From: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org, Joe Simmons-Talbott <josimmon@redhat.com>
Subject: Re: [PATCH] vfscanf-internal: Remove potentially unbounded allocas
Date: Mon, 3 Jul 2023 11:39:35 -0300 [thread overview]
Message-ID: <3baffc46-3a2f-85ca-f5a4-b1857d7c44a7@linaro.org> (raw)
In-Reply-To: <20230703143829.2256518-1-adhemerval.zanella@linaro.org>
On 03/07/23 11:38, Adhemerval Zanella wrote:
> From: Joe Simmons-Talbott <josimmon@redhat.com>
Sorry about that, I forgot to adjust the author.
>
> Some locales define a list of mapping pairs of alternate digits and
> separators for input digits (to_inpunct). This require the scanf
> to create a list of all possible inputs for the optional type
> modifier 'I'.
>
> Checked on x86_64-linux-gnu.
> ---
> stdio-common/Makefile | 3 ++
> stdio-common/tst-scanf-to_inpunct.c | 78 ++++++++++++++++++++++++++++
> stdio-common/vfscanf-internal.c | 40 ++++++++-------
> wcsmbs/Makefile | 3 ++
> wcsmbs/tst-wscanf-to_inpunct.c | 79 +++++++++++++++++++++++++++++
> 5 files changed, 186 insertions(+), 17 deletions(-)
> create mode 100644 stdio-common/tst-scanf-to_inpunct.c
> create mode 100644 wcsmbs/tst-wscanf-to_inpunct.c
>
> diff --git a/stdio-common/Makefile b/stdio-common/Makefile
> index 8871ec7668..f6d9017ff1 100644
> --- a/stdio-common/Makefile
> +++ b/stdio-common/Makefile
> @@ -231,6 +231,7 @@ tests := \
> tst-scanf-binary-gnu11 \
> tst-scanf-binary-gnu89 \
> tst-scanf-round \
> + tst-scanf-to_inpunct \
> tst-setvbuf1 \
> tst-sprintf \
> tst-sprintf-errno \
> @@ -347,6 +348,7 @@ LOCALES := \
> de_DE.ISO-8859-1 \
> de_DE.UTF-8 \
> en_US.ISO-8859-1 \
> + fa_IR.UTF-8 \
> hi_IN.UTF-8 \
> ja_JP.EUC-JP \
> ps_AF.UTF-8 \
> @@ -366,6 +368,7 @@ $(objpfx)tst-swprintf.out: $(gen-locales)
> $(objpfx)tst-vfprintf-mbs-prec.out: $(gen-locales)
> $(objpfx)tst-vfprintf-width-i18n.out: $(gen-locales)
> $(objpfx)tst-grouping3.out: $(gen-locales)
> +$(objpfx)tst-scanf-to_inpunct.out: $(gen-locales)
> endif
>
> tst-printf-bz18872-ENV = MALLOC_TRACE=$(objpfx)tst-printf-bz18872.mtrace \
> diff --git a/stdio-common/tst-scanf-to_inpunct.c b/stdio-common/tst-scanf-to_inpunct.c
> new file mode 100644
> index 0000000000..32236ac2dc
> --- /dev/null
> +++ b/stdio-common/tst-scanf-to_inpunct.c
> @@ -0,0 +1,78 @@
> +/* Test scanf for languages with mapping pairs of alternate digits and
> + separators.
