From: Robin Dapp <rdapp@linux.ibm.com>
To: GCC Patches <gcc-patches@gcc.gnu.org>
Subject: [PATCH 1/2] c-family: Copy DECL_USER_ALIGN even if DECL_ALIGN is similar.
Date: Mon, 3 May 2021 16:53:14 +0200 [thread overview]
Message-ID: <dbec56ab-a57a-d84f-1286-2f46d049bcae@linux.ibm.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 2573 bytes --]
Hi,
on s390 a warning test fails:
inline int ATTR ((cold, aligned (8)))
finline_hot_noret_align (int);
inline int ATTR ((warn_unused_result))
finline_hot_noret_align (int);
inline int ATTR ((aligned (4)))
finline_hot_noret_align (int); /* { dg-warning "ignoring attribute
.aligned \\(4\\). because it conflicts with attribute .aligned \\(8\\)."
This test actually uncovered two problems. First, on s390 the default
function alignment is 8 bytes. When the second decl above is merged
with the first one, DECL_USER_ALIGN is only copied if DECL_ALIGN (old) >
DECL_ALIGN (new). Subsequently, when merging the third decl, no warning
is emitted since DECL_USER_ALIGN is unset.
This patch also copies DECL_USER_ALIGN if DECL_ALIGN (old) == DECL_ALIGN
(new) && DECL_USER_ALIGN (olddecl) != DECL_USER_ALIGN (newdecl)).
Then, while going through the related files I also noticed that we emit
a wrong warning for:
inline int ATTR ((aligned (32)))
finline_align (int);
inline int ATTR ((aligned (4)))
finline_align (int); /* { dg-warning "ignoring attribute .aligned
\\(4\\). because it conflicts with attribute .aligned \\(32\\)." "" } */
What we emit is
warning: ignoring attribute ‘aligned (4)’ because it conflicts with
attribute ‘aligned (8)’ [-Wattributes].
This is due to the short circuit evaluation in c-attribs.c:
&& ((curalign = DECL_ALIGN (decl)) > bitalign
|| ((lastalign = DECL_ALIGN (last_decl)) > bitalign)))
where lastalign is only initialized when curalign > bitalign. On s390
this is not the case and lastalign is used zero-initialized in the
following code.
On top, the following condition
else if (!warn_if_not_aligned_p
&& TREE_CODE (decl) == FUNCTION_DECL
&& DECL_ALIGN (decl) > bitalign)
seems to be fully covered by
else if (TREE_CODE (decl) == FUNCTION_DECL
&& ((curalign = DECL_ALIGN (decl)) > bitalign
|| ((lastalign = DECL_ALIGN (last_decl)) > bitalign)))
and so is essentially dead. I therefore removed it as there does not
seem to be a test anywhere for the error message ( "alignment for %q+D
was previously specified as %d and may not be decreased") either.
Bootstrapped and regtested on s390, ppc64 and amd64.
Is it OK?
Regards
Robin
--
gcc/c-family/ChangeLog:
* c-attribs.c (common_handle_aligned_attribute): Re-init
lastalign and remove case.
gcc/c/ChangeLog:
* c-decl.c (merge_decls): Copy DECL_USER_ALIGN even for similar
DECL_ALIGN.
gcc/cp/ChangeLog:
* decl.c (duplicate_decls): Dito.
[-- Attachment #2: 0001-c-family-Copy-DECL_USER_ALIGN-even-if-DECL_ALIGN-is-.patch --]
[-- Type: text/x-patch, Size: 4030 bytes --]
From f3a9f8b9bd1e28c23f2efcdb061959852a91deb6 Mon Sep 17 00:00:00 2001
From: Robin Dapp <rdapp@linux.ibm.com>
Date: Tue, 20 Apr 2021 10:51:18 +0200
Subject: [PATCH 1/2] c-family: Copy DECL_USER_ALIGN even if DECL_ALIGN is
similar.
When re-declaring a function with differing attributes DECL_USER_ALIGN
is usually not merged/copied when DECL_ALIGN is similar. On s390 this
will cause a warning message not to be shown. Similarly, a note is not
shown when we short-circuit an alignment initialization in
common_handle_aligned_attribute ().
Fix this by copying DECL_USER_ALIGN even if DECL_ALIGN is similar as
well as reinitializing the short-circuited initialization.
