glibc/libio/tst-freopen.c

113 lines
2.7 KiB
C
Raw Normal View History

/* Test freopen with mmap stdio.
Copyright (C) 2002-2018 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Jakub Jelinek <jakub@redhat.com>, 2002.
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
<http://www.gnu.org/licenses/>. */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
#include <support/check.h>
#include <support/temp_file.h>
static int fd;
static char *name;
static void
do_prepare (int argc, char *argv[])
{
fd = create_temp_file ("tst-freopen.", &name);
TEST_VERIFY_EXIT (fd != -1);
}
#define PREPARE do_prepare
/* Basic tests for freopen. */
static void
do_test_basic (void)
{
const char * const test = "Let's test freopen.\n";
char temp[strlen (test) + 1];
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FILE *f = fdopen (fd, "w");
if (f == NULL)
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("fdopen: %m");
fputs (test, f);
fclose (f);
f = fopen (name, "r");
if (f == NULL)
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("fopen: %m");
if (fread (temp, 1, strlen (test), f) != strlen (test))
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("fread: %m");
temp [strlen (test)] = '\0';
if (strcmp (test, temp))
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("read different string than was written: (%s, %s)",
test, temp);
f = freopen (name, "r+", f);
if (f == NULL)
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("freopen: %m");
if (fseek (f, 0, SEEK_SET) != 0)
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("fseek: %m");
if (fread (temp, 1, strlen (test), f) != strlen (test))
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("fread: %m");
temp [strlen (test)] = '\0';
if (strcmp (test, temp))
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
FAIL_EXIT1 ("read different string than was written: (%s, %s)",
test, temp);
fclose (f);
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
}
/* Test for BZ#21398, where it tries to freopen stdio after the close
of its file descriptor. */
static void
do_test_bz21398 (void)
{
(void) close (STDIN_FILENO);
FILE *f = freopen (name, "r", stdin);
if (f == NULL)
FAIL_EXIT1 ("freopen: %m");
TEST_VERIFY_EXIT (ferror (f) == 0);
char buf[128];
char *ret = fgets (buf, sizeof (buf), stdin);
TEST_VERIFY_EXIT (ret != NULL);
TEST_VERIFY_EXIT (ferror (f) == 0);
}
static int
do_test (void)
{
do_test_basic ();
do_test_bz21398 ();
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
return 0;
}
Modify several tests to use test-skeleton.c This patch modifies several test cases to use test-skeleton.c. It was generated by a bash script written for this purpose and thus excludes several other tests which I deemed worth a visual inspection before making the change. I intend to follow up with individual patches to the tests skipped by the script. The script itself resides at http://git.io/WODAmg and should reproduce this very patch when run against master. ChangeLog: 2014-10-30 Arjun Shankar <arjun.is@lostca.se> * catgets/test-gencat.c: Use test-skeleton.c. * catgets/tst-catgets.c: Likewise. * csu/tst-empty.c: Likewise. * elf/tst-audit2.c: Likewise. * elf/tst-global1.c: Likewise. * elf/tst-pathopt.c: Likewise. * elf/tst-piemod1.c: Likewise. * elf/tst-tls10.c: Likewise. * elf/tst-tls11.c: Likewise. * elf/tst-tls12.c: Likewise. * gnulib/tst-gcc.c: Likewise. * iconvdata/tst-e2big.c: Likewise. * iconvdata/tst-loading.c: Likewise. * iconv/tst-iconv1.c: Likewise. * iconv/tst-iconv2.c: Likewise. * inet/test-inet6_opt.c: Likewise. * inet/tst-gethnm.c: Likewise. * inet/tst-network.c: Likewise. * inet/tst-ntoa.c: Likewise. * intl/tst-codeset.c: Likewise. * intl/tst-gettext2.c: Likewise. * intl/tst-gettext3.c: Likewise. * intl/tst-ngettext.c: Likewise. * intl/tst-translit.c: Likewise. * io/test-stat.c: Likewise. * libio/test-fmemopen.c: Likewise. * libio/tst-freopen.c: Likewise. * libio/tst-sscanf.c: Likewise. * libio/tst-ungetwc1.c: Likewise. * libio/tst-ungetwc2.c: Likewise. * libio/tst-widetext.c: Likewise. * localedata/tst-ctype.c: Likewise. * localedata/tst-digits.c: Likewise. * localedata/tst-leaks.c: Likewise. * localedata/tst-mbswcs1.c: Likewise. * localedata/tst-mbswcs2.c: Likewise. * localedata/tst-mbswcs3.c: Likewise. * localedata/tst-mbswcs4.c: Likewise. * localedata/tst-mbswcs5.c: Likewise. * localedata/tst-setlocale.c: Likewise. * localedata/tst-trans.c: Likewise. * localedata/tst-wctype.c: Likewise. * localedata/tst-xlocale1.c: