X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=blobdiff_plain;f=TODO;h=e9ab9de6e7936a35b2407b639ffb9f0b7fe0b92a;hp=2bade7b5499fcc335ff40afc31d30e6e2c8c70d7;hb=4bba9156da3e1df2cee24d10d7cd88c776ef4179;hpb=7a2a0b907b5cc60f5d9a871997d7d6e7f62bf4d8 diff --git a/TODO b/TODO index 2bade7b54..e9ab9de6e 100644 --- a/TODO +++ b/TODO @@ -1,237 +1,188 @@ -* udev-kernel.socket + udev.control.socket seems not - to work, udevd is started but no fd is passed - -F15: +Bugfixes: * swap units that are activated by one name but shown in the kernel under another are semi-broken -F15 External: - -* NFS, networkmanager ordering issue (PENDING) - * NM should pull in network.target (PENDING) https://bugzilla.redhat.com/show_bug.cgi?id=692008 -* bluetooth should be possible to disable (PENDING) - * make anaconda write timeout=0 for encrypted devices -* fix broken Sockets=syslog-ng.socket packaging +* service: pid file reading after reload doesn't work, since we don't reset the pid variable + +* make sure timeouts are applied to Type=oneshot services. + +* Dangling symlinks of .automount unit files in .wants/ directories, set up + automount points even when the original .automount file did not exist + anymore. Only the .mount unit was still around. + +* make polkit checks async + +* properly handle .mount unit state tracking when two mount points are stacked one on top of another on the exact same mount point. Features: -* possibly apply systemd-sysctl per network device subtrees on hotplug +* logind: selinux is borked... -* add conditions for file executability +* logind: sends SessionNew on Lock()? -* allow creating fifos with tmpfiles +* logind: allow showing logout dialog from system -* implement Register= switch in .socket units to enable registration - in Avahi, RPC and other socket registration services. +* document that %% can be used to write % in a string that is specifier extended -* Remove kill_mode parameter to kill bus calls +* check utf8 everywhere -* Fix nspawn to not read-only mount /selinux on the host system +* when an instanced service exits, remove its parent cgroup too if possible. -* make sure people don't leave processes around after ExecStartPre= +* Make libselinux, libattr, libcap, libdl dependencies only of the tools which actually need them. -* make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early +* as Tom Gundersen pointed out there's a always a dep loop if people use crypto file systems with random keys -* add loginctl, i.e. a systemctl for logind introspection +* unset container= in PID1? -* support presets +* automatically escape unit names passed on the service (i.e. think "systemctl start serial-getty.service@serial/by-path/jshdfjsdfhkjh" being automatically escaped as necessary. -* kernel: add /proc/sys file exposing CAP_LAST_CAP? +* if we can not get user quota for tmpfs, mount a separate tmpfs instance + for every user in /run/user/$USER with a configured maximum size -* kernel: add device_type = "fb", "fbcon" to class "graphics" +* default to actual 32bit PIDs, via /proc/sys/kernel/pid_max -* understand https://bugzilla.redhat.com/show_bug.cgi?id=672194 +* add an option to make mounts private/shareable and so on, enable this for root by default -* readahead: use BTRFS_IOC_DEFRAG_RANGE instead of BTRFS_IOC_DEFRAG ioctl, with START_IO +* internal restart counter for units (focus on auto-respawn) -* readahead: check whether a btrfs volume includes ssd by checking mount flag "ssd" +* finer-grained auto-respawn settings (rate-limit) -* support sd_notify() style notificatio when reload is finished (RELOADED=1) +* be able to specify a forced restart of service A where service B depends on, in case B + needs to be auto-respawned? -* verify that the AF_UNIX sockets of a service in the fs still exist - when we start a service in order to avoid confusion when a user - assumes starting a service is enough to make it accessible +* Something is wrong with symlink handling of "autovt@.service" in "systemctl list-unit-files" -* Make it possible to set the keymap independently from the font on - the kernel cmdline. Right now setting one resets also the other. +* introduce sd_log_open() to create a connection to the syslog bridge via sd-daemon.h -* add dbus call to convert snapshot into target +* when a bus name of a service disappears