summaryrefslogtreecommitdiffstats
path: root/docs (unfollow)
Commit message (Collapse)AuthorFilesLines
2020-10-01sysusers: look at login.defs when setting the default range to allocate usersZbigniew Jędrzejewski-Szmek18-25/+143
Also, even if login.defs are not present, don't start allocating at 1, but at SYSTEM_UID_MIN. Fixes #9769. The test is adjusted. Actually, it was busted before, because sysusers would never use SYSTEM_GID_MIN, so if SYSTEM_GID_MIN was different than SYSTEM_UID_MIN, the tests would fail. On all "normal" systems the two are equal, so we didn't notice. Since sysusers now always uses the minimum of the two, we only need to substitute one value.
2020-10-01test-sysusers: fix how paths are calculatedZbigniew Jędrzejewski-Szmek1-8/+8
We were looking at ${f%.*}, i.e. the $f with any suffix starting with a dot removed. This worked fine for paths like /some/path/test-11.input. It also worked for paths like /some/path/inline (there were no dots, so we got $f back unscathed). But in the ubuntu CI the package is built in a temporary directory like /tmp/autopkgtest-lxc.nnnfqb26/downtmp/build.UfW/ (yes, it has a dot, even two.). That still worked for the first case, but in the second case we truncated things after the first dot, and we would try to get /tmp/autopkgtest-lxc.nnnfqb26/downtmp/build and try to load /tmp/autopkgtest-lxc.nnnfqb26/downtmp/build.expected-password, which obviously didn't work as expected. To avoid this issue, do the suffix removal only when we know that there really is a suffix. A second minor issue was that we would try to copy $1.expected-*, and sometimes $1 would be given, and sometimes not. Effectively we were relying on there not being any files matching .expected-*. There weren't any such files, but let's avoid this ugliness and always pass $1.
2020-10-01test-sysusers: sort examplesZbigniew Jędrzejewski-Szmek1-4/+4
This shouldn't affect the outcome, but makes outputs easier to compare.
2020-10-01test/TEST-21-SYSUSERS: turn into a unit testZbigniew Jędrzejewski-Szmek61-134/+148
All this test does is manipulate text files in a subdir specified with --testroot. It can be a normal unittest without the overhead of creating a machine image. As a bonus, also test the .standalone version.
2020-10-01basic/fileio: add chase_symlinks_and_fopen_unlocked()Zbigniew Jędrzejewski-Szmek3-1/+45
2020-10-01shared/uid-range: reduce scope of iterator variablesZbigniew Jędrzejewski-Szmek1-11/+5
2020-10-01Also parse the minimum uid/gid valuesZbigniew Jędrzejewski-Szmek6-35/+58
We don't (and shouldn't I think) look at them when determining the type of the user, but they should be used during user/group allocation. (For example, an admin may specify SYS_UID_MIN==200 to allow statically numbered users that are shared with other systems in the range 1–199.)
2020-10-01Look at /etc/login.defs for the system_max_[ug]id valuesZbigniew Jędrzejewski-Szmek7-14/+175
It makes little sense to make the boundary between systemd and user guids configurable. Nevertheless, a completely fixed compile-time define is not enough in two scenarios: - the systemd_uid_max boundary has moved over time. The default used to be 500 for a long time. Systems which are upgraded over time might have users in the wrong range, but changing existing systems is complicated and expensive (offline disks, backups, remote systems, read-only media, etc.) - systems are used in a heterogenous enviornment, where some vendors pick one value and others another. So let's make this boundary overridable using /etc/login.defs. Fixes #3855, #10184.
2020-09-25coredump: use uid_is_system() when appropriateZbigniew Jędrzejewski-Szmek1-1/+1
2020-09-25Move {uid,gid}_is_*() from basic to sharedZbigniew Jędrzejewski-Szmek7-25/+30
Those are functions that express policy, and nothing in basic/ uses (or should use) them.
