public inbox for
help / color / mirror / Atom feed
From: Adhemerval Zanella <>
Subject: [glibc] string: Suppress -Wmaybe-unitialized for wordcopy [BZ #19444]
Date: Thu, 12 Jan 2023 12:24:21 +0000 (GMT)	[thread overview]
Message-ID: <> (raw);h=402853be1db5a233ec688f8c5343b73191629fbd

commit 402853be1db5a233ec688f8c5343b73191629fbd
Author: Adhemerval Zanella <>
Date:   Wed Jan 11 18:27:29 2023 -0300

    string: Suppress -Wmaybe-unitialized for wordcopy [BZ #19444]
    When compiling with GCC 6+ the sparc build warns that some variables
    might be used uninitialized.  However it does not seem the fact, since
    the variables are really initialized (and also other targets that use the
    same code, like powerpc, do not warn about it).
    So suppress the warning for now.
    Changes from v1:
    * Update patch description and the explanation for the suppresion.
    Reviewed-by: Carlos O'Donell <>

 string/wordcopy.c | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/string/wordcopy.c b/string/wordcopy.c
index ae5ccd793c..a8a40c9f01 100644
--- a/string/wordcopy.c
+++ b/string/wordcopy.c
@@ -19,7 +19,19 @@
 #include <stddef.h>
+#include <libc-diag.h>
+/* Compiling with -O1 might warn that 'a2' and 'a3' may be used
+   uninitialized.  There are only two ways to arrive at labels 'do4', 'do3'
+   or 'do1', all of which use 'a2' or 'a3' in the MERGE macro: either from
+   the earlier switch case statement or via a loop iteration.  In all cases
+   the switch statement or previous loop sets both 'a2' and 'a3'.
+   Since the usage is within the MERGE macro we disable the
+   warning in the definition, but only in this file.  */
+DIAG_IGNORE_NEEDS_COMMENT (6, "-Wmaybe-uninitialized");
 #include <memcopy.h>
 /* _wordcopy_fwd_aligned -- Copy block beginning at SRCP to
    block beginning at DSTP with LEN `op_t' words (not LEN bytes!).
@@ -94,7 +106,15 @@ WORDCOPY_FWD_ALIGNED (long int dstp, long int srcp, size_t len)
       a0 = ((op_t *) srcp)[0];
+      /* Compiling with -O1 may warn that 'a1' may be used uninitialized.
+	 There are only two ways to arrive at label 'do8' and they are via a
+	 do-while loop iteration or directly via the earlier switch 'case 1:'
+	 case. The switch case always sets 'a1' and all previous loop
+	 iterations will also have set 'a1' before the use.  */
+      DIAG_IGNORE_NEEDS_COMMENT (6, "-Wmaybe-uninitialized");
       ((op_t *) dstp)[0] = a1;
       a1 = ((op_t *) srcp)[1];
       ((op_t *) dstp)[1] = a0;
@@ -291,7 +311,11 @@ WORDCOPY_BWD_ALIGNED (long int dstp, long int srcp, size_t len)
       a0 = ((op_t *) srcp)[7];
+      /* Check the comment on WORDCOPY_FWD_ALIGNED.  */
+      DIAG_IGNORE_NEEDS_COMMENT (6, "-Wmaybe-uninitialized");
       ((op_t *) dstp)[7] = a1;
       a1 = ((op_t *) srcp)[6];
       ((op_t *) dstp)[6] = a0;

                 reply	other threads:[~2023-01-12 12:24 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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:

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

  git send-email \ \ \ \

* 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).