varasm: Redo mergeable section support [PR121438]

We increased the switch conversion array decl alignment
for better mergeability but it turns out that we increase
the alignment on targets which don't support mergeable sections
(e.g. NVPTX). Also after the fix for PR 121394, it becomes
obvious that we can place any sized into the mergeable section
instead of increasing the alignment.
This implements that and now also fixes PR 121438 as we don't
need to increase the alignment for the mergeable decls that
were being created by the C++ front-end.

Bootstrapped and tested on x86_64-linux-gnu.

	PR middle-end/121438
	PR middle-end/121444
gcc/ChangeLog:

	* output.h (MAX_ALIGN_MERGABLE): Rename to ...
	(MAX_MERGEABLE_BITSIZE): This.
	* tree-switch-conversion.cc (switch_conversion::build_one_array): Don't
	increase the alignment.
	* varasm.cc (mergeable_string_section): Use MAX_MERGEABLE_BITSIZE
	instead of MAX_ALIGN_MERGABLE. Also replace `/ 8` with `/ BITS_PER_UNIT`.
	(mergeable_constant_section): Select the mergeable section based on
	the bitsize rather than the alignment. Make sure the align is less
	than the entity size.

Signed-off-by: Andrew Pinski <quic_apinski@quicinc.com>

3 files changed, 13 insertions, 21 deletions

diff --git a/gcc/output.h b/gcc/output.h
index 51c2d36..835a259 100644
--- a/gcc/output.h
+++ b/gcc/output.h
@@ -545,8 +545,8 @@ extern GTY(()) section *bss_noswitch_section;
 extern GTY(()) section *in_section;
 extern GTY(()) bool in_cold_section_p;
 
-/* MAX bit alignment for mergable sections. */
-#define MAX_ALIGN_MERGABLE 256
+/* MAX size for mergeable sections in bits. */
+#define MAX_MERGEABLE_BITSIZE 256
 
 extern section *get_unnamed_section (unsigned int, void (*) (const char *),
 				     const char *);
diff --git a/gcc/tree-switch-conversion.cc b/gcc/tree-switch-conversion.cc
index 04b357f..62eddcd 100644
--- a/gcc/tree-switch-conversion.cc
+++ b/gcc/tree-switch-conversion.cc
@@ -55,7 +55,6 @@ Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA
 #include "hwint.h"
 #include "internal-fn.h"
 #include "diagnostic-core.h"
-#include "output.h"
 
 /* ??? For lang_hooks.types.type_for_mode, but is there a word_mode
    type in the GIMPLE type system that is language-independent?  */
@@ -1031,19 +1030,10 @@ switch_conversion::build_one_array (int num, tree arr_index_type,
       TREE_CONSTANT (decl) = 1;
       TREE_READONLY (decl) = 1;
       DECL_IGNORED_P (decl) = 1;
-      /* The decl is mergable since we don't take the address ever and
+      /* The decl is mergeable since we don't take the address ever and
 	 just reading from it. */
       DECL_MERGEABLE (decl) = 1;
 
-      /* Increase the alignments as needed. */
-      if (tree_to_uhwi (DECL_SIZE (decl)) > DECL_ALIGN (decl))
-	{
-	  unsigned HOST_WIDE_INT s = tree_to_uhwi (DECL_SIZE (decl));
-	  /* Only support up to the max supported for merging. */
-	  if (s <= MAX_ALIGN_MERGABLE)
-	    SET_DECL_ALIGN (decl, HOST_WIDE_INT_1U << ceil_log2 (s));
-	}
-
       if (offloading_function_p (cfun->decl))
 	DECL_ATTRIBUTES (decl)
 	  = tree_cons (get_identifier ("omp declare target"), NULL_TREE,
diff --git a/gcc/varasm.cc b/gcc/varasm.cc
index ee32cf1..0d78f5b 100644
--- a/gcc/varasm.cc
+++ b/gcc/varasm.cc
@@ -871,7 +871,7 @@ mergeable_string_section (tree decl ATTRIBUTE_UNUSED,
   if (HAVE_GAS_SHF_MERGE && flag_merge_constants
       && TREE_CODE (decl) == STRING_CST
       && TREE_CODE (TREE_TYPE (decl)) == ARRAY_TYPE
-      && align <= MAX_ALIGN_MERGABLE
+      && align <= MAX_MERGEABLE_BITSIZE
       && (len = int_size_in_bytes (TREE_TYPE (decl))) > 0
       && TREE_STRING_LENGTH (decl) == len)
     {
@@ -885,7 +885,7 @@ mergeable_string_section (tree decl ATTRIBUTE_UNUSED,
 
       mode = SCALAR_INT_TYPE_MODE (TREE_TYPE (TREE_TYPE (decl)));
       modesize = GET_MODE_BITSIZE (mode);
-      if (modesize >= 8 && modesize <= MAX_ALIGN_MERGABLE
+      if (modesize >= 8 && modesize <= MAX_MERGEABLE_BITSIZE
 	  && (modesize & (modesize - 1)) == 0)
 	{
 	  if (align < modesize)
@@ -906,8 +906,8 @@ mergeable_string_section (tree decl ATTRIBUTE_UNUSED,
 	  if (i == len - unit || (unit == 1 && i == len))
 	    {
 	      sprintf (name, "%s.str%d.%d", prefix,
-		       modesize / 8, (int) (align / 8));
-	      flags |= (modesize / 8) | SECTION_MERGE | SECTION_STRINGS;
+		       modesize / BITS_PER_UNIT, (int) (align / BITS_PER_UNIT));
+	      flags |= (modesize / BITS_PER_UNIT) | SECTION_MERGE | SECTION_STRINGS;
 	      return get_section (name, flags, NULL);
 	    }
 	}
@@ -923,17 +923,19 @@ mergeable_constant_section (unsigned HOST_WIDE_INT size_bits,
 			    unsigned HOST_WIDE_INT align,
 			    unsigned int flags)
 {
+  unsigned HOST_WIDE_INT newsize;
+  newsize = HOST_WIDE_INT_1U << ceil_log2 (size_bits);
   if (HAVE_GAS_SHF_MERGE && flag_merge_constants
-      && size_bits <= align
+      && newsize <= MAX_MERGEABLE_BITSIZE
       && align >= 8
-      && align <= MAX_ALIGN_MERGABLE
+      && align <= newsize
       && (align & (align - 1)) == 0)
     {
       const char *prefix = function_mergeable_rodata_prefix ();
       char *name = (char *) alloca (strlen (prefix) + 30);
 
-      sprintf (name, "%s.cst%d", prefix, (int) (align / 8));
-      flags |= (align / 8) | SECTION_MERGE;
+      sprintf (name, "%s.cst%d", prefix, (int) (newsize / BITS_PER_UNIT));
+      flags |= (newsize / BITS_PER_UNIT) | SECTION_MERGE;
       return get_section (name, flags, NULL);
     }
   return readonly_data_section;