2020-09-24hwdb: add Medion Akoya E2221T MD60691 (#17147)germanztz1-0/+5
2020-09-24Translated using Weblate (Chinese (Simplified))Charles Lee1-29/+13
Currently translated at 63.1% (118 of 187 strings) Co-authored-by: Charles Lee <lchopn@gmail.com> Translate-URL: https://translate.fedoraproject.org/projects/systemd/master/zh_CN/ Translation: systemd/master
2020-09-24random-util: Add an environment variable to disable RDRAND.Kyle Huey2-0/+13
SYSTEMD_RDRAND=0 will prevent using RDRAND even on systems whose CPUID claims to support it. All other values have no effect. Fixes: #17112
2020-09-23update TODOLennart Poettering1-4/+0
2020-09-23tree-wide: port remaining umount() calls to umount_verbose()Lennart Poettering1-4/+4
2020-09-23tree-wide: switch remaining mount() invocations over to mount_nofollow_verbose()Lennart Poettering6-116/+127
(Well, at least the ones where that makes sense. Where it does't make sense are the ones that re invoked on the root path, which cannot possibly be a symlink.)
2020-09-23mount-util: rework umount_verbose() to take log level and flags argLennart Poettering10-20/+33
Let's make umount_verbose() more like mount_verbose_xyz(), i.e. take log level and flags param. In particular the latter matters, since we typically don't actually want to follow symlinks when unmounting.
2020-09-23mount-util: switch most mount_verbose() code over to not follow symlinksLennart Poettering14-103/+140
2020-09-23mount-util: make sure get_mount_flags() ever follows symlinksLennart Poettering1-3/+11
It doesn't follow symlinks when we check for the mount flags via libmount, hence it shouldn't do either when we go via statvfs().
2020-09-23mount-util: add helpers for mount() without following symlinksLennart Poettering2-0/+53
2020-09-23shutdown: also fsync() DM devices before going downLennart Poettering1-15/+20
Similar reasons as the previous patch.
2020-09-23shutdown: also fsync() MD devices when going downLennart Poettering1-0/+3
Let's make this explicit, just in case this suffers by the same issues as the loopback devices, and drops in-flight IO when we disassemble it.
2020-09-23shutdown: don't compare pointers with 0Lennart Poettering1-2/+2
2020-09-23shutdown: fsync() before detaching loopback devicesLennart Poettering1-1/+8
This is a follow-up for cae1e8fb88c5a6b0960a2d0be3df8755f0c78462: we also call the detach ioctls in the shutdown code, hence add the fsync()s there too, just to be safe.
2020-09-23fs-util: check for /proc mounted in access_fd()Lennart Poettering1-5/+14
It's a very basic function very similar to other cases where we go via /proc/self/fd/, hence do the explicit proc_mounted() check here too.
2020-09-23fs-util: use strna() on returned strings of fd_get_path() if we don't check ↵Lennart Poettering1-2/+2
its return value Let's make sure to use strna() on the strings returned by fd_get_path() where we knowingly ignore any failures. We got this right in most cases, but two were missing.
2020-09-23basic: update fd_get_path() to use proc_mounted() helperLennart Poettering2-7/+5
We use it pretty much everywhere else, hence use it here too. This also changes the error generated from EOPNOTSUPP to ENOSYS, to match the other cases where we do such a check. One user checked for EOPNOTSUPP which is updated to check for ENOSYS instead.
2020-09-23fs-util: drop chmod_and_chown_unsafe() which is unused nowLennart Poettering3-92/+0
2020-09-23sysusers: modernize file backup logic a bitLennart Poettering1-24/+21
Let's use _cleanup_ magic to clean up files, let's fully operate by fds whenever we can.
2020-09-23update TODOLennart Poettering1-16/+7
2020-09-23run: let systemd resolve the path with RootDirectory=/RootImage=Zbigniew Jędrzejewski-Szmek2-6/+12
Fixes #13338.
2020-09-23core: don't try to load units from non-absolute pathsZbigniew Jędrzejewski-Szmek1-3/+2
The error message disagreed with the check that was actually performed. Adjust the check.
