gcc/0057-Backport-Add-support-for-__builtin_bswap128.patch

From 96afd5b761a74e9eef40a2e843810c503c669de8 Mon Sep 17 00:00:00 2001
From: Eric Botcazou <ebotcazou@gcc.gnu.org>
Date: Thu, 28 May 2020 00:31:15 +0200
Subject: [PATCH 09/35] [Backport] Add support for __builtin_bswap128

Reference: https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=fe7ebef7fe4f9acb79658ed9db0749b07efc3105

This patch introduces a new builtin named __builtin_bswap128 on targets
where TImode is supported, i.e. 64-bit targets only in practice.  The
implementation simply reuses the existing double word path in optab, so
no routine is added to libgcc (which means that you get two calls to
_bswapdi2 in the worst case).

gcc/ChangeLog:

	* builtin-types.def (BT_UINT128): New primitive type.
	(BT_FN_UINT128_UINT128): New function type.
	* builtins.def (BUILT_IN_BSWAP128): New GCC builtin.
	* doc/extend.texi (__builtin_bswap128): Document it.
	* builtins.c (expand_builtin): Deal with BUILT_IN_BSWAP128.
	(is_inexpensive_builtin): Likewise.
	* fold-const-call.c (fold_const_call_ss): Likewise.
	* fold-const.c (tree_call_nonnegative_warnv_p): Likewise.
	* tree-ssa-ccp.c (evaluate_stmt): Likewise.
	* tree-vect-stmts.c (vect_get_data_ptr_increment): Likewise.
	(vectorizable_call): Likewise.
	* optabs.c (expand_unop): Always use the double word path for it.
	* tree-core.h (enum tree_index): Add TI_UINT128_TYPE.
	* tree.h (uint128_type_node): New global type.
	* tree.c (build_common_tree_nodes): Build it if TImode is supported.

gcc/testsuite/ChangeLog:

	* gcc.dg/builtin-bswap-10.c: New test.
	* gcc.dg/builtin-bswap-11.c: Likewise.
	* gcc.dg/builtin-bswap-12.c: Likewise.
	* gcc.target/i386/builtin-bswap-5.c: Likewise.
---
 gcc/builtin-types.def |  4 ++++
 gcc/builtins.c        |  2 ++
 gcc/builtins.def      |  2 ++
 gcc/doc/extend.texi   | 10 ++++++++--
 gcc/fold-const-call.c |  1 +
 gcc/fold-const.c      |  2 ++
 gcc/optabs.c          |  5 ++++-
 gcc/tree-core.h       |  1 +
 gcc/tree-ssa-ccp.c    |  1 +
 gcc/tree-vect-stmts.c |  5 +++--
 gcc/tree.c            |  2 ++
 gcc/tree.h            |  1 +
 12 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/gcc/builtin-types.def b/gcc/builtin-types.def
index c7aa691b2..c46b1bc5c 100644
--- a/gcc/builtin-types.def
+++ b/gcc/builtin-types.def
@@ -73,6 +73,9 @@ DEF_PRIMITIVE_TYPE (BT_UINT8, unsigned_char_type_node)
 DEF_PRIMITIVE_TYPE (BT_UINT16, uint16_type_node)
 DEF_PRIMITIVE_TYPE (BT_UINT32, uint32_type_node)
 DEF_PRIMITIVE_TYPE (BT_UINT64, uint64_type_node)
+DEF_PRIMITIVE_TYPE (BT_UINT128, uint128_type_node
+				? uint128_type_node
+				: error_mark_node)
 DEF_PRIMITIVE_TYPE (BT_WORD, (*lang_hooks.types.type_for_mode) (word_mode, 1))
 DEF_PRIMITIVE_TYPE (BT_UNWINDWORD, (*lang_hooks.types.type_for_mode)
 				    (targetm.unwind_word_mode (), 1))
@@ -300,6 +303,7 @@ DEF_FUNCTION_TYPE_1 (BT_FN_UINT8_FLOAT, BT_UINT8, BT_FLOAT)
 DEF_FUNCTION_TYPE_1 (BT_FN_UINT16_UINT16, BT_UINT16, BT_UINT16)
 DEF_FUNCTION_TYPE_1 (BT_FN_UINT32_UINT32, BT_UINT32, BT_UINT32)
 DEF_FUNCTION_TYPE_1 (BT_FN_UINT64_UINT64, BT_UINT64, BT_UINT64)
+DEF_FUNCTION_TYPE_1 (BT_FN_UINT128_UINT128, BT_UINT128, BT_UINT128)
 DEF_FUNCTION_TYPE_1 (BT_FN_UINT64_FLOAT, BT_UINT64, BT_FLOAT)
 DEF_FUNCTION_TYPE_1 (BT_FN_BOOL_INT, BT_BOOL, BT_INT)
 DEF_FUNCTION_TYPE_1 (BT_FN_BOOL_PTR, BT_BOOL, BT_PTR)
diff --git a/gcc/builtins.c b/gcc/builtins.c
index 10b6fd3bb..1b1c75cc1 100644
--- a/gcc/builtins.c
+++ b/gcc/builtins.c
@@ -8015,6 +8015,7 @@ expand_builtin (tree exp, rtx target, rtx subtarget, machine_mode mode,
     case BUILT_IN_BSWAP16:
     case BUILT_IN_BSWAP32:
     case BUILT_IN_BSWAP64:
+    case BUILT_IN_BSWAP128:
       target = expand_builtin_bswap (target_mode, exp, target, subtarget);
       if (target)
 	return target;
@@ -11732,6 +11733,7 @@ is_inexpensive_builtin (tree decl)
       case BUILT_IN_BSWAP16:
       case BUILT_IN_BSWAP32:
       case BUILT_IN_BSWAP64:
+      case BUILT_IN_BSWAP128:
       case BUILT_IN_CLZ:
       case BUILT_IN_CLZIMAX:
       case BUILT_IN_CLZL:
diff --git a/gcc/builtins.def b/gcc/builtins.def
index fa8b0641a..ee67ac15d 100644
--- a/gcc/builtins.def
+++ b/gcc/builtins.def
@@ -834,6 +834,8 @@ DEF_GCC_BUILTIN        (BUILT_IN_APPLY_ARGS, "apply_args", BT_FN_PTR_VAR, ATTR_L
 DEF_GCC_BUILTIN        (BUILT_IN_BSWAP16, "bswap16", BT_FN_UINT16_UINT16, ATTR_CONST_NOTHROW_LEAF_LIST)
 DEF_GCC_BUILTIN        (BUILT_IN_BSWAP32, "bswap32", BT_FN_UINT32_UINT32, ATTR_CONST_NOTHROW_LEAF_LIST)
 DEF_GCC_BUILTIN        (BUILT_IN_BSWAP64, "bswap64", BT_FN_UINT64_UINT64, ATTR_CONST_NOTHROW_LEAF_LIST)
+DEF_GCC_BUILTIN        (BUILT_IN_BSWAP128, "bswap128", BT_FN_UINT128_UINT128, ATTR_CONST_NOTHROW_LEAF_LIST)
+
 DEF_EXT_LIB_BUILTIN    (BUILT_IN_CLEAR_CACHE, "__clear_cache", BT_FN_VOID_PTR_PTR, ATTR_NOTHROW_LEAF_LIST)
 /* [trans-mem]: Adjust BUILT_IN_TM_CALLOC if BUILT_IN_CALLOC is changed.  */
 DEF_LIB_BUILTIN        (BUILT_IN_CALLOC, "calloc", BT_FN_PTR_SIZE_SIZE, ATTR_MALLOC_WARN_UNUSED_RESULT_SIZE_1_2_NOTHROW_LEAF_LIST)
diff --git a/gcc/doc/extend.texi b/gcc/doc/extend.texi
index 9c7345959..a7bd772de 100644
--- a/gcc/doc/extend.texi
+++ b/gcc/doc/extend.texi
@@ -13727,14 +13727,20 @@ exactly 8 bits.
 