> + Copyright (C) 2023 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <https://www.gnu.org/licenses/>. */
> +
> +#include <array_length.h>
> +#include <stdio.h>
> +#include <support/support.h>
> +#include <support/check.h>
> +
> +/* fa_IR defines to_inpunct for numbers. */
> +static const struct
> +{
> + int n;
> + const char *str;
> +} inputs[] =
> +{
> + { 1, "\xdb\xb1" },
> + { 2, "\xdb\xb2" },
> + { 3, "\xdb\xb3" },
> + { 4, "\xdb\xb4" },
> + { 5, "\xdb\xb5" },
> + { 6, "\xdb\xb6" },
> + { 7, "\xdb\xb7" },
> + { 8, "\xdb\xb8" },
> + { 9, "\xdb\xb9" },
> + { 10, "\xdb\xb1\xdb\xb0" },
> + { 11, "\xdb\xb1\xdb\xb1" },
> + { 12, "\xdb\xb1\xdb\xb2" },
> + { 13, "\xdb\xb1\xdb\xb3" },
> + { 14, "\xdb\xb1\xdb\xb4" },
> + { 15, "\xdb\xb1\xdb\xb5" },
> + { 16, "\xdb\xb1\xdb\xb6" },
> + { 17, "\xdb\xb1\xdb\xb7" },
> + { 18, "\xdb\xb1\xdb\xb8" },
> + { 19, "\xdb\xb1\xdb\xb9" },
> + { 20, "\xdb\xb2\xdb\xb0" },
> + { 30, "\xdb\xb3\xdb\xb0" },
> + { 40, "\xdb\xb4\xdb\xb0" },
> + { 50, "\xdb\xb5\xdb\xb0" },
> + { 60, "\xdb\xb6\xdb\xb0" },
> + { 70, "\xdb\xb7\xdb\xb0" },
> + { 80, "\xdb\xb8\xdb\xb0" },
> + { 90, "\xdb\xb9\xdb\xb0" },
> + { 100, "\xdb\xb1\xdb\xb0\xdb\xb0" },
> + { 1000, "\xdb\xb1\xdb\xb0\xdb\xb0\xdb\xb0" },
> +};
> +
> +static int
> +do_test (void)
> +{
> + xsetlocale (LC_ALL, "fa_IR.UTF-8");
> +
> + for (int i = 0; i < array_length (inputs); i++)
> + {
> + int n;
> + sscanf (inputs[i].str, "%Id", &n);
> + TEST_COMPARE (n, inputs[i].n);
> + }
> +
> + return 0;
> +}
> +
> +#include <support/test-driver.c>
> diff --git a/stdio-common/vfscanf-internal.c b/stdio-common/vfscanf-internal.c
> index bfb9baa21a..ba4b289de6 100644
> --- a/stdio-common/vfscanf-internal.c
> +++ b/stdio-common/vfscanf-internal.c
> @@ -1455,13 +1455,14 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> int from_level;
> int to_level;
> int level;
> + enum { num_digits_len = 10 };
> #ifdef COMPILE_WSCANF
> - const wchar_t *wcdigits[10];
> - const wchar_t *wcdigits_extended[10];
> + const wchar_t *wcdigits[num_digits_len];
> #else
> - const char *mbdigits[10];
> - const char *mbdigits_extended[10];
> + const char *mbdigits[num_digits_len];
> #endif
> + CHAR_T *digits_extended[num_digits_len] = { NULL };
> +
> /* "to_inpunct" is a map from ASCII digits to their
> equivalent in locale. This is defined for locales
> which use an extra digits set. */
> @@ -1482,18 +1483,18 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> /* Adding new level for extra digits set in locale file. */
> ++to_level;
>
> - for (n = 0; n < 10; ++n)
> + for (n = 0; n < num_digits_len; ++n)
> {
> #ifdef COMPILE_WSCANF
> wcdigits[n] = (const wchar_t *)
> _NL_CURRENT (LC_CTYPE, _NL_CTYPE_INDIGITS0_WC + n);
>
> wchar_t *wc_extended = (wchar_t *)
> - alloca ((to_level + 2) * sizeof (wchar_t));
> + malloc ((to_level + 2) * sizeof (wchar_t));
> __wmemcpy (wc_extended, wcdigits[n], to_level);
> wc_extended[to_level] = __towctrans (L'0' + n, map);
> wc_extended[to_level + 1] = '\0';
> - wcdigits_extended[n] = wc_extended;
> + digits_extended[n] = wc_extended;
> #else
> mbdigits[n]
> = curctype->values[_NL_CTYPE_INDIGITS0_MB + n].string;
> @@ -1524,14 +1525,13 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> size_t mbdigits_len = last_char - mbdigits[n];
>
> /* Allocate memory for extended multibyte digit. */
> - char *mb_extended;
> - mb_extended = (char *) alloca (mbdigits_len + mblen + 1);
