public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jason Merrill <jason@redhat.com>
To: gcc-patches@gcc.gnu.org
Subject: [PATCH RFA] c-family: attribute ((aligned, mode)) [PR100545]
Date: Wed, 27 Apr 2022 11:19:57 -0400	[thread overview]
Message-ID: <20220427151957.795214-1-jason@redhat.com> (raw)

The problem here was that handle_mode_attribute clobbered the changes of any
previous attribute, only copying type qualifiers to the new type.  And
common_handle_aligned_attribute had previously set up the typedef, so when
we later called set_underlying_type it saw DECL_ORIGINAL_TYPE set and just
returned, even though handle_mode_attribute had messed up the TREE_TYPE.
So, let's fix handle_mode_attribute to copy attributes, alignment, and
typedefness to the new type.

Tested x86_64-pc-linux-gnu, OK for trunk now or after 12.1?

	PR c/100545

gcc/c-family/ChangeLog:

	* c-attribs.cc (handle_mode_attribute): Copy attributes, aligned,
	and typedef.
	* c-common.cc (set_underlying_type): Add assert.

gcc/testsuite/ChangeLog:

	* c-c++-common/attr-mode-1.c: New test.
---
 gcc/c-family/c-attribs.cc                | 15 ++++++++++++++-
 gcc/c-family/c-common.cc                 |  7 ++++---
 gcc/testsuite/c-c++-common/attr-mode-1.c |  3 +++
 3 files changed, 21 insertions(+), 4 deletions(-)
 create mode 100644 gcc/testsuite/c-c++-common/attr-mode-1.c

diff --git a/gcc/c-family/c-attribs.cc b/gcc/c-family/c-attribs.cc
index 111a33f405a..26876d0f309 100644
--- a/gcc/c-family/c-attribs.cc
+++ b/gcc/c-family/c-attribs.cc
@@ -2199,7 +2199,20 @@ handle_mode_attribute (tree *node, tree name, tree args,
 	  return NULL_TREE;
 	}
 
-      *node = build_qualified_type (typefm, TYPE_QUALS (type));
+      /* Copy any quals and attributes to the new type.  */
+      *node = build_type_attribute_qual_variant (typefm, TYPE_ATTRIBUTES (type),
+						 TYPE_QUALS (type));
+      if (TYPE_USER_ALIGN (type))
+	*node = build_aligned_type (*node, TYPE_ALIGN (type));
+      if (typedef_variant_p (type))
+	{
+	  /* Set up the typedef all over again.  */
+	  tree decl = TYPE_NAME (type);
+	  DECL_ORIGINAL_TYPE (decl) = NULL_TREE;
+	  TREE_TYPE (decl) = *node;
+	  set_underlying_type (decl);
+	  *node = TREE_TYPE (decl);
+	}
     }
 
   return NULL_TREE;
diff --git a/gcc/c-family/c-common.cc b/gcc/c-family/c-common.cc
index bb0544eeaea..730faa9e87f 100644
--- a/gcc/c-family/c-common.cc
+++ b/gcc/c-family/c-common.cc
@@ -8153,15 +8153,16 @@ check_missing_format_attribute (tree ltype, tree rtype)
 void
 set_underlying_type (tree x)
 {
-  if (x == error_mark_node)
+  if (x == error_mark_node || TREE_TYPE (x) == error_mark_node)
     return;
   if (DECL_IS_UNDECLARED_BUILTIN (x) && TREE_CODE (TREE_TYPE (x)) != ARRAY_TYPE)
     {
       if (TYPE_NAME (TREE_TYPE (x)) == 0)
 	TYPE_NAME (TREE_TYPE (x)) = x;
     }
-  else if (TREE_TYPE (x) != error_mark_node
-	   && DECL_ORIGINAL_TYPE (x) == NULL_TREE)
+  else if (DECL_ORIGINAL_TYPE (x))
+    gcc_checking_assert (TYPE_NAME (TREE_TYPE (x)) == x);
+  else
     {
       tree tt = TREE_TYPE (x);
       DECL_ORIGINAL_TYPE (x) = tt;
diff --git a/gcc/testsuite/c-c++-common/attr-mode-1.c b/gcc/testsuite/c-c++-common/attr-mode-1.c
new file mode 100644
index 00000000000..59b20cd99e8
--- /dev/null
+++ b/gcc/testsuite/c-c++-common/attr-mode-1.c
@@ -0,0 +1,3 @@
+// PR c/100545
+
+typedef int fatp_t __attribute__((aligned, mode(SI)));

base-commit: 6a460a2007dd9c527c5f9d5bbbedb852db7c1373
-- 
2.27.0


             reply	other threads:[~2022-04-27 15:20 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-27 15:19 Jason Merrill [this message]
2022-04-27 16:58 ` Marek Polacek
2022-04-27 17:02 ` Joseph Myers
2022-04-27 23:48   ` Jason Merrill
2022-04-29 19:20     ` Jason Merrill
2022-04-29 19:46       ` Joseph Myers

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220427151957.795214-1-jason@redhat.com \
    --to=jason@redhat.com \
    --cc=gcc-patches@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).