public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Szabolcs Nagy <szabolcs.nagy@arm.com>
To: libc-alpha@sourceware.org
Subject: [PATCH] Fix aliasing violation in __vfscanf_internal [BZ #26690]
Date: Wed,  7 Oct 2020 18:50:49 +0100	[thread overview]
Message-ID: <20201007175049.32564-1-szabolcs.nagy@arm.com> (raw)

Internal stdio code uses both CHAR_T and UCHAR_T strings.
But the internal helper read_int was only written for UCHAR_T.

A CHAR_T object can alias UCHAR_T, but CHAR_T * cannot alias
UCHAR_T *.  This means a cast like (UCHAR_T **)&pc is likely
a bug in the code (currently GCC does not warn about this
see PR97321). The fix introduces a read_int_char variant and
removes the problematic casts.

(The mixed use of CHAR_T and UCHAR_T may be a design mistake
in stdio: if everything used char and wchar_t consistently
then aliasing violation would be much less likely, but fixing
that requires more refactoring.)

Fixes bug 26690.
---
 stdio-common/vfscanf-internal.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/stdio-common/vfscanf-internal.c b/stdio-common/vfscanf-internal.c
index 95b46dcbeb..d832493623 100644
--- a/stdio-common/vfscanf-internal.c
+++ b/stdio-common/vfscanf-internal.c
@@ -135,6 +135,16 @@
 
 #include "printf-parse.h" /* Use read_int.  */
 
+/* Same as read_int, but for CHAR_T * instead of UCHAR_T * string.  */
+static int
+read_int_char (const CHAR_T * *pstr)
+{
+  const UCHAR_T *ustr = (const UCHAR_T *) *pstr;
+  int retval = read_int (&ustr);
+  *pstr = (const CHAR_T *) ustr;
+  return retval;
+}
+
 #define encode_error() do {						      \
 			  __set_errno (EILSEQ);				      \
 			  goto errout;					      \
@@ -486,7 +496,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
       /* Check for a positional parameter specification.  */
       if (ISDIGIT ((UCHAR_T) *f))
 	{
-	  argpos = read_int ((const UCHAR_T **) &f);
+	  argpos = read_int_char (&f);
 	  if (*f == L_('$'))
 	    ++f;
 	  else
@@ -522,7 +532,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
       /* Find the maximum field width.  */
       width = 0;
       if (ISDIGIT ((UCHAR_T) *f))
-	width = read_int ((const UCHAR_T **) &f);
+	width = read_int_char (&f);
     got_width:
       if (width == 0)
 	width = -1;
-- 
2.17.1


             reply	other threads:[~2020-10-07 17:51 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-07 17:50 Szabolcs Nagy [this message]
2020-10-07 18:08 ` Florian Weimer
2020-10-07 18:10   ` Adhemerval Zanella
2020-10-07 18:16     ` Florian Weimer
2020-10-08  7:27       ` Szabolcs Nagy

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=20201007175049.32564-1-szabolcs.nagy@arm.com \
    --to=szabolcs.nagy@arm.com \
    --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).