---
gcc/c-family/c-attribs.c | 23 ++++-------------------
gcc/c/c-decl.c | 3 +++
gcc/cp/decl.c | 5 +++++
3 files changed, 12 insertions(+), 19 deletions(-)
diff --git a/gcc/c-family/c-attribs.c b/gcc/c-family/c-attribs.c
index c1f652d1dc9..d132b6fd3b6 100644
--- a/gcc/c-family/c-attribs.c
+++ b/gcc/c-family/c-attribs.c
@@ -2317,6 +2317,10 @@ common_handle_aligned_attribute (tree *node, tree name, tree args, int flags,
&& ((curalign = DECL_ALIGN (decl)) > bitalign
|| ((lastalign = DECL_ALIGN (last_decl)) > bitalign)))
{
+ /* Re-init lastalign in case we short-circuit the condition,
+ i.e. curalign > bitalign. */
+ lastalign = DECL_ALIGN (last_decl);
+
/* Either a prior attribute on the same declaration or one
on a prior declaration of the same function specifies
stricter alignment than this attribute. */
@@ -2366,25 +2370,6 @@ common_handle_aligned_attribute (tree *node, tree name, tree args, int flags,
This formally comes from the c++11 specification but we are
doing it for the GNU attribute syntax as well. */
*no_add_attrs = true;
- else if (!warn_if_not_aligned_p
- && TREE_CODE (decl) == FUNCTION_DECL
- && DECL_ALIGN (decl) > bitalign)
- {
- /* Don't warn for function alignment here if warn_if_not_aligned_p
- is true. It will be warned about later. */
- if (DECL_USER_ALIGN (decl))
- {
- /* Only reject attempts to relax/override an alignment
- explicitly specified previously and accept declarations
- that appear to relax the implicit function alignment for
- the target. Both increasing and increasing the alignment
- set by -falign-functions setting is permitted. */
- error ("alignment for %q+D was previously specified as %d "
- "and may not be decreased", decl,
- DECL_ALIGN (decl) / BITS_PER_UNIT);
- *no_add_attrs = true;
- }
- }
else if (warn_if_not_aligned_p
&& TREE_CODE (decl) == FIELD_DECL
&& !DECL_C_BIT_FIELD (decl))
diff --git a/gcc/c/c-decl.c b/gcc/c/c-decl.c
index 3ea4708c507..2ea5051e9cd 100644
--- a/gcc/c/c-decl.c
+++ b/gcc/c/c-decl.c
@@ -2620,6 +2620,9 @@ merge_decls (tree newdecl, tree olddecl, tree newtype, tree oldtype)
SET_DECL_ALIGN (newdecl, DECL_ALIGN (olddecl));
DECL_USER_ALIGN (newdecl) |= DECL_USER_ALIGN (olddecl);
}
+ else if (DECL_ALIGN (olddecl) == DECL_ALIGN (newdecl)
+ && DECL_USER_ALIGN (olddecl) != DECL_USER_ALIGN (newdecl))
+ DECL_USER_ALIGN (newdecl) |= DECL_USER_ALIGN (olddecl);
if (DECL_WARN_IF_NOT_ALIGN (olddecl)
> DECL_WARN_IF_NOT_ALIGN (newdecl))
SET_DECL_WARN_IF_NOT_ALIGN (newdecl,
diff --git a/gcc/cp/decl.c b/gcc/cp/decl.c
index 60dc2bf182d..ad07b1a4e63 100644
--- a/gcc/cp/decl.c
+++ b/gcc/cp/decl.c
@@ -2795,6 +2795,11 @@ duplicate_decls (tree newdecl, tree olddecl, bool hiding, bool was_hidden)
SET_DECL_ALIGN (newdecl, DECL_ALIGN (olddecl));
DECL_USER_ALIGN (newdecl) |= DECL_USER_ALIGN (olddecl);
}
+ else if (DECL_ALIGN (olddecl) == DECL_ALIGN (newdecl)
+ && DECL_USER_ALIGN (olddecl) != DECL_USER_ALIGN (newdecl))
+ {
+ DECL_USER_ALIGN (newdecl) |= DECL_USER_ALIGN (olddecl);
+ }
DECL_USER_ALIGN (olddecl) = DECL_USER_ALIGN (newdecl);
if (DECL_WARN_IF_NOT_ALIGN (olddecl)
> DECL_WARN_IF_NOT_ALIGN (newdecl))
--
2.23.0
next reply other threads:[~2021-05-03 14:53 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-03 14:53 Robin Dapp [this message]
2021-05-10 6:32 ` Robin Dapp
2021-05-17 14:03 ` Robin Dapp
2021-05-19 22:03 ` Martin Sebor
2021-05-21 19:13 ` Jason Merrill
2021-05-25 10:38 ` Robin Dapp
2021-05-25 15:15 ` Martin Sebor
2021-05-25 15:56 ` Jason Merrill
2021-06-01 13:20 ` Robin Dapp
2021-06-01 19:11 ` Jason Merrill
2021-06-09 8:47 ` Robin Dapp
2021-06-09 21:24 ` Jason Merrill
2021-06-01 13:24 ` Robin Dapp
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=dbec56ab-a57a-d84f-1286-2f46d049bcae@linux.ibm.com \
--to=rdapp@linux.ibm.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).