Nikolay Borisov [Mon, 13 May 2019 12:44:18 +0000 (15:44 +0300)]
libblkid: Set BLKID_BID_FL_VERIFIED in case revalidation is not needed
If blkid_verify deems that device revalidation is not needed since BLKID_PROBE_MIN
seconds haven't elapsed since the last revalidation of the device it returns
a blkid_dev. However, if this device has been read from the cache file on disk then
it wont' have BLKID_BID_FL_VERIFIED bit set. This in turn could lead to a later
call to blkid_dev_get free this device in case its part of a multi device
configuration. This is particularly relevant to btrfs raid configurations.
Namely this was exhibited by btrfs-progs which use blkid for device enumeration.
In case of a multi-device filesystem (i.e raid10) running xfstest btrfs/011
would fail sporadically since cached devices read from cache were not revalidated
due to being recently validated (according to timestamp in the cache file) at
the same time their in-memory blkid_dev structures didn't have BLKID_BID_FL_VERIFIED
set. This lead to their untimely removal from the cache from blkid_get_dev.
Fix this by setting the BLKID_BID_FL_VERIFIED when returning from blkid_verify
with a valid device irrespective of whether full revalidation was performed or
the device is deemed valid due to being recently validated.
[kzak@redhat.com: - remove also check for BLKID_BID_FL_VERIFIED, the
function needs to check for elapsed time only]
Signed-off-by: Nikolay Borisov <nborisov@suse.com> Signed-off-by: Karel Zak <kzak@redhat.com>
Nikolay Borisov [Mon, 13 May 2019 12:44:17 +0000 (15:44 +0300)]
libblkid: Don't check BLKID_PROBE_INTERVAL in blkid_verify
That constant is set to 200 seconds and is already check in probe_all().
It essentially controls how often blkid_probe_all can do a full cache
revalidation. Since blkid_verify is called from within probe_all() iff at least
BLKID_PROBE_INTERVAL seconds have elapsed it makes no sense to check this value
in blkid_verify.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Sami Kerola [Tue, 7 May 2019 19:18:08 +0000 (20:18 +0100)]
fstrim: add --quiet option to suppress error messages
When fstrim interacts with NTFS it result can be error reporting bad file
descriptor. That seems to be a bug in NTFS. While waiting driver to get on
top of the issue and be commonly available lets add to fstrim option to make
it be more silent about errno 9 aka EBADF, Bad file descriptor.
Reported-by: https://github.com/moviuro Proposed-by: Dave Reisner <dreisner@archlinux.org>
Reference: https://bugs.archlinux.org/task/62288
Addresses: https://github.com/karelzak/util-linux/issues/789 Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Romain Izard [Fri, 3 May 2019 13:42:22 +0000 (15:42 +0200)]
lib/loopdev.c: Retry LOOP_SET_STATUS64 on EAGAIN
A recent bugfix in the Linux kernel made it possible for the
LOOP_SET_STATUS64 ioctl to fail when called with a non-zero offset,
with an EAGAIN errno:
5db470e229e2 loop: drop caches if offset or block_size are changed
This fix changes a silent failure (where mount could sometimes access
the backing loop image through the cache without the specified offset)
to an explicit failure, and it has also been backported on stable
branches.
On a 5.0 kernel, other changes to the loop driver make it hard to get
generate the EAGAIN error, but this bugfix has also been backported to
stables branches, without these changes. At least with the 4.14 stable
branch, the EAGAIN error can be quickly generated with the following loop:
while mount -o loop,offset=239 disk point && umount point; do :; done
Retry the ioctl when it fails with EAGAIN, which means that mount or
losetup will eventually succeed when encountering this case.
[kzak@redhat.com: - use our local portable xusleep()]
Signed-off-by: Romain Izard <romain.izard.pro@gmail.com> Signed-off-by: Karel Zak <kzak@redhat.com>
misc-utils/hardlink.c: In function ‘process_path’:
misc-utils/hardlink.c:287:30: warning: operand of ?: changes signedness from ‘off_t’ {aka ‘long int’} to ‘long unsigned int’ due to unsignedness of other operand [-Wsign-compare]
misc-utils/hardlink.c: In function ‘main’:
misc-utils/hardlink.c:455:5: warning: ‘exclude_pattern’ may be used uninitialized in this function [-Wmaybe-uninitialized]
Sami Kerola [Wed, 1 May 2019 19:04:24 +0000 (20:04 +0100)]
hardlink: retire NIOBUF in favour of more common BUFSIZ
Reason to retire NIOBUF is that it is obscure local definition, while BUFSIZ
is well understood and commonly used constant. Besized sizes of these are
not far off, the NIOBUF was 4096 bytes and BUFSIZ tends to be 8192 bytes.
