#endif
>  
>  DEF(brcond_i32, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
>  #if TCG_TARGET_REG_BITS == 32
> @@ -168,6 +171,9 @@ DEF(sar_i64, 1, 2, 0, 0)
>  DEF(rotl_i64, 1, 2, 0, 0)
>  DEF(rotr_i64, 1, 2, 0, 0)
>  #endif
> +#ifdef TCG_TARGET_HAS_deposit_i64
> +DEF(deposit_i64, 1, 2, 2, 0)
> +#endif
>  
>  DEF(brcond_i64, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
>  #ifdef TCG_TARGET_HAS_ext8s_i64
> -- 
> 1.7.2.3
> 
> 
>
Edgar Iglesias - Jan. 12, 2011, 11 a.m.
On Mon, Jan 10, 2011 at 07:23:42PM -0800, Richard Henderson wrote:
> Signed-off-by: Richard Henderson <rth@twiddle.net>
> ---
>  tcg/README    |   14 ++++++++++++
>  tcg/tcg-op.h  |   64 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tcg/tcg-opc.h |    6 +++++
>  3 files changed, 84 insertions(+), 0 deletions(-)
> 
> diff --git a/tcg/README b/tcg/README
> index a18a87f..a50ecc6 100644
> --- a/tcg/README
> +++ b/tcg/README
> @@ -285,6 +285,20 @@ the four high order bytes are set to zero.
>  Indicate that the value of t0 won't be used later. It is useful to
>  force dead code elimination.
>  
> +* deposit_i32/i64 dest, t1, t2, pos, loc
                                        ^^^
                                  Should maybe be len.

> +
> +Deposit T2 as a bitfield into T1, placing the result in DEST.
> +The bitfield is described by POS/LOC, which are immediate values:
                                    ^^^
                                    LEN


The rest looks good, thanks.

