Sumanth Korikkar [Thu, 16 Oct 2025 15:38:04 +0000 (17:38 +0200)]
chmem: add chmem documentation for dynamic (de)configuration of memory
Describe chmem configure, deconfigure and memmap-on-memory options:
ff18dcb19aab ("s390/sclp: Add support for dynamic (de)configuration of
memory") s390 kernel no longer pre-adds all standby memory at boot.
Instead, users must explicitly configure a block before it can be used
for online/offline actions. At configuration time, users can dynamically
decide whether to use optional memmap-on-memory for each memory block,
where value of 1 allocates metadata (such as struct pages array) from
the hotplug memory itself, enabling hot-add operations even under memory
pressure. A value of 0 stores metadata in regular system memory, which
may require additional free memory, but enables continuous physical
memory across memory blocks.
Add documentation to reflect the following options:
* chmem --configure 128M --memmap-on-memory 1
* chmem --deconfigure 128M
* chmem --enable 128M # implicitly configure memory if supported by
architecture and online it
* chmem --disable 128M # offline memory and implicitly deconfigure if
supported by the architecture.
Just like online and offline actions, memory configuration and
deconfiguration can be controlled through similar options. Also,
memmap-on-memory setting can be changed, only when the memory block is
in deconfigured state. This means, it is usable only via --configure
option.
Reviewed-by: Maria Eisenhaendler <maria1@de.ibm.com> Signed-off-by: Sumanth Korikkar <sumanthk@linux.ibm.com>
Sumanth Korikkar [Thu, 16 Oct 2025 15:38:03 +0000 (17:38 +0200)]
chmem: add support for dynamic (de)configuration of hotplug memory
Extend chmem to use the new s390 kernel interface for configuring and
deconfiguring hotpluggable memory blocks, with memmap-on-memory support.
Background:
On s390, memmap-on-memory was introduced to ensure that the struct page
array (metadata) for hotpluggable standby memory is allocated from the
memory block itself. This allowed hot-add operations even under memory
pressure, particularly in cases with a strong imbalance between
boot-time online memory and standby memory.
The original design, however, had few limitations:
* All hotpluggable standby memory was added at boot.
* The use of memmap-on-memory was global and static, decided at boot
time. Either all standby blocks used it, or none of them did.
* memmap-on-memory choice could not be changed at runtime, limiting
flexibility. For example, when continuous physical memory was required
later across memory blocks.
The s390 kernel ff18dcb19aab ("s390/sclp: Add support for dynamic
(de)configuration of memory") no longer pre-adds all standby memory at
boot. Instead, users must explicitly configure a block before it can be
used for online/offline actions. At configuration time, users can
dynamically decide whether to use optional memmap-on-memory for each
memory block, where value of 1 allocates metadata (such as struct pages
array) from the hotplug memory itself, enabling hot-add operations even
under memory pressure. A value of 0 stores metadata in regular system
memory and enables continuous physical memory across memory blocks.
s390 kernel sysfs interface to configure/deconfigure memory with
memmap-on-memory support looks as shown below:
lsmem -o RANGE,SIZE,STATE,BLOCK,CONFIGURED,MEMMAP-ON-MEMORY
RANGE SIZE STATE BLOCK CONFIGURED MEMMAP-ON-MEMORY
0x00000000-0x7fffffff 2G online 0-15 yes no
0x80000000-0xffffffff 2G offline 16-31 no yes
Memory block size: 128M
Total online memory: 2G
Total offline memory: 2G
Memmap on memory parameter: yes
Just like online and offline actions, memory configuration and
deconfiguration can be controlled through similar options. Also,
memmap-on-memory setting can be changed, only when the memory block is
in deconfigured state. This means, it is usable only via --configure
option.
Sumanth Korikkar [Thu, 16 Oct 2025 15:38:02 +0000 (17:38 +0200)]
lsmem: add support to display dynamic (de)configuration of memory
Extend lsmem to display (de)configured blocks and memmap_on_memory
state. With the new s390 kernel interface (linux-next) ff18dcb19aab
("s390/sclp: Add support for dynamic (de)configuration of memory"),
standby memory blocks are no longer pre-added at boot, but must be
explicitly configured before being eligible for online/offline
operations. At configuration time, users can also decide whether to use
memmap-on-memory per block.