Proposed-by: Karel Zak <kzak@redhat.com>
Reference: https://github.com/karelzak/util-linux/pull/783 Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Karel Zak [Wed, 24 Apr 2019 16:31:08 +0000 (18:31 +0200)]
hardlink: remove \r from output
* remove \r from internationalized messages
* remove \r from all output to make it easy to use (see for example
output file from "hardlink -vv --dry-run . &> log")
* remove unnecessary formatting stuff from output, just keep is simple
and stupid...
Karel Zak [Wed, 24 Apr 2019 10:24:15 +0000 (12:24 +0200)]
tests: auto-enable ASAN option if necessary
Let's detect ASAN LDFLAGS in top level Makefile to make sure we call
tests with --memcheck-asan if build-system has been configured with
--enable-asan.
Stanislav Brabec [Wed, 24 Apr 2019 09:16:53 +0000 (11:16 +0200)]
lslogins: Fix discrepancies of SYS_UID_MIN
util-linux does not contain useradd. Its most popular implementation
comes from shadow. SYS_UID_MIN is one of common parameters. Its
hardcoded fallback value is equal to 101 in shadow useradd (see
shadow-4.6/libmisc/find_new_uid.c: get_ranges()), but 201 in
login-utils/lslogins.c.
Let lslogins use the same fallback as useradd from shadow.
Hopefully most distros define its custom value of SYS_UID_MIN in
/etc/login.defs, so this problem is not visible.
login-utils/lslogins.1 does not mention its default at all. Add a
reference and improve text of lslogins(1) to prevent off-by-one
interpretation.
Signed-off-by: Stanislav Brabec <sbrabec@suse.cz> Signed-off-by: Karel Zak <kzak@redhat.com>
Karel Zak [Tue, 16 Apr 2019 11:42:34 +0000 (13:42 +0200)]
include/closestream: add close_stdout_atexit()
It seems better to have a way to control when atexit(close_stdout()) is
used, because close stdout means that for example ASAN (or another
into binary integrated tool) is not able to print the final summary.
Wang Shilong [Sun, 14 Apr 2019 03:31:56 +0000 (12:31 +0900)]
fstrim: get realpath before trim ioctl
Original motivation is we want to run fstrim command
on Lustre[1] osd server mount point directly, however
our server mount point doesn't export osd directory
to users, and it will cause following command fail:
$fstrim -v /mnt/mds/
But following succeed:
$fstrim -v /mnt/mds
We could improve this a bit by getting realpath
before trapping kernel, this also give benifits
to normal use cases.
Cc: Andreas Dilger <adilger@dilger.ca> Cc: Shuichi Ihara <sihara@ddn.com>
[1] http://wiki.lustre.org/Main_Page Signed-off-by: Wang Shilong <wshilong@ddn.com>
Karel Zak [Mon, 25 Mar 2019 15:39:43 +0000 (16:39 +0100)]
su: change error message
We use PAM and if pam_strerror() returns nothing we have no clue why
authentication failed. It's mistake to blame incorrect password if
there are many possible reason...
Addresses: https://github.com/karelzak/util-linux/issues/778 Signed-off-by: Karel Zak <kzak@redhat.com>
Karel Zak [Wed, 20 Mar 2019 12:12:25 +0000 (13:12 +0100)]
taskset: fix cpuset list parser
taskset hangs when executed with badly formatted cpuset list, for
example:
$ taskset -c 0--1 true
The current cpuset list parser is pretty weak as based on scanf()
without strings verification ("-1" as input for "%u" returns
unexpected number). It seems faster and better to use strtoul() and
isdigit().
Addresses: https://github.com/karelzak/util-linux/issues/77 Signed-off-by: Karel Zak <kzak@redhat.com>
Karel Zak [Tue, 19 Mar 2019 10:09:02 +0000 (11:09 +0100)]
lscpu: report more usable cache sizes
The current version reports cache size as reported by /sys, it means
real size of the one piece of the cache. This way provides minimal
overview about all system as the cache is often shared between CPUs.
It seems better to report all size of the caches in the summary
output. It also does not make sense to report sizes per core (or
socket) as CPU topology may be pretty complicated.
The final solution (not implemented yet) will be to have --list-caches
where we can report all details like all-size, item-size, per-core size,
ways, type, etc.
Addresses: https://github.com/karelzak/util-linux/issues/663 Signed-off-by: Karel Zak <kzak@redhat.com>
projects / thirdparty / util-linux.git / log