Acked-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>


> +
> +  LEN - the length of the bitfield
> +  POS - the position of the first bit, counting from the LSB
> +
> +For example, pos=8, len=4 indicates a 4-bit field at bit 8.
> +This operation would be equivalent to
> +
> +  dest = (t1 & ~0x0f00) | ((t2 << 8) & 0x0f00)
> +
> +
>  ********* Conditional moves
>  
>  * setcond_i32/i64 cond, dest, t1, t2
> diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
> index 3ee0a58..207a89f 100644
> --- a/tcg/tcg-op.h
> +++ b/tcg/tcg-op.h
> @@ -254,6 +254,30 @@ static inline void tcg_gen_op5i_i64(TCGOpcode opc, TCGv_i64 arg1, TCGv_i64 arg2,
>      *gen_opparam_ptr++ = arg5;
>  }
>  
> +static inline void tcg_gen_op5ii_i32(TCGOpcode opc, TCGv_i32 arg1,
> +                                     TCGv_i32 arg2, TCGv_i32 arg3,
> +                                     TCGArg arg4, TCGArg arg5)
> +{
> +    *gen_opc_ptr++ = opc;
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg1);
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg2);
> +    *gen_opparam_ptr++ = GET_TCGV_I32(arg3);
> +    *gen_opparam_ptr++ = arg4;
> +    *gen_opparam_ptr++ = arg5;
> +}
> +
> +static inline void tcg_gen_op5ii_i64(TCGOpcode opc, TCGv_i64 arg1,
> +                                     TCGv_i64 arg2, TCGv_i64 arg3,
> +                                     TCGArg arg4, TCGArg arg5)
> +{
> +    *gen_opc_ptr++ = opc;
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg1);
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg2);
> +    *gen_opparam_ptr++ = GET_TCGV_I64(arg3);
> +    *gen_opparam_ptr++ = arg4;
> +    *gen_opparam_ptr++ = arg5;
> +}
> +
>  static inline void tcg_gen_op6_i32(TCGOpcode opc, TCGv_i32 arg1, TCGv_i32 arg2,
>                                     TCGv_i32 arg3, TCGv_i32 arg4, TCGv_i32 arg5,
>                                     TCGv_i32 arg6)
> @@ -2071,6 +2095,44 @@ static inline void tcg_gen_rotri_i64(TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2)
>      }
>  }
>  
> +static inline void tcg_gen_deposit_i32(TCGv_i32 ret, TCGv_i32 arg1,
> +				       TCGv_i32 arg2, unsigned int ofs,
> +				       unsigned int len)
> +{
> +#ifdef TCG_TARGET_HAS_deposit_i32
> +  tcg_gen_op5ii_i32(INDEX_op_deposit_i32, ret, arg1, arg2, ofs, len);
> +#else
> +  uint32_t mask = (1u << len) - 1;
> +  TCGv_i32 t1 = tcg_temp_new_i32 ();
> +
> +  tcg_gen_andi_i32(t1, arg2, mask);
> +  tcg_gen_shli_i32(t1, t1, ofs);
> +  tcg_gen_andi_i32(ret, arg1, ~(mask << ofs));
> +  tcg_gen_or_i32(ret, ret, t1);
> +
> +  tcg_temp_free_i32(t1);
> +#endif
> +}
> +
> +static inline void tcg_gen_deposit_i64(TCGv_i64 ret, TCGv_i64 arg1,
> +				       TCGv_i64 arg2, unsigned int ofs,
> +				       unsigned int len)
> +{
> +#ifdef TCG_TARGET_HAS_deposit_i64
> +  tcg_gen_op5ii_i64(INDEX_op_deposit_i64, ret, arg1, arg2, ofs, len);
> +#else
> +  uint64_t mask = (1ull << len) - 1;
> +  TCGv_i64 t1 = tcg_temp_new_i64 ();
> +
> +  tcg_gen_andi_i64(t1, arg2, mask);
> +  tcg_gen_shli_i64(t1, t1, ofs);
> +  tcg_gen_andi_i64(ret, arg1, ~(mask << ofs));
> +  tcg_gen_or_i64(ret, ret, t1);
> +
> +  tcg_temp_free_i64(t1);
> +#endif
> +}
> +
>  /***************************************/
>  /* QEMU specific operations. Their type depend on the QEMU CPU
>     type. */
> @@ -2384,6 +2446,7 @@ static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
>  #define tcg_gen_rotli_tl tcg_gen_rotli_i64
>  #define tcg_gen_rotr_tl tcg_gen_rotr_i64
>  #define tcg_gen_rotri_tl tcg_gen_rotri_i64
> +#define tcg_gen_deposit_tl tcg_gen_deposit_i64
>  #define tcg_const_tl tcg_const_i64
>  #define tcg_const_local_tl tcg_const_local_i64
>  #else
> @@ -2454,6 +2517,7 @@ static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
>  #define tcg_gen_rotli_tl tcg_gen_rotli_i32
>  #define tcg_gen_rotr_tl tcg_gen_rotr_i32
>  #define tcg_gen_rotri_tl tcg_gen_rotri_i32
> +#define tcg_gen_deposit_tl tcg_gen_deposit_i32
>  #define tcg_const_tl tcg_const_i32
>  #define tcg_const_local_tl tcg_const_local_i32
>  #endif
> diff --git a/tcg/tcg-opc.h b/tcg/tcg-opc.h
> index 2a98fed..2c7ca1a 100644
> --- a/tcg/tcg-opc.h
> +++ b/tcg/tcg-opc.h
> @@ -78,6 +78,9 @@ DEF(sar_i32, 1, 2, 0, 0)
>  DEF(rotl_i32, 1, 2, 0, 0)
>  DEF(rotr_i32, 1, 2, 0, 0)
>  #endif
> +#ifdef TCG_TARGET_HAS_deposit_i32
> +DEF(deposit_i32, 1, 2, 2, 0)
> +#endif
>  
>  DEF(brcond_i32, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
>  #if TCG_TARGET_REG_BITS == 32
> @@ -168,6 +171,9 @@ DEF(sar_i64, 1, 2, 0, 0)
>  DEF(rotl_i64, 1, 2, 0, 0)
>  DEF(rotr_i64, 1, 2, 0, 0)
>  #endif
> +#ifdef TCG_TARGET_HAS_deposit_i64
> +DEF(deposit_i64, 1, 2, 2, 0)
> +#endif
>  
>  DEF(brcond_i64, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
>  #ifdef TCG_TARGET_HAS_ext8s_i64
> -- 
> 1.7.2.3
> 
>