Add CONFIGURED column : indicate if a memory block has been explicitly
configured.
Add MEMMAP-ON-MEMORY column : indicate if a memory block uses
memmap-on-memory.
Sumanth Korikkar [Thu, 16 Oct 2025 15:38:01 +0000 (17:38 +0200)]
lsmem: display global memmap on memory parameter
Display the output of global memmap-on-memory parameter for memory
hotplug. Retrieve the details via
/sys/module/memory_hotplug/parameters/memmap_on_memory.
lsmem
RANGE SIZE STATE REMOVABLE BLOCK
0x0000000000000000-0x00000001ffffffff 8G online yes 0-63
Memory block size: 128M
Total online memory: 8G
Total offline memory: 0B
Memmap on memory parameter: yes
cgoesche [Sun, 2 Nov 2025 16:55:09 +0000 (11:55 -0500)]
su: pass arguments after <user> to shell
The su(1) manpage describes how the arguments after <user>
are passed to the invoked shell. However this is empirically
wrong, as option flags after <user> are interpreted by su(1)
and will eventually never be passed or yield an error that
terminates the program due to an unrecognized option flag.
To fix this we can change getopt(3)'s scanning mode with a '+'
prefixed to 'optstring', this will make it so that getopt(3)
stops processing argv elements on the first occurrence of a
non-option argument, e.g. '-' or '<user>'.
Additionally, if the argument that directly follows '-' is an
option flag, su(1) will assume that this argument and the ones
that follow, are to be passed to a shell invoked by the root user.
losetup: remove the --verbose flag, as it doesn't actually do anything
Fourteen years ago, commit c7e0925def rewrote the `losetup` tool,
removing all references to the 'verbose' variable. Three years
later, commit 60cb2c3720 removed the line 'verbose = 1' because
the compiler complained that the variable was set but never used.
Munehisa Kamata [Tue, 28 Oct 2025 19:54:17 +0000 (12:54 -0700)]
wdctl: remove -d option leftover
-d option was removed in commit f56338b43973 ("wdctl: allow to specify more
than one device"), but the optstring wasn't updated at that time and wdctl
can still accept the option halfway as below:
$ wdctl -d
wdctl: option requires an argument -- 'd'
whereas it should say:
wdctl: invalid option -- 'd'
So update the optstring.
Fixes: f56338b43973 ("wdctl: allow to specify more than one device") Signed-off-by: Munehisa Kamata <kamatam@amazon.com>
fincore: do not fall back to mincore if cachestat fails with EPERM
cachestat() and mincore() both require that the tested file is
(potentially) writable by the current user. If this permission check
fails, cachestat() will return EPERM while mincore() will simply mark
all pages as resident in core, as a proper EPERM would violate its API
contract. But when cachestat() fails with EPERM we know that mincore()
will not return real data, so instead show an error message.
Reported-by: Christian Hesse <mail@eworm.de> Signed-off-by: Thomas Weißschuh <thomas@t-8ch.de>
Karel Zak [Thu, 30 Oct 2025 11:11:43 +0000 (12:11 +0100)]
libfdisk: (dos) fix off-by-one in maximum last sector calculation
The get_disk_ranges() function incorrectly capped the last usable
sector at UINT_MAX, which could cause an overflow when calculating
partition size for MBR partition tables.
MBR stores partition size as a 32-bit value with maximum UINT_MAX.
The partition size is calculated as: size = stop - start + 1
For a partition starting at sector 0:
- If stop = UINT_MAX: size = UINT_MAX + 1 (overflow!)
- If stop = UINT_MAX - 1: size = UINT_MAX (correct maximum)
This fixes the inconsistency where dos_init() correctly warns about
disks larger than UINT_MAX sectors (2TiB - 512 bytes for 512-byte
sectors), but get_disk_ranges() allowed creating partitions that
would overflow the 32-bit size field.
Addresses: https://issues.redhat.com/browse/RHEL-122367 Signed-off-by: Karel Zak <kzak@redhat.com>
Bastian Krause [Fri, 24 Oct 2025 15:32:20 +0000 (17:32 +0200)]
hwclock: skip RTC_PARAM_SET for --param-set with unchanged value
Parameters set with `hwclock --param-set` tend to be persisted in the
RTC's EEPROM. Writing the same value over and over again can wear out
the EEPROM (e.g. on each boot).
