Thomas Weißschuh [Sun, 21 Apr 2024 18:00:31 +0000 (20:00 +0200)]
findmnt: always zero-terminate SOURCES data
libsmartcols expects it's data fields to be zero terminated.
See the call to strlen() in scols_column_greatest_wrap().
ul_buffer however does not guarantee that termination,
ul_buffer_append_strings() discard the zero-termination.
Always zero-terminate in get_data_col_sources() and drop the now
unnecessary variable "i".
Closes: https://github.com/util-linux/util-linux/issues/2980 Signed-off-by: Thomas Weißschuh <thomas@t-8ch.de>
(cherry picked from commit 6594679e9af805075c282da35b1b2a8d767e403b)
Gbp-Pq: Topic upstream
Gbp-Pq: Name findmnt-always-zero-terminate-SOURCES-data.patch
lslocks: don't abort gathering per-process information even if opening a /proc/[0-9]* fails
If a process ($pid) taking a lock is gone while running lslocks,
gathering per-process information for the process may fail in opening
/proc/$pid. Though lslocks should work with incomplete information,
the original code stopped gathering per-process information for the
other processes.
As a result of the original behavior, tests/ts/lslocks/lslocks really
failed in an environment where multiple test cases ran simultaneously.
lsns: tolerate lsns_ioctl(fd, NS_GET_{PARENT,USERNS}) failing with ENOSYS
With the original code, "lsns/filedesc" test case failed on
"build (qemu-user, s390x)" and "build (qemu-user, riscv64)".
On the platforms, lsns_ioctl(fd, NS_GET_{PARENT,USERNS}) failed
with ENOSYS. The error stoped the iteration for gathering
information from /proc/[0-9]+. As a result, lsns printed
nothing. We don't expect this behavior.
3dd79293b5b655da9d913dedd8facb08959a7826 added -I option.
However, the code used bit flags (FL_DF and FL_DF_INODES) wrongly;
the code broke the output of -D option.
Commit 5d71d711d07a ("libblkid: topolicy/ioctl: use union for multiple data types")
incorrectly assumed that set_ulong and set_int refer to the type
returned by the kernel. Instead the different function pointer names
refer to the types of the function pointers.
However all ioctls, except for the later added BLKGETDISKSEQ, return
32bit integers.
This made libblkid also interpret the upper 32bits too, leading to
garbage values.
Introduce a new member 'kernel_size' to also handle the 64bit
BLKGETDISKSEQ.
Karel Zak [Mon, 1 Apr 2024 10:14:50 +0000 (12:14 +0200)]
libblkid: Fix segfault when blkid.conf doesn't exist
* Move 'line' and 'uevent' to the beginning of the LIBECONF code.
* Remove unwanted space between function name and arguments.
* Check for 'line' pointer before dereferencing.
References: https://github.com/util-linux/util-linux/pull/2883 Signed-off-by: Karel Zak <kzak@redhat.com>
(cherry picked from commit 7e357241b413a01c37b0b4d064bc0a47e3259361)
Gbp-Pq: Topic upstream
Gbp-Pq: Name libblkid-Fix-segfault-when-blkid.conf-doesn-t-exist.patch
Thomas Weißschuh [Sat, 13 Apr 2024 07:42:39 +0000 (09:42 +0200)]
Revert "lib/pager: Apply pager-specific fixes only when needed"
Less is not always detectable, as it can be hidden behind symlinks or
wrapper as for example in a default Debian installation.
Also testing for the literal string "less" does not match full path
specifications like "/usr/bin/less".
projects / util-linux.git / log