Likewise. * login/tst-grantpt.c: Likewise. * malloc/tst-calloc.c: Likewise. * malloc/tst-malloc.c: Likewise. * malloc/tst-mallocstate.c: Likewise. * malloc/tst-mcheck.c: Likewise. * malloc/tst-mtrace.c: Likewise. * malloc/tst-obstack.c: Likewise. * math/atest-exp2.c: Likewise. * math/atest-exp.c: Likewise. * math/atest-sincos.c: Likewise. * math/test-matherr.c: Likewise. * math/test-misc.c: Likewise. * math/test-powl.c: Likewise. * math/tst-definitions.c: Likewise. * misc/tst-dirname.c: Likewise. * misc/tst-efgcvt.c: Likewise. * misc/tst-fdset.c: Likewise. * misc/tst-hsearch.c: Likewise. * misc/tst-mntent2.c: Likewise. * nptl/tst-sem7.c: Likewise. * nptl/tst-sem8.c: Likewise. * nptl/tst-sem9.c: Likewise. * nss/test-netdb.c: Likewise. * posix/tst-fnmatch.c: Likewise. * posix/tst-getlogin.c: Likewise. * posix/tst-gnuglob.c: Likewise. * posix/tst-mmap.c: Likewise. * pwd/tst-getpw.c: Likewise. * resolv/tst-inet_ntop.c: Likewise. * rt/tst-timer.c: Likewise. * stdio-common/test-fseek.c: Likewise. * stdio-common/test-popen.c: Likewise. * stdio-common/test-vfprintf.c: Likewise. * stdio-common/tst-cookie.c: Likewise. * stdio-common/tst-fileno.c: Likewise. * stdio-common/tst-gets.c: Likewise. * stdio-common/tst-obprintf.c: Likewise. * stdio-common/tst-perror.c: Likewise. * stdio-common/tst-sprintf2.c: Likewise. * stdio-common/tst-sprintf3.c: Likewise. * stdio-common/tst-sprintf.c: Likewise. * stdio-common/tst-swprintf.c: Likewise. * stdio-common/tst-tmpnam.c: Likewise. * stdio-common/tst-unbputc.c: Likewise. * stdio-common/tst-wc-printf.c: Likewise. * stdlib/tst-environ.c: Likewise. * stdlib/tst-fmtmsg.c: Likewise. * stdlib/tst-limits.c: Likewise. * stdlib/tst-rand48-2.c: Likewise. * stdlib/tst-rand48.c: Likewise. * stdlib/tst-random2.c: Likewise. * stdlib/tst-random.c: Likewise. * stdlib/tst-strtol.c: Likewise. * stdlib/tst-strtoll.c: Likewise. * stdlib/tst-tls-atexit.c: Likewise. * stdlib/tst-xpg-basename.c: Likewise. * string/test-ffs.c: Likewise. * string/tst-bswap.c: Likewise. * string/tst-inlcall.c: Likewise. * string/tst-strtok.c: Likewise. * string/tst-strxfrm.c: Likewise. * sysdeps/x86_64/tst-audit10.c: Likewise. * sysdeps/x86_64/tst-audit3.c: Likewise. * sysdeps/x86_64/tst-audit4.c: Likewise. * sysdeps/x86_64/tst-audit5.c: Likewise. * time/tst-ftime_l.c: Likewise. * time/tst-getdate.c: Likewise. * time/tst-mktime3.c: Likewise. * time/tst-mktime.c: Likewise. * time/tst-posixtz.c: Likewise. * time/tst-strptime2.c: Likewise. * time/tst-strptime3.c: Likewise. * wcsmbs/tst-btowc.c: Likewise. * wcsmbs/tst-mbrtowc.c: Likewise. * wcsmbs/tst-mbsrtowcs.c: Likewise. * wcsmbs/tst-wchar-h.c: Likewise. * wcsmbs/tst-wcpncpy.c: Likewise. * wcsmbs/tst-wcrtomb.c: Likewise. * wcsmbs/tst-wcsnlen.c: Likewise. * wcsmbs/tst-wcstof.c: Likewise.
2014-11-05 10:54:08 +01:00
libio: Avoid dup already opened file descriptor [BZ#21393] As described in BZ#21398 (close as dup of 21393) report current freopen implementation fails when one tries to freopen STDIN_FILENO, STDOUT_FILENO, or STDERR_FILENO. Although on bug report the discussion leads to argue if a close followed by a freopen on the standard file is a valid operation, the underlying issue is not really the check for dup3 returned value, but rather calling it if the returned file descriptor is equal as the input one. So for a quality of implementation this patch avoid calling dup3 for the aforementioned case. It also adds a dup3 error case check for the two possible failures, with one being Linux only: EINTR and EBUSY. The EBUSY issue is better explained on this stackoverflow thread [1], but in a short it is due the internal Linux implementation which allows a race condition window for dup2 due the logic dissociation of file descriptor allocation and actual VFS 'install' operation. For both outliers failures all allocated memory is freed and a NULL FILE* is returned. With this patch the example on BZ#21398 is now actually possible (I used as the testcase for the bug report). Checked on x86_64-linux-gnu. [BZ #21393] * libio/freopen.c (freopen): Avoid dup already opened file descriptor and add a check for dup3 failure. * libio/freopen64.c (freopen64): Likewise. * libio/tst-freopen.c (do_test): Rename to do_test_basic and use libsupport. (do_test_bz21398): New test. * manual/stdio.texi (freopen): Add documentation of EBUSY failure. [1] http://stackoverflow.com/questions/23440216/race-condition-when-using-dup2
2017-05-05 16:31:38 +02:00
#include <support/test-driver.c>