from the bus make sure to queue further activation requests -* move nss-myhostname into systemd +* something like ConditionExec= or ExecStartPre= without failure state -* figure out a standard place to configure timezone name, inform myllynen@redhat.com +* service restart retry configuration -* add dbus call to convert snapshot into target, and a dbus call to generate target from current state +* tmpfiles: apply "x" on "D" too (see patch from William Douglas) +* tmpfiles: support generation of char/block devices, symlinks and one-line files (think sysfs) -* detect LXC with $container=lxc +* Introduce ControlGroupPersistant=yes to set +t on the tasks file when creating the cgroup -* add prefix match to sysctl, tmpfiles, ... +* don't set $HOME in services unless requested -* drop /.readahead on bigger upgrades with yum +* remove hacks in vala code and require libnotify 0.7 -* add inode stat() check to readahead to suppress preloading changed files +* hide PAM/TCPWrap options in fragment parser when compile time disabled -* allow list of paths in config_parse_condition_path() +* when we automatically restart a service, ensure we retsart its rdeps, too. -* introduce dbus calls for enabling/disabling a service +* allow Type=simple with PIDFile= + https://bugzilla.redhat.com/show_bug.cgi?id=723942 -* support notifications for services being enabled/disabled +* move PAM code into its own binary -* show enablement status in systemctl status +* warn if the user stops a service but not its associated socket -* consider services with any kind of link in /etc/systemd/system enabled +* logind: spawn user@..service on login -* teach systemctl to enable unit files in arbitrary directories +* logind: non-local X11 server handling -* In systemctl make sure both is-enabled and is-active print a string, or neither. +* implement Register= switch in .socket units to enable registration + in Avahi, RPC and other socket registration services. -* Implement: - systemctl mask - systemctl unmask - Also support --temp to make this temporary by placing mask links in /run. +* make sure people don't leave processes around after ExecStartPre= -* add support for /bin/mount -s +* make sure systemd-ask-password-wall does not shutdown systemd-ask-password-console too early -* GC unreferenced jobs (such as .device jobs) +* readahead: use BTRFS_IOC_DEFRAG_RANGE instead of BTRFS_IOC_DEFRAG ioctl, with START_IO -* add JoinControllers= to system.conf to mount certain cgroup - controllers together in order to guarantee atomic creation/addition - of cgroups +* readahead: check whether a btrfs volume includes ssd by checking mount flag "ssd" -* avoid DefaultStandardOutput=syslog to have any effect on StandardInput=socket services +* support sd_notify() style notification when reload begins (RELOADING=1), reload is finished (READY=1) -* cgroup_notify_empty(): recursively check groups up the tree, too +* support sd_notify() style notification when shutting down, to make auto-exit bus services work (STOPPING=1) -* fix alsa mixer restore to not print error when no config is stored +* verify that the AF_UNIX sockets of a service in the fs still exist + when we start a service in order to avoid confusion when a user + assumes starting a service is enough to make it accessible + +* Make it possible to set the keymap independently from the font on + the kernel cmdline. Right now setting one resets also the other. -* fix upstart reboot compat call +* move nss-myhostname into systemd + +* and a dbus call to generate target from current state + +* detect LXC with $container=lxc + +* drop /.readahead on bigger upgrades with yum + +* add inode nr check to readahead to suppress preloading changed files + +* add support for /bin/mount -s + +* GC unreferenced jobs (such as .device jobs) + +* cgroup_notify_empty(): recursively check groups up the tree, too * when failing to start a service due to ratelimiting, try again later, if restart=always is set * write blog stories about: - enabling dbus services - status update - - /etc/sysconfig and /etc/default - how to make changes to sysctl and sysfs attributes + - remote access + - cgroup best pratices to avoid stepping on each others toes + - how to pass throw-away units to systemd, or dynamically change properties of existing units + - how to integrate cgconfig