> + char *mb_extended = malloc (mbdigits_len + mblen + 1);
>
> /* And get the mbdigits + extra_digit string. */
> *(char *) __mempcpy (__mempcpy (mb_extended, mbdigits[n],
> mbdigits_len),
> extra_mbdigit, mblen) = '\0';
> - mbdigits_extended[n] = mb_extended;
> + digits_extended[n] = mb_extended;
> #endif
> }
> }
> @@ -1541,7 +1541,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> {
> /* In this round we get the pointer to the digit strings
> and also perform the first round of comparisons. */
> - for (n = 0; n < 10; ++n)
> + for (n = 0; n < num_digits_len; ++n)
> {
> /* Get the string for the digits with value N. */
> #ifdef COMPILE_WSCANF
> @@ -1553,7 +1553,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> DIAG_IGNORE_NEEDS_COMMENT (4.7, "-Wmaybe-uninitialized");
>
> if (__glibc_unlikely (map != NULL))
> - wcdigits[n] = wcdigits_extended[n];
> + wcdigits[n] = digits_extended[n];
> else
> wcdigits[n] = (const wchar_t *)
> _NL_CURRENT (LC_CTYPE, _NL_CTYPE_INDIGITS0_WC + n);
> @@ -1574,7 +1574,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> int avail = width > 0 ? width : INT_MAX;
>
> if (__glibc_unlikely (map != NULL))
> - mbdigits[n] = mbdigits_extended[n];
> + mbdigits[n] = digits_extended[n];
> else
> mbdigits[n]
> = curctype->values[_NL_CTYPE_INDIGITS0_MB + n].string;
> @@ -1617,13 +1617,13 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> #endif
> }
>
> - if (n == 10)
> + if (n == num_digits_len)
> {
> /* Have not yet found the digit. */
> for (level = from_level + 1; level <= to_level; ++level)
> {
> /* Search all ten digits of this level. */
> - for (n = 0; n < 10; ++n)
> + for (n = 0; n < num_digits_len; ++n)
> {
> #ifdef COMPILE_WSCANF
> if (c == (wint_t) *wcdigits[n])
> @@ -1679,7 +1679,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> }
> }
>
> - if (n < 10)
> + if (n < num_digits_len)
> c = L_('0') + n;
> else if (flags & GROUP)
> {
> @@ -1708,7 +1708,7 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
> {
> __set_errno (ENOMEM);
> done = EOF;
> - goto errout;
> + break;
> }
>
> if (*cmpp != '\0')
> @@ -1742,6 +1742,12 @@ __vfscanf_internal (FILE *s, const char *format, va_list argptr,
>
> c = inchar ();
> }
> +
> + for (n = 0; n < num_digits_len; n++)
> + free (digits_extended[n]);
> +
> + if (done == EOF)
> + goto errout;
> }
> else
> /* Read the number into workspace. */
> diff --git a/wcsmbs/Makefile b/wcsmbs/Makefile
> index 22192985e1..ed2660a524 100644
> --- a/wcsmbs/Makefile
> +++ b/wcsmbs/Makefile
> @@ -175,6 +175,7 @@ tests := \
> tst-wscanf-binary-c2x \
> tst-wscanf-binary-gnu11 \
> tst-wscanf-binary-gnu89 \
> + tst-wscanf-to_inpunct \
> wcsatcliff \
> wcsmbs-tst1 \
> # tests
> @@ -186,6 +187,7 @@ LOCALES := \
> de_DE.ISO-8859-1 \
> de_DE.UTF-8 \
> en_US.ANSI_X3.4-1968 \
> + fa_IR.UTF-8 \
> hr_HR.ISO-8859-2 \
> ja_JP.EUC-JP \
> tr_TR.ISO-8859-9 \
> @@ -207,6 +209,7 @@ $(objpfx)tst-c16-surrogate.out: $(gen-locales)
> $(objpfx)tst-c32-state.out: $(gen-locales)
> $(objpfx)test-c8rtomb.out: $(gen-locales)
> $(objpfx)test-mbrtoc8.out: $(gen-locales)
> +$(objpfx)tst-wscanf-to_inpunct.out: $(gen-locales)
> endif
>
> $(objpfx)tst-wcstod-round: $(libm)
> diff --git a/wcsmbs/tst-wscanf-to_inpunct.c b/wcsmbs/tst-wscanf-to_inpunct.c
> new file mode 100644
> index 0000000000..72f2a1a422
> --- /dev/null
> +++ b/wcsmbs/tst-wscanf-to_inpunct.c
> @@ -0,0 +1,79 @@
> +/* Test scanf for languages with mapping pairs of alternate digits and
> + separators.