2020-09-23catalog: add missing @ to German catalog fileLennart Poettering1-1/+1
2020-09-23sd-radv: fix indentationLennart Poettering1-10/+10
Fix-up for 97efde65d81b1501a3737f408f1234c6c25da61c
2020-09-23machine: fix error code propagation in two casesLennart Poettering1-2/+2
2020-09-23man: sd_bus_message_new: fix typoRoy Chen (陳彥廷)1-1/+1
2020-09-23sd-path: include 'search' in search pathnamesRonan Pigott1-4/+4
2020-09-23zsh: add systemd-path completionsRonan Pigott2-0/+9
2020-09-22shared/conf-parser: drop redundant cast to booleanZbigniew Jędrzejewski-Szmek1-1/+1
parse_boolean returns either 0 or 1 or error, and we checked for errors earlier already.
2020-09-22logind: use _cleanup_ in one more placeZbigniew Jędrzejewski-Szmek1-16/+6
2020-09-22sd-{login,netlink,network}: use TAKE_FD() in more placesZbigniew Jędrzejewski-Szmek3-21/+7
2020-09-22test-bus-chat: add missing return value in checkZbigniew Jędrzejewski-Szmek1-1/+1
2020-09-22various: remove assignments of unread variablesZbigniew Jędrzejewski-Szmek3-3/+1
2020-09-22activate: reduce scope of iterator variableZbigniew Jędrzejewski-Szmek1-7/+4
2020-09-22core/namespace: drop bitfield annotations from boolean fieldsZbigniew Jędrzejewski-Szmek1-13/+13
Such microoptimization makes sense when the structure is used in many many copies, but here's it's not, and the few bytes we save are not worth the extra code the compiler has to generate: return ns_info->mount_apivfs || ns_info->protect_control_groups || ns_info->protect_kernel_tunables || ... before: 49b187: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b18b: 0f b6 00 movzbl (%rax),%eax 49b18e: 83 e0 80 and $0xffffff80,%eax 49b191: 84 c0 test %al,%al 49b193: 75 32 jne 49b1c7 <namespace_info_mount_apivfs+0x80> 49b195: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b199: 0f b6 00 movzbl (%rax),%eax 49b19c: 83 e0 08 and $0x8,%eax 49b19f: 84 c0 test %al,%al 49b1a1: 75 24 jne 49b1c7 <namespace_info_mount_apivfs+0x80> 49b1a3: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b1a7: 0f b6 00 movzbl (%rax),%eax 49b1aa: 83 e0 10 and $0x10,%eax 49b1ad: 84 c0 test %al,%al 49b1af: 75 16 jne 49b1c7 <namespace_info_mount_apivfs+0x80> after: 49b024: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b028: 0f b6 40 07 movzbl 0x7(%rax),%eax 49b02c: 84 c0 test %al,%al 49b02e: 75 2e jne 49b05e <namespace_info_mount_apivfs+0x7a> 49b030: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b034: 0f b6 40 03 movzbl 0x3(%rax),%eax 49b038: 84 c0 test %al,%al 49b03a: 75 22 jne 49b05e <namespace_info_mount_apivfs+0x7a> 49b03c: 48 8b 45 f8 mov -0x8(%rbp),%rax 49b040: 0f b6 40 04 movzbl 0x4(%rax),%eax 49b044: 84 c0 test %al,%al 49b046: 75 16 jne 49b05e <namespace_info_mount_apivfs+0x7a>
2020-09-22ptyfwd: use ERRNO_IS_DISCONNECT() when checking for disconnection on foreign fdsLennart Poettering1-2/+3
2020-09-22tree-wide: use ERRNO_IS_PRIVILEGE() whereever appropriateLennart Poettering5-13/+15
2020-09-22dissect-image: do not refuse verity GPT without /usr partitionLuca Boccassi1-1/+1
Only enforce that /usr verity partition is present if a /usr partition is there
2020-09-22dissect-image: use correct path variable when reading verity signaturesLuca Boccassi1-2/+2
Copypasta slipped in via https://github.com/systemd/systemd/pull/17101
2020-09-22Try stopping MD RAID devices in shutdown tooHubert Kario3-10/+146
Currently the systemd-shutdown command attempts to stop swaps, DM (crypt, LVM2) and loop devices, but it doesn't attempt to stop MD RAID devices, which means that if the RAID is set up on crypt, loop, etc. device, it won't be able to stop those underlying devices. This code extends the shutdown application to also attempt stopping the MD RAID devices. Signed-off-by: Hubert Kario <hubert@kario.pl>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-02 02:16:34 +0100