and suchlike with systemd + - security properties + - xinetd conversion + - details of instantiated @.service setups * allow port=0 in .socket units -* rename systemd-logger to systemd-stdio-syslog-bridge - -* take BSD file lock on tty devices when using them? - -* avoid any flag files, or readahead files in /, we need to support r/o / - or / on tmpfs like Android setups. - * move readahead files into /var, look for them with .path units * teach dbus to activate all services it finds in /etc/systemd/services/org-*.service -* get process transport into dbus for systemctl -P/-H - -* document default dependencies - -* support systemd.whitelist=/systemd.blacklist= on the kernel command - line. - -* Find a way to replace /var/run, /var/lock directories with - symlinks during an RPM package upgrade (filesystem.rpm or systemd.rpm). - (lua code to create symlinks right away for new installations is in filesytem.rpm now) - We soon want to get rid of var-run.mount var-lock.mount units: - if mountpoint /run ; then - umount /var/run || : - else - mount --move /var/run /run || mount --bind /var/run /run - fi - mv /var/run /var/.run.save - ln -s /run /var/run - echo "R /var/.run.save" > /etc/tmpfiles.d/remove-run-save.conf +* support systemd.mask= on the kernel command line. * when key file cannot be found, read it from kbd in cryptsetup -* add switch to systemctl to show enabled but not running services. Or - another switch that shows service that have been running since - booting but aren't running anymore. - * reuse mkdtemp namespace dirs in /tmp? * recreate systemd's D-Bus private socket file on SIGUSR2 -* be more specific what failed: - ... - Unmounting file systems. - Not all file systems unmounted, 1 left. - Disabling swaps. - Detaching loop devices. - Detaching DM devices. - Cannot finalize remaining file systems and devices, trying to kill remaining processes. - Unmounting file systems. - Not all file systems unmounted, 1 left. - Cannot finalize remaining file systems and devices, giving up. - ... - -* check for compiled-in, but not active selinux, and don't print any warnings - about policy loading. Probably check for available selinux in /proc/filesystems, - and check for active selinux with getcon_raw() == "kernel" - * Support --test based on current system state -* show failure error string in "systemctl status" - -* make sure timeouts are applied to Type=oneshot services. - -* detect LXC environment - * investigate whether the gnome pty helper should be moved into systemd, to provide cgroup support. -* perhaps add "systemctl reenable" as combination of "systemctl disable" and "systemctl enable" - -* need a way to apply mount options of api vfs from systemd unit files - (or some other modern source?) instead of fstab? - * maybe introduce ExecRestartPre= -* figure out what happened to bluez patch +* configurable jitter for timer events -* Patch systemd-fsck to use -C and pass console fd to it +* timer events with system resume -* configurable jitter for timer events +* timer events on calendar time * Support ProcessNeededForShutdown=true to allow stuff like mdmon to be killed very late after the rootfs is read only? If implement pass this to shutdown binary via command line argument. -* use pivot_root on shutdown so that we can unmount the root directory. - - copy binaries to, and pivot_root() to the "shutdown tmpfs" - - solve "mdadm --wait-clean" problem that way - * dot output for --test showing the 'initial transaction' * calendar time support in timer, iCalendar semantics for the timer stuff (RFC2445) http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=99ee5315dac6211e972fa3f23bcc9a0343ff58c4 -* systemd --user - - get PR_SET_ANCHOR merged: http://lkml.org/lkml/2010/2/2/165 - -* add VT tracking: - - provide CK functionality - - start getty only when actual vt switch happens (same model as - socket on-demand activation). allocate the next free tty and - start a getty there. this way, pressing alt-f[1-12] will switch - through running X and getty sessions, and any unallocated - activated tty will start a new getty. the hardcoding of - getty[1-6] will entirely go away. - - http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=fbc92a3455577ab17615cbcb91826399061bd789 - * implicitly import "defaults" settings file into all types +* exec