> + Copyright (C) 2023 Free Software Foundation, Inc.
> + This file is part of the GNU C Library.
> +
> + The GNU C Library is free software; you can redistribute it and/or
> + modify it under the terms of the GNU Lesser General Public
> + License as published by the Free Software Foundation; either
> + version 2.1 of the License, or (at your option) any later version.
> +
> + The GNU C Library is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> + Lesser General Public License for more details.
> +
> + You should have received a copy of the GNU Lesser General Public
> + License along with the GNU C Library; if not, see
> + <https://www.gnu.org/licenses/>. */
> +
> +#include <array_length.h>
> +#include <stdio.h>
> +#include <support/support.h>
> +#include <support/check.h>
> +#include <wchar.h>
> +
> +/* fa_IR defines to_inpunct for numbers. */
> +static const struct input_t
> +{
> + int n;
> + const wchar_t str[5];
> +} inputs[] =
> +{
> + { 1, { 0x000006f1, L'\0' } },
> + { 2, { 0x000006f2, L'\0' } },
> + { 3, { 0x000006f3, L'\0' } },
> + { 4, { 0x000006f4, L'\0' } },
> + { 5, { 0x000006f5, L'\0' } },
> + { 6, { 0x000006f6, L'\0' } },
> + { 7, { 0x000006f7, L'\0' } },
> + { 8, { 0x000006f8, L'\0' } },
> + { 9, { 0x000006f9, L'\0' } },
> + { 10, { 0x000006f1, 0x000006f0, L'\0' } },
> + { 11, { 0x000006f1, 0x000006f1, L'\0' } },
> + { 12, { 0x000006f1, 0x000006f2, L'\0' } },
> + { 13, { 0x000006f1, 0x000006f3, L'\0' } },
> + { 14, { 0x000006f1, 0x000006f4, L'\0' } },
> + { 15, { 0x000006f1, 0x000006f5, L'\0' } },
> + { 16, { 0x000006f1, 0x000006f6, L'\0' } },
> + { 17, { 0x000006f1, 0x000006f7, L'\0' } },
> + { 18, { 0x000006f1, 0x000006f8, L'\0' } },
> + { 19, { 0x000006f1, 0x000006f9, L'\0' } },
> + { 20, { 0x000006f2, 0x000006f0, L'\0' } },
> + { 30, { 0x000006f3, 0x000006f0, L'\0' } },
> + { 40, { 0x000006f4, 0x000006f0, L'\0' } },
> + { 50, { 0x000006f5, 0x000006f0, L'\0' } },
> + { 60, { 0x000006f6, 0x000006f0, L'\0' } },
> + { 70, { 0x000006f7, 0x000006f0, L'\0' } },
> + { 80, { 0x000006f8, 0x000006f0, L'\0' } },
> + { 90, { 0x000006f9, 0x000006f0, L'\0' } },
> + { 100, { 0x000006f1, 0x000006f0, 0x000006f0, L'\0' } },
> + { 1000, { 0x000006f1, 0x000006f0, 0x000006f0, 0x000006f0, L'\0' } },
> +};
> +
> +static int
> +do_test (void)
> +{
> + xsetlocale (LC_ALL, "fa_IR.UTF-8");
> +
> + for (int i = 0; i < array_length (inputs); i++)
> + {
> + int n;
> + swscanf (inputs[i].str, L"%Id", &n);
> + TEST_COMPARE (n, inputs[i].n);
> + }
> +
> + return 0;
> +}
> +
> +#include <support/test-driver.c>
next prev parent reply other threads:[~2023-07-03 14:39 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-03 14:38 Adhemerval Zanella
2023-07-03 14:39 ` Adhemerval Zanella Netto [this message]
2023-07-03 18:18 ` Joe Simmons-Talbott
2023-07-03 18:52 ` Adhemerval Zanella Netto
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=3baffc46-3a2f-85ca-f5a4-b1857d7c44a7@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=josimmon@redhat.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).