 @deftypefn {Built-in Function} uint32_t __builtin_bswap32 (uint32_t x)
 Similar to @code{__builtin_bswap16}, except the argument and return types
-are 32 bit.
+are 32-bit.
 @end deftypefn
 
 @deftypefn {Built-in Function} uint64_t __builtin_bswap64 (uint64_t x)
 Similar to @code{__builtin_bswap32}, except the argument and return types
-are 64 bit.
+are 64-bit.
 @end deftypefn
 
+@deftypefn {Built-in Function} uint128_t __builtin_bswap128 (uint128_t x)
+Similar to @code{__builtin_bswap64}, except the argument and return types
+are 128-bit.  Only supported on targets when 128-bit types are supported.
+@end deftypefn
+
+
 @deftypefn {Built-in Function} Pmode __builtin_extend_pointer (void * x)
 On targets where the user visible pointer size is smaller than the size
 of an actual hardware address this function returns the extended user
diff --git a/gcc/fold-const-call.c b/gcc/fold-const-call.c
index 6150d7ada..da01759d9 100644
--- a/gcc/fold-const-call.c
+++ b/gcc/fold-const-call.c
@@ -1032,6 +1032,7 @@ fold_const_call_ss (wide_int *result, combined_fn fn, const wide_int_ref &arg,
     case CFN_BUILT_IN_BSWAP16:
     case CFN_BUILT_IN_BSWAP32:
     case CFN_BUILT_IN_BSWAP64:
+    case CFN_BUILT_IN_BSWAP128:
       *result = wide_int::from (arg, precision, TYPE_SIGN (arg_type)).bswap ();
       return true;
 
diff --git a/gcc/fold-const.c b/gcc/fold-const.c
index 6e635382f..78227a83d 100644
--- a/gcc/fold-const.c
+++ b/gcc/fold-const.c
@@ -13889,8 +13889,10 @@ tree_call_nonnegative_warnv_p (tree type, combined_fn fn, tree arg0, tree arg1,
     CASE_CFN_POPCOUNT:
     CASE_CFN_CLZ:
     CASE_CFN_CLRSB:
+    case CFN_BUILT_IN_BSWAP16:
     case CFN_BUILT_IN_BSWAP32:
     case CFN_BUILT_IN_BSWAP64:
+    case CFN_BUILT_IN_BSWAP128:
       /* Always true.  */
       return true;
 
diff --git a/gcc/optabs.c b/gcc/optabs.c
index 049a18ceb..c3751fdf7 100644
--- a/gcc/optabs.c
+++ b/gcc/optabs.c
@@ -2896,8 +2896,11 @@ expand_unop (machine_mode mode, optab unoptab, rtx op0, rtx target,
 	  if (temp)
 	    return temp;
 