settings override +* writable cgroups dbus properties for live changes + +* read config fragments for all units from /lib/systemd/system/foobar.service.d/ to override/extend specific settings * port over to LISTEN_FDS/LISTEN_PID: - - uuidd HAVEPATCH - rpcbind (/var/run/rpcbind.sock!) HAVEPATCH - cups HAVEPATCH - postfix, saslauthd @@ -240,31 +191,38 @@ Features: - bluetoothd (/var/run/sdp! @/org/bluez/audio!) - distccd -* fingerprint.target, wireless.target, gps.target, netdevice.target +* auditd service files -* set_put(), hashmap_put() return values check. i.e. == 0 doesn't free()! +* fingerprint.target, wireless.target, gps.target, netdevice.target * io priority during initialization * if a service fails too often, make the service enter failed mode, and the socket, too. -* systemctl list-jobs - show dependencies +* new deps "Triggers" and "TriggeredBy" -* auditd service files +* systemctl list-jobs - show dependencies * add systemctl switch to dump transaction without executing it * suspend, resume support? -* readahead: btrfs/LVM SSD detection +* drop cap bounding set in readahead and other services -* add separate man page for [Install] settings +External: -* allow runtime changing of log level and target +* dbus: + - get process transport into dbus for systemctl -P/-H (PENDING) + - dbus --user + - natively watch for dbus-*.service symlinks (PENDING) + - allow specification of socket mode/umask when allocating DBusServer + - allow disabling of fd passing when connecting a AF_UNIX connection + - allow disabling of UID passing for AUTH EXTERNAL -* drop cap bounding set in readahead and other services +* systemd --user + PR_SET_CHILD_REAPER patch: https://lkml.org/lkml/2011/7/28/426 -External: +* fix alsa mixer restore to not print error when no config is stored * udisks should not use udisks-part-id, instead use blkid. also not probe /dev/loopxxx @@ -276,30 +234,19 @@ External: * patch kernel for xattr support in /dev, /proc/, /sys and /sys/fs/cgroup? +* NTP: the kernel's 11-minutes-mode syncs the system time to the RTC, but only + in an ~30 minutes window. It does not adjust larger differences. Find a way + to tell the kernel, to always do a full time sync when the RTC is in UTC and + we are in 11-minutes-mode. When we trust the system time to NTP we also want + the RTC to sync up. + * patch kernel for cpu feature modalias for autoloading aes/kvm/... - http://git.kernel.org/?p=linux/kernel/git/ak/linux-misc-2.6.git;a=shortlog;h=refs/heads/cpuid-match - (Rafael J. Wysocki's sysdev rework is on the way. After that CPUs can be exported a proper bus.) - -* procps, psmisc, sysvinit-tools, hostname → util-linux-ng - -https://bugzilla.redhat.com/show_bug.cgi?id=614245 -- plymouth -https://bugzilla.redhat.com/show_bug.cgi?id=612789 -- umount /cgroup on halt -https://bugzilla.redhat.com/show_bug.cgi?id=612728 -- /etc/rc.d/init.d/functions -https://bugzilla.redhat.com/show_bug.cgi?id=612712 -- pam_systemd -https://bugs.freedesktop.org/show_bug.cgi?id=29193 -- accountsservice -https://bugs.freedesktop.org/show_bug.cgi?id=29194 -- ConsoleKit -https://bugs.freedesktop.org/show_bug.cgi?id=29205 -- udisks -http://article.gmane.org/gmane.linux.bluez.kernel/6479 -- bluez -http://www.spinics.net/lists/linux-nfs/msg14371.html -- rpcbind -https://bugzilla.redhat.com/show_bug.cgi?id=617328 -- ntp -https://bugzilla.redhat.com/show_bug.cgi?id=617320 -- at -https://bugzilla.redhat.com/show_bug.cgi?id=617326 -- fprintd -https://bugzilla.redhat.com/show_bug.cgi?id=617333 -- yum -https://bugzilla.redhat.com/show_bug.cgi?id=617317 -- acpid -https://bugzilla.redhat.com/show_bug.cgi?id=617327 -- gpm -https://bugzilla.redhat.com/show_bug.cgi?id=617330 -- pcsc-lite -https://bugzilla.redhat.com/show_bug.cgi?id=617321 -- audit -https://bugzilla.redhat.com/show_bug.cgi?id=617316 -- abrt + (patches in linux-next, on the way to the next kernel) + +* kernel: add /proc/sys file exposing CAP_LAST_CAP? sysconf? + merged: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=73efc0394e148d0e15583e13712637831f926720 + +* kernel: add device_type = "fb", "fbcon" to class "graphics" Regularly: @@ -313,4 +260,4 @@ Regularly: * pahole -* CFLAGS="-Wl,--gc-sections -Wl,--print-gc-sections -ffunction-sections -fdata-sections" +* set_put(), hashmap_put() return values check. i.e. == 0 doesn't free()!