So read the current value first. Only if the parameter is changed,
actually write the new value.
This allows for easier integrations, especially since there is no
machine-readable way of retrieving the current value via hwclock.
A lot of documentation on the Internet seems to assume "defaults" is the
/correct/ default value when no other options are intended.
Documentation/example.files/fstab does not have an entry for swap, but
it shows "defaults" for other file systems.
It seems prudent to align on a single variant, at least in the
documentation, even if both are accepted by swapon.
Signed-off-by: Chris Hofstaedtler <zeha@debian.org>
Karel Zak [Wed, 22 Oct 2025 08:42:36 +0000 (10:42 +0200)]
mount: add note about systemd and --all historical context
Add information that mount -a was originally designed for init scripts
but many modern systemd-based distributions use systemd units instead
for mounting filesystems on boot in a more sophisticated way.
Karel Zak [Wed, 22 Oct 2025 08:36:32 +0000 (10:36 +0200)]
mount: improve --all documentation regarding swap areas
The --all option description did not mention that swap entries in
fstab are silently ignored. Add a note clarifying this behavior and
pointing users to swapon --all for enabling swap devices and files.
Karel Zak [Tue, 21 Oct 2025 09:45:54 +0000 (11:45 +0200)]
losetup: improve command line option processing
The options --detach, --remove, and --set-capacity are mutually
exclusive. We can assume the device name follows the options, which is
better than assuming the device name is specified as an option's
argument. This also allows the use of the existing mutually-exclusive
check.
Karel Zak [Tue, 21 Oct 2025 08:58:36 +0000 (10:58 +0200)]
lostup: report EACCES on loop-control
Don't start scanning /dev/loop* after a failed access to
/dev/loop-control. It's a waste of time, and the user does not receive
a relevant error message.
Karel Zak [Thu, 16 Oct 2025 10:16:27 +0000 (12:16 +0200)]
losetup: improve --remove documentation
Rewrite the --remove option description to clearly explain the
difference between --detach and --remove. Add a new LOOP DEVICE
LIFECYCLE section that describes the three stages: creation,
detachment, and removal, with corresponding ioctl names.
Karel Zak [Thu, 16 Oct 2025 09:56:08 +0000 (11:56 +0200)]
losetup: make --remove a long-only option with mutual exclusivity
Change --remove from '-R, --remove' to a long-only option '--remove'.
This makes it consistent with other administrative options and adds
mutual exclusivity with other major actions like -d, -D, -a, -c, -f,
-j, -l, and -O.
Add a new static helper function loopcxt_get_device_nr() to extract
the loop device number from the device path. This eliminates code
duplication in loopcxt_remove_device() and loopcxt_add_device().
The helper function supports both /dev/loop<N> and /dev/loop/<N>
formats and provides consistent error handling with debug logging.
Karel Zak [Thu, 16 Oct 2025 09:22:47 +0000 (11:22 +0200)]
losetup: add error feedback for --remove command
The remove_loop() function was silently failing without providing
any feedback to the user. This also fixes incorrect indentation
(spaces instead of tabs).
Signed-off-by: Karel Zak <kzak@redhat.com> Reported-by: Benno Schulenberg <bensberg@telfort.nl>
Karel Zak [Wed, 15 Oct 2025 13:46:06 +0000 (15:46 +0200)]
treewide: use is_dotdir_dirent() helper
This simplifies the code by using the is_dotdir_dirent() helper
function instead of manual strcmp() checks for "." and ".." directory
entries across multiple utilities.
Karel Zak [Wed, 15 Oct 2025 13:30:06 +0000 (15:30 +0200)]
tests: (configs) add sorting and masking test
Add a new test that verifies:
- Main config file appears first
- Drop-in files are sorted alphabetically
- Files from /etc mask same-named files from /usr
- Priority ordering is correct (etc > usr)
The test creates mmm.conf in both /etc and /usr to verify
that the /etc version takes precedence and /usr version is
masked (not included in output).