+	  /* We do not provide a 128-bit bswap in libgcc so force the use of
+	     a double bswap for 64-bit targets.  */
 	  if (GET_MODE_SIZE (int_mode) == 2 * UNITS_PER_WORD
-	      && optab_handler (unoptab, word_mode) != CODE_FOR_nothing)
+	      && (UNITS_PER_WORD == 64
+		  || optab_handler (unoptab, word_mode) != CODE_FOR_nothing))
 	    {
 	      temp = expand_doubleword_bswap (mode, op0, target);
 	      if (temp)
diff --git a/gcc/tree-core.h b/gcc/tree-core.h
index eb01c2434..058e046aa 100644
--- a/gcc/tree-core.h
+++ b/gcc/tree-core.h
@@ -600,6 +600,7 @@ enum tree_index {
   TI_UINT16_TYPE,
   TI_UINT32_TYPE,
   TI_UINT64_TYPE,
+  TI_UINT128_TYPE,
 
   TI_VOID,
 
diff --git a/gcc/tree-ssa-ccp.c b/gcc/tree-ssa-ccp.c
index 952fd9cd4..dcdf10369 100644
--- a/gcc/tree-ssa-ccp.c
+++ b/gcc/tree-ssa-ccp.c
@@ -2005,6 +2005,7 @@ evaluate_stmt (gimple *stmt)
 	    case BUILT_IN_BSWAP16:
 	    case BUILT_IN_BSWAP32:
 	    case BUILT_IN_BSWAP64:
+	    case BUILT_IN_BSWAP128:
 	      val = get_value_for_expr (gimple_call_arg (stmt, 0), true);
 	      if (val.lattice_val == UNDEFINED)
 		break;
diff --git a/gcc/tree-vect-stmts.c b/gcc/tree-vect-stmts.c
index b872cfc8d..4636b7ba2 100644
--- a/gcc/tree-vect-stmts.c
+++ b/gcc/tree-vect-stmts.c
@@ -3085,7 +3085,7 @@ vect_get_data_ptr_increment (dr_vec_info *dr_info, tree aggr_type,
   return iv_step;
 }
 
-/* Check and perform vectorization of BUILT_IN_BSWAP{16,32,64}.  */
+/* Check and perform vectorization of BUILT_IN_BSWAP{16,32,64,128}.  */
 
 static bool
 vectorizable_bswap (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
@@ -3454,7 +3454,8 @@ vectorizable_call (stmt_vec_info stmt_info, gimple_stmt_iterator *gsi,
       else if (modifier == NONE
 	       && (gimple_call_builtin_p (stmt, BUILT_IN_BSWAP16)
 		   || gimple_call_builtin_p (stmt, BUILT_IN_BSWAP32)
-		   || gimple_call_builtin_p (stmt, BUILT_IN_BSWAP64)))
+		   || gimple_call_builtin_p (stmt, BUILT_IN_BSWAP64)
+		   || gimple_call_builtin_p (stmt, BUILT_IN_BSWAP128)))
 	return vectorizable_bswap (stmt_info, gsi, vec_stmt, slp_node,
 				   vectype_in, cost_vec);
       else
diff --git a/gcc/tree.c b/gcc/tree.c
index 84a440b35..3e6647ae0 100644
--- a/gcc/tree.c
+++ b/gcc/tree.c
@@ -10394,6 +10394,8 @@ build_common_tree_nodes (bool signed_char)
   uint16_type_node = make_or_reuse_type (16, 1);
   uint32_type_node = make_or_reuse_type (32, 1);
   uint64_type_node = make_or_reuse_type (64, 1);
+  if (targetm.scalar_mode_supported_p (TImode))
+    uint128_type_node = make_or_reuse_type (128, 1);
 
   /* Decimal float types. */
   if (targetm.decimal_float_supported_p ())
diff --git a/gcc/tree.h b/gcc/tree.h
index 328a2d5d2..bddc6e528 100644
--- a/gcc/tree.h
+++ b/gcc/tree.h
@@ -4035,6 +4035,7 @@ tree_strip_any_location_wrapper (tree exp)
 #define uint16_type_node		global_trees[TI_UINT16_TYPE]
 #define uint32_type_node		global_trees[TI_UINT32_TYPE]
 #define uint64_type_node		global_trees[TI_UINT64_TYPE]
+#define uint128_type_node		global_trees[TI_UINT128_TYPE]
 
 #define void_node			global_trees[TI_VOID]
 
-- 
2.27.0.windows.1