Patch

diff --git a/tcg/README b/tcg/README
index a18a87f..a50ecc6 100644
--- a/tcg/README
+++ b/tcg/README
@@ -285,6 +285,20 @@  the four high order bytes are set to zero.
 Indicate that the value of t0 won't be used later. It is useful to
 force dead code elimination.
 
+* deposit_i32/i64 dest, t1, t2, pos, loc
+
+Deposit T2 as a bitfield into T1, placing the result in DEST.
+The bitfield is described by POS/LOC, which are immediate values:
+
+  LEN - the length of the bitfield
+  POS - the position of the first bit, counting from the LSB
+
+For example, pos=8, len=4 indicates a 4-bit field at bit 8.
+This operation would be equivalent to
+
+  dest = (t1 & ~0x0f00) | ((t2 << 8) & 0x0f00)
+
+
 ********* Conditional moves
 
 * setcond_i32/i64 cond, dest, t1, t2
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index 3ee0a58..207a89f 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -254,6 +254,30 @@  static inline void tcg_gen_op5i_i64(TCGOpcode opc, TCGv_i64 arg1, TCGv_i64 arg2,
     *gen_opparam_ptr++ = arg5;
 }
 
+static inline void tcg_gen_op5ii_i32(TCGOpcode opc, TCGv_i32 arg1,
+                                     TCGv_i32 arg2, TCGv_i32 arg3,
+                                     TCGArg arg4, TCGArg arg5)
+{
+    *gen_opc_ptr++ = opc;
+    *gen_opparam_ptr++ = GET_TCGV_I32(arg1);
+    *gen_opparam_ptr++ = GET_TCGV_I32(arg2);
+    *gen_opparam_ptr++ = GET_TCGV_I32(arg3);
+    *gen_opparam_ptr++ = arg4;
+    *gen_opparam_ptr++ = arg5;
+}
+
+static inline void tcg_gen_op5ii_i64(TCGOpcode opc, TCGv_i64 arg1,
+                                     TCGv_i64 arg2, TCGv_i64 arg3,
+                                     TCGArg arg4, TCGArg arg5)
+{
+    *gen_opc_ptr++ = opc;
+    *gen_opparam_ptr++ = GET_TCGV_I64(arg1);
+    *gen_opparam_ptr++ = GET_TCGV_I64(arg2);
+    *gen_opparam_ptr++ = GET_TCGV_I64(arg3);
+    *gen_opparam_ptr++ = arg4;
+    *gen_opparam_ptr++ = arg5;
+}
+
 static inline void tcg_gen_op6_i32(TCGOpcode opc, TCGv_i32 arg1, TCGv_i32 arg2,
                                    TCGv_i32 arg3, TCGv_i32 arg4, TCGv_i32 arg5,
                                    TCGv_i32 arg6)
@@ -2071,6 +2095,44 @@  static inline void tcg_gen_rotri_i64(TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2)
     }
 }
 
