| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
==10918==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7fffd795b680 at pc 0x0000004447c6 bp 0x7fffd795b3e0 sp 0x7fffd795ab78
WRITE of size 129 at 0x7fffd795b680 thread T0
#0 0x4447c5 in scanf_common(void*, int, bool, char const*, __va_list_tag*) (/home/src/util-linux/.libs/lt-mount+0x4447c5)
#1 0x445892 in sscanf (/home/src/util-linux/.libs/lt-mount+0x445892)
#2 0x7fe78709a3d3 in get_filesystems /home/src/util-linux/libmount/src/utils.c:581:7
#3 0x7fe78709a1ba in mnt_get_filesystems /home/src/util-linux/libmount/src/utils.c:622:7
#4 0x7fe7870aa78f in do_mount_by_pattern /home/src/util-linux/libmount/src/context_mount.c:833:7
#5 0x7fe7870a9534 in mnt_context_do_mount /home/src/util-linux/libmount/src/context_mount.c:951:9
#6 0x7fe7870aab2b in mnt_context_mount /home/src/util-linux/libmount/src/context_mount.c:1051:8
#7 0x4ba9f5 in main /home/src/util-linux/sys-utils/mount.c:1107:7
#8 0x7fe785caa03f in __libc_start_main (/usr/lib/libc.so.6+0x2003f)
#9 0x4b9f9c in _start (/home/src/util-linux/.libs/lt-mount+0x4b9f9c)
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current command line parser will stop at the first occurrence of an
option, however the kernel does the opposite. So if you have:
root=/dev/sda1 root=/dev/sda2
When you look for "root", the kernel will use /dev/sda2, but util-linux
uses /dev/sda1.
Further, if args are passed to custom init programs, the parser will
pick those up as kernel options. So if you have:
root=/dev/sda1 -- /foo bar=yes
The kernel will stop at the "--" and pass the rest to userland. But if
you look for "bar", util-linux will incorrectly return "yes".
Ultimately, there's no way for util-linux to exactly parse the command
line the same way as the kernel -- we don't know exactly which ones the
kernel picks up and which it passes on to userland (either as env vars
or as command line args). The kernel passes all unrecognized options.
These updates are simple best effort.
URL: https://bugs.gentoo.org/526754
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
|
|
|
|
|
| |
This issue affects util-linux portability to GNU/HURD
Reported-by: Pino Toscano (from Red Hat)
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
| |
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
| |
* libmount/src/utils.c (BTRFS_TEST_MAGIC): Conditionally add define
which is used since commit v2.24-243-g6a52473.
Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The umount(8) always parses /proc/self/mountinfo to get fstype and to
merge kernel mount options with userspace mount options from
/run/mount/utab. This behavior is overkill in many cases and it's
pretty expensive as kernel has to always compose *whole* mountinfo.
This performance disadvantage is visible for crazy use-cases with huge
number of mountpoints and frequently called umount(8).
It seems that we can bypass /proc/self/mountinfo by statfs() to get
filesystem type (statfs.f_type magic) and analyze /run/mount/utab
before we parse mountinfo.
This optimization is not used when:
* umount(8) executed by non-root (as user= in utab is expected)
* umount --lazy / --force (target is probably unreachable NFS, then
use statfs() is pretty bad idea)
* target is not a directory (e.g. umount /dev/sda1)
* there is (deprecated) writeable mtab
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
| |
After mkostemp() failed, umask() and free() might alter the errno
to another value. Not sure those calls really changes the errno
or not. But let's be more conservative.
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When unmounting some mountpoints as an unprivileged user (via the
'user' option in fstab), the umount fails. Debug output of 'umount
/opt' reveals:
17760: libmount: CXT: [0x22890e0]: do umount
17760: libmount: UTILS: moving to /opt parent
17760: libmount: CXT: current directory moved to / [last_component='opt']
17760: libmount: CXT: [0x22890e0]: umount(2) [target='pt', flags=0x00000000]
valgrind shows the problem:
==23544== Source and destination overlap in memcpy(0x58d1370, 0x58d1371, 4)
==23544== at 0x4C2BBC3: memcpy@@GLIBC_2.14 (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==23544== by 0x4E537C3: mnt_chdir_to_parent (utils.c:168)
==23544== by 0x4E45E4C: mnt_context_do_umount (context_umount.c:601)
==23544== by 0x4E46513: mnt_context_umount (context_umount.c:855)
==23544== by 0x403476: umount_one (umount.c:299)
==23544== by 0x402B34: main (umount.c:629)
==23544==
ref: https://bugs.archlinux.org/task/36968
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
|
|
|
|
|
|
|
|
| |
Avoid code dublication in libmount and time-util.
Proposed-by: Karel Zak <kzak@redhat.com>
Reference: http://markmail.org/message/h7zexvqsieqngtmx
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
| |
Signed-off-by: Ondrej Oprala <ooprala@redhat.com>
|
|
|
|
|
| |
Reported-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
| |
# mount DUMMY=filename.img /mnt
The 'DUMMY=filename.img' is a filename and should not be
interpreted as tag name. The valid tag names are LABEL, UUID,
PARTLABEL and PARTUUID only.
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We use
mnt_optstr_append_option(&o, mnt_fs_get_vfs_options(fs), NULL);
in mount.nfs, unfortunately mnt_optstr_append_option() has been marked
ass nonnull(1, 2). That's wrong because append and prepend should
robust enough to accept NULL as option name.
The patch also removes almost all __attribute__((nonnull). It seems
better to rely on assert() to have usable feedback. In many cases
(nonnull) is premature optimization for the library. This attribute
makes sense for things like strlen() or so...
Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=948274
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
| |
This is more robust implementation of mnt_get_mountpoint() that does
not ignore bind mountpoints (mount --bind /mnt /mnt) as it does not
depend on st_dev numbers.
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
| |
- ignore empty files
- ignore empty tables
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
| |
Reported-by: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The above function infloops when the name to search for can only
be found at the beginning of /proc/cmdline but doesn't match,
e.g. when searching for "ro" in "root=/dev/sda1 quiet vga=0x31a".
* libmount/src/utils.c (mnt_get_kernel_cmdline_option): Replace
while by for loop to ensure the pointer p is incremented.
Signed-off-by: Bernhard Voelker <mail@bernhard-voelker.de>
|
|
|
|
|
|
|
|
|
|
| |
- use __attribute__((nonnull) for functions where we not able to
return an return code ("is", "has" and some "get" functions).
- use __attribute__((nonnull) for small functions where we
always modify any of the function argument (some mnt_optstr_* functions)
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
| |
Signed-off-by: Ondrej Oprala <ooprala@redhat.com>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
| |
- single line with '*' in /etc/filesystems means that libmount has to
read /proc/filesystems, otherwise /proc/filesystems has to be ignored
- mount(2) ENODEV is no reason to break the do_mount_by_pattern()
loop when trying to mount by /{etc,proc}/filesystems
Reported-by: NeilBrown <neilb@suse.de>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
| |
Reported-by: Juergen Daubert <jue@jue.li>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
| |
It's the responsibility of anyone adding to this list in the future to
ensure that the list remains sorted.
While we're at it, expand this list of known pseudofs types.
Signed-off-by: Dave Reisner <dreisner@archlinux.org>
|
| |
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Examples:
* findmnt --fstab -O noauto
Returns all entries where is not "auto" option (including entries
with "noauto" option.
* findmnt --fstab -O +noauto
Returns all entries where is literally "noauto" option.
Addresses: https://bugzilla.redhat.com/show_bug.cgi?id=758457
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|