Karel Zak [Wed, 15 Oct 2025 13:18:05 +0000 (15:18 +0200)]
lib/configs: simplify merge error checking
Combine the three config_merge_list() calls into a single
conditional statement to reduce repetitive error checking.
The calls are short-circuited on first failure.
Karel Zak [Wed, 15 Oct 2025 13:15:02 +0000 (15:15 +0200)]
lib/configs: eliminate counter variable
Remove the counter variable and return list_count_entries()
directly. This simplifies the code by calculating the count
only when needed (on success path) rather than tracking it
throughout the function.
Karel Zak [Wed, 15 Oct 2025 13:11:25 +0000 (15:11 +0200)]
lib/configs: add head parameter to configs_refer_filename()
Allow configs_refer_filename() to add entries at either the head
or tail of the list by adding a 'head' parameter. This simplifies
the code for adding the main config file, eliminating the need to
add to the tail and then move to the head.
When head=1, use list_add() to prepend to the list.
When head=0, use list_add_tail() to append to the list.
Karel Zak [Wed, 15 Oct 2025 12:59:44 +0000 (14:59 +0200)]
lib/configs: merge new_list_entry() and configs_add_filename()
Introduce configs_refer_filename() which merges the functionality
of new_list_entry() and configs_add_filename() while avoiding
unnecessary string duplication.
The new function takes ownership of the filename pointer directly
instead of using strdup(), which eliminates one allocation per
config file entry. Callers no longer need to free the filename
after adding it to the list, as ownership is transferred to the
list element.
This improves both performance and code clarity by making the
ownership semantics explicit.
Karel Zak [Wed, 15 Oct 2025 12:52:29 +0000 (14:52 +0200)]
lib/configs: refactor directory list merging
The original implementation used complex nested loops to merge
configuration file lists from different directories. This commit
simplifies the code by introducing a new config_merge_list()
function that handles the merging logic.
Changes:
- Add config_merge_list() to merge lists with duplicate detection
- Add config_cmp() comparison function using strcoll() (consistent
with alphasort() behavior from scandirat())
- Add configs_add_filename() helper to reduce code duplication
- Simplify ul_configs_file_list() by replacing ~120 lines of
merging logic with 3 calls to config_merge_list()
- Remove intermediate etc_run_file_list, merge directly to output
- Update read_dir() to return 0/-ENOMEM instead of entry count
- Use list_count_entries() to get final count
The new config_merge_list() moves entries directly from source
lists to the destination list without extra allocations, making
it more efficient than the previous approach.
Karel Zak [Tue, 14 Oct 2025 11:20:42 +0000 (13:20 +0200)]
tests: (configs) add file listings to expected output
Add sorted file listings to the configs test output to help debug
platform-specific test failures. Each subtest now outputs the actual
files created in the test directory before running the config helper,
making it easier to identify issues with filesystem ordering or file
creation problems.
Karel Zak [Thu, 9 Oct 2025 20:19:40 +0000 (22:19 +0200)]
lib/configs: introduce config_mk_path() helper
Refactor main_configs() and read_dir() to use a new helper function
that combines path construction with existence and type checking.
This reduces code duplication and simplifies the logic.
Karel Zak [Thu, 9 Oct 2025 20:04:19 +0000 (22:04 +0200)]
tests: add regression tests for ul_configs_file_list()
Add comprehensive test script for configuration file list functionality
with the following test cases:
- main-etc: Main config file in /etc
- main-usr: Main config file fallback to /usr
- dropin-etc: Drop-in files from /etc
- dropin-usr: Drop-in files from /usr
- combined: Main config + drop-ins from multiple directories
- masking: Same basename in multiple directories (masking behavior)
- no-project: Configuration without project subdirectory
The tests verify proper file discovery, priority ordering across /etc,
/run, and /usr directories, and file masking behavior according to the
Configuration Files Specification.
Karel Zak [Thu, 9 Oct 2025 20:02:24 +0000 (22:02 +0200)]
lib/configs: add test program for ul_configs_file_list()
Add a test program with command line options to test the configuration
file list functionality. The test program allows specifying custom paths
for /etc, /run, and /usr directories, project name, config name, and suffix.
This enables testing the priority ordering and file discovery logic
of ul_configs_file_list() from the command line.
projects / thirdparty / util-linux.git / log