+static inline void tcg_gen_deposit_i32(TCGv_i32 ret, TCGv_i32 arg1,
+				       TCGv_i32 arg2, unsigned int ofs,
+				       unsigned int len)
+{
+#ifdef TCG_TARGET_HAS_deposit_i32
+  tcg_gen_op5ii_i32(INDEX_op_deposit_i32, ret, arg1, arg2, ofs, len);
+#else
+  uint32_t mask = (1u << len) - 1;
+  TCGv_i32 t1 = tcg_temp_new_i32 ();
+
+  tcg_gen_andi_i32(t1, arg2, mask);
+  tcg_gen_shli_i32(t1, t1, ofs);
+  tcg_gen_andi_i32(ret, arg1, ~(mask << ofs));
+  tcg_gen_or_i32(ret, ret, t1);
+
+  tcg_temp_free_i32(t1);
+#endif
+}
+
+static inline void tcg_gen_deposit_i64(TCGv_i64 ret, TCGv_i64 arg1,
+				       TCGv_i64 arg2, unsigned int ofs,
+				       unsigned int len)
+{
+#ifdef TCG_TARGET_HAS_deposit_i64
+  tcg_gen_op5ii_i64(INDEX_op_deposit_i64, ret, arg1, arg2, ofs, len);
+#else
+  uint64_t mask = (1ull << len) - 1;
+  TCGv_i64 t1 = tcg_temp_new_i64 ();
+
+  tcg_gen_andi_i64(t1, arg2, mask);
+  tcg_gen_shli_i64(t1, t1, ofs);
+  tcg_gen_andi_i64(ret, arg1, ~(mask << ofs));
+  tcg_gen_or_i64(ret, ret, t1);
+
+  tcg_temp_free_i64(t1);
+#endif
+}
+
 /***************************************/
 /* QEMU specific operations. Their type depend on the QEMU CPU
    type. */
@@ -2384,6 +2446,7 @@  static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
 #define tcg_gen_rotli_tl tcg_gen_rotli_i64
 #define tcg_gen_rotr_tl tcg_gen_rotr_i64
 #define tcg_gen_rotri_tl tcg_gen_rotri_i64
+#define tcg_gen_deposit_tl tcg_gen_deposit_i64
 #define tcg_const_tl tcg_const_i64
 #define tcg_const_local_tl tcg_const_local_i64
 #else
@@ -2454,6 +2517,7 @@  static inline void tcg_gen_qemu_st64(TCGv_i64 arg, TCGv addr, int mem_index)
 #define tcg_gen_rotli_tl tcg_gen_rotli_i32
 #define tcg_gen_rotr_tl tcg_gen_rotr_i32
 #define tcg_gen_rotri_tl tcg_gen_rotri_i32
+#define tcg_gen_deposit_tl tcg_gen_deposit_i32
 #define tcg_const_tl tcg_const_i32
 #define tcg_const_local_tl tcg_const_local_i32
 #endif
diff --git a/tcg/tcg-opc.h b/tcg/tcg-opc.h
index 2a98fed..2c7ca1a 100644
--- a/tcg/tcg-opc.h
+++ b/tcg/tcg-opc.h
@@ -78,6 +78,9 @@  DEF(sar_i32, 1, 2, 0, 0)
 DEF(rotl_i32, 1, 2, 0, 0)
 DEF(rotr_i32, 1, 2, 0, 0)
 #endif
+#ifdef TCG_TARGET_HAS_deposit_i32
+DEF(deposit_i32, 1, 2, 2, 0)
+#endif
 
 DEF(brcond_i32, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
 #if TCG_TARGET_REG_BITS == 32
@@ -168,6 +171,9 @@  DEF(sar_i64, 1, 2, 0, 0)
 DEF(rotl_i64, 1, 2, 0, 0)
 DEF(rotr_i64, 1, 2, 0, 0)
 #endif
+#ifdef TCG_TARGET_HAS_deposit_i64
+DEF(deposit_i64, 1, 2, 2, 0)
+#endif
 
 DEF(brcond_i64, 0, 2, 2, TCG_OPF_BB_END | TCG_OPF_SIDE_EFFECTS)
 #ifdef TCG_TARGET_HAS_ext8s_i64