public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Carlos O'Donell <carlos@redhat.com>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: libc-alpha@sourceware.org
Subject: Re: [PATCH 2/6] posix: Suppress -Os warnings on fnmatch
Date: Wed, 5 Oct 2022 09:47:48 -0400	[thread overview]
Message-ID: <Yz2LBLnzKy1zj3V1@fedora> (raw)
In-Reply-To: <20220921135108.3324737-3-adhemerval.zanella@linaro.org>

On Wed, Sep 21, 2022 at 10:51:04AM -0300, Adhemerval Zanella via Libc-alpha wrote:
> GCC with -Os issues some may uninitialized warnings on fnmatch code.
> All of the variables are already set when they are accessed on the
> loop prior.
> 
> Checked on x86_64-linux-gnu and i686-linux-gnu.

LGTM.

This touches gnulib code and may require merging upstream or harmonizing
in a different way. For now this looks ok. There are some alternatives,
but this is OK.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>

> ---
>  posix/fnmatch_loop.c | 31 +++++++++++++++++++++++++++++++
>  1 file changed, 31 insertions(+)
> 
> diff --git a/posix/fnmatch_loop.c b/posix/fnmatch_loop.c
> index 9445ed9c58..23eeffc79c 100644
> --- a/posix/fnmatch_loop.c
> +++ b/posix/fnmatch_loop.c
> @@ -530,6 +530,14 @@ FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
>                                {
>                                  /* Compare the byte sequence but only if
>                                     this is not part of a range.  */
> +
> +				/* The compiler might warn that idx may be
> +				   used uninitialized, however it will be
> +				   reached iff elem < table_size which means
> +				   that it was properly set in the loop
> +				   above.   */
> +                                DIAG_PUSH_NEEDS_COMMENT;
> +                                DIAG_IGNORE_Os_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
>                                  if (! is_range
>  
>  # if WIDE_CHAR_VERSION
> @@ -542,11 +550,19 @@ FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
>                                      n += c1 - 1;
>                                      goto matched;
>                                    }
> +                                DIAG_POP_NEEDS_COMMENT;

OK.

>  
>                                  /* Get the collation sequence value.  */
>                                  is_seqval = true;
>  # if WIDE_CHAR_VERSION
> +                                /* The compile might warn that wextra may be
> +				   used uninitialized and similar to 'idx'
> +				   above it will be properly set by the loop.
> +				   */
> +                                DIAG_PUSH_NEEDS_COMMENT;
> +                                DIAG_IGNORE_Os_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
>                                  cold = wextra[1 + wextra[0]];
> +                                DIAG_POP_NEEDS_COMMENT;

OK.

>  # else
>                                  idx += 1 + extra[idx];
>                                  /* Adjust for the alignment.  */
> @@ -723,9 +739,24 @@ FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
>                                      /* Get the collation sequence value.  */
>                                      is_seqval = true;
>  # if WIDE_CHAR_VERSION
> +                                    /* The compiler might warn that wextra may
> +                                       be used uninitialized, however it will
> +                                       be reached iff elem < table_size which
> +                                       means that it was properly set in the
> +                                       loop above.   */
> +                                    DIAG_PUSH_NEEDS_COMMENT;
> +                                    DIAG_IGNORE_Os_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
>                                      cend = wextra[1 + wextra[0]];
> +                                    DIAG_POP_NEEDS_COMMENT;

OK.

>  # else
> +				    /* The compile might warn that idx may
> +				       be used uninitialized and similar to
> +				       wextra above it will be properly set by
> +				       the loop.   */
> +                                    DIAG_PUSH_NEEDS_COMMENT;
> +                                    DIAG_IGNORE_Os_NEEDS_COMMENT (8, "-Wmaybe-uninitialized");
>                                      idx += 1 + extra[idx];
> +                                    DIAG_POP_NEEDS_COMMENT;

OK.

>                                      /* Adjust for the alignment.  */
>                                      idx = (idx + 3) & ~3;
>                                      cend = *((int32_t *) &extra[idx]);
> -- 
> 2.34.1
> 


  reply	other threads:[~2022-10-05 13:47 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-21 13:51 [PATCH 0/6] Fix -Os build Adhemerval Zanella
2022-09-21 13:51 ` [PATCH 1/6] locale: prevent maybe-uninitialized errors with -Os [BZ #19444] Adhemerval Zanella
2022-10-05 13:39   ` Carlos O'Donell
2022-09-21 13:51 ` [PATCH 2/6] posix: Suppress -Os warnings on fnmatch Adhemerval Zanella
2022-10-05 13:47   ` Carlos O'Donell [this message]
2022-10-05 20:01     ` Paul Eggert
2022-09-21 13:51 ` [PATCH 3/6] posix: Suppress -Os may be used uninitialized warnings on regexec Adhemerval Zanella
2022-10-05 13:49   ` Carlos O'Donell
2022-09-21 13:51 ` [PATCH 4/6] rt: Initialize mq_send input on tst-mqueue{5,6} Adhemerval Zanella
2022-10-05 13:51   ` Carlos O'Donell
2022-09-21 13:51 ` [PATCH 5/6] sunrpc: Suppress GCC -Os warning on user2netname Adhemerval Zanella
2022-10-05 13:52   ` Carlos O'Donell
2022-09-21 13:51 ` [PATCH 6/6] x86: Fix -Os build (BZ #29576) Adhemerval Zanella
2022-10-05 14:10   ` Carlos O'Donell
2022-10-05 17:21     ` H.J. Lu
2022-10-05 17:38       ` Adhemerval Zanella Netto
2022-10-05 18:06         ` H.J. Lu
2022-10-05 20:43           ` Adhemerval Zanella Netto
2022-10-05 21:27             ` H.J. Lu
2022-10-05 13:40 ` [PATCH 0/6] Fix -Os build Carlos O'Donell

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=Yz2LBLnzKy1zj3V1@fedora \
    --to=carlos@redhat.com \
    --cc=adhemerval.zanella@linaro.org \
    --cc=libc-alpha@sourceware.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).