X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=blobdiff_plain;f=TODO;h=45d7b4ebc7c859a718cab035ec36041802ec17da;hp=8660f16f9c9ce60b5245e536d9885fef72e2adb7;hb=e252dd8f0ee40ac1c20d524f3622641c7907919f;hpb=060ed82ec24d942c5f519e3dae45e9e2bfb227d8 diff --git a/TODO b/TODO index 8660f16f9..45d7b4ebc 100644 --- a/TODO +++ b/TODO @@ -1,67 +1,155 @@ -Bugs: +F15: + +* swap units that are activated by one name but shown in the kernel under another are semi-broken * isolate multi-user.target doesn't start a getty@tty1 if we run it from graphical.target -* when plymouth is disabled the console password entry stuff seems to be borked - https://bugzilla.redhat.com/show_bug.cgi?id=655538 +* NFS, networkmanager ordering issue (PENDING) + +* add fstab fields to add wait timeouts, change Wants to Requires by local-fs.target + +* hook emergency.target into local-fs.target in some way as OnFailure with isolate + +* bind mounts are ignored + +* 0595f9a1c182a84581749823ef47c5f292e545f9 is borked, freezes shutdown + (path: after installing inotify watches, recheck file again to fix race) + +* NM should pull in network.target, ntpd should pull in rtc-set.target. + +* bluetooth should be possible to disable -* fsck.btrfs non-existance should not cause mounting to fail +* fix alsa mixer restore to not print error when no config is stored Features: -* perhaps add "systemctl reenable" as combination of "systemctl disable" and "systemctl enable" +* don't trim empty cgroups + https://bugzilla.redhat.com/show_bug.cgi?id=678555 -* need a way to apply mount options of api vfs from systemd unit files instead of fstab +* serialize used job ids and max job id -* udisks should not use udisks-part-id, instead use blkid. also not probe /dev/loopxxx +* expose monotonic timestamps on the bus and make systemd-analyze use it -* snd-seq should go, https://bugzilla.redhat.com/show_bug.cgi?id=676095 +* write blog stories about: + - chroot, nspawn and friends + - the blame game: systemd-analyze + - enabling dbus services + - status update -* mount.tmpfs should be optimized, https://bugzilla.redhat.com/show_bug.cgi?id=676100 +* allow port = 0 in .socket units -* mdmon is irre +* rename systemd-logger to systemd-stdio-syslog-bridge -* gnome-shell python script/glxinfo/is-accelerated wech +* introduce /usr/lib/binfmt.d/, /usr/lib/tmpfiles.d/ -* PID heuristik bei Type=forking ausmachbar machen +* in pam_systemd: don't rely on /proc/self/loginuid in a container -* maybe introduce ExecRestartPre= +* take BSD file lock on tty devices when using them? -* reload PID file after reload, allow dynamically changing main PIDs +* tmpfiles should allow two identical lines + https://bugzilla.redhat.com/show_bug.cgi?id=690253 -* figure out what happened to bluez patch +* avoid any flag files, or readahead files in /, we need to support r/o / + or / on tmpfs like Android setups. -* introduce StandardOutput=syslog+console and StandardOutput=kmsg+console to support fsck output at boot +* teach dbus to activate all services it finds in /etc/systemd/services/org-*.service -* Patch systemd-fsck to use -C and pass console fd to it +* get process transport into dbus for systemctl -P/-H -* support remote/ssh systemctl/systemadm, and local privileged access +* document default dependencies -* finish syslog socket stuff +* support systemd.whitelist=/systemd.blacklist= on the kernel command + line. -* configurable jitter for timer events +* Find a way to replace /var/run, /var/lock directories with + symlinks during an RPM package upgrade (filesystem.rpm or systemd.rpm). + We soon want to get rid of var-run.mount var-lock.mount units. + +* when key file cannot be found, read it from kbd in cryptsetup + +* get rid of random file name in generator directory? + /run/systemd/generator-IH1vFu + +* fix SD_WARNING syslog stuff in src/sd-daemon.h to include the + LOG_DAEMON(3) facility value. Never use the LOG_KERNEL(0) facility. + +* 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? + +* don't strip facility from kmsg log messages as soon as that is possible: + http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=9d90c8d9cde929cbc575098e825d7c29d9f45054 + +* 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" + +* optionally create watched directories in .path units + +* Support --test based on current system state + +* consider services with no [Install] section and stored in /lib enabled by "systemctl is-enabled" + +* consider services with any kind of link in /etc/systemd/system enabled -* support caching password questions in plymouth and on the console - https://bugzilla.redhat.com/show_bug.cgi?id=655538 +* show failure error string in "systemctl status" -* exclude java hsp files by default - https://bugzilla.redhat.com/show_bug.cgi?id=527425 +* make sure timeouts are applied to Type=oneshot services. -* load EnvironmentFile= when starting services, not when reloading configuration - https://bugzilla.redhat.com/show_bug.cgi?id=661282 +* Maybe implement "systemctl mask" and "systemctl unmask", but not + document it? When doing that add switch to make this temporary by + placing mask links in /dev. + Consider moving the actual fs operations into systemd behind a D-Bus + interface, to make namespaces/containers/remote connections work properly. -* drop IN_ATTRIB from inotify watches for .path units where possible to avoid - lots of wakeups due to /dev changing when we watch a subdir of /dev. +* detect LXC environment + +* invoke vhangup() before and after invoking getty + http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=3c95c985fa91ecf6a0e29622bbdd13dcfc5ce9f1 + +* support "auto" and "comment=systemd.automount" at the same time for an fstab entry + +* Maybe store in unit files whether a service should be enabled by default on package installation + (belongs into a distro pattern though, not in an upstream package's service file) + +* 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 + +* Patch systemd-fsck to use -C and pass console fd to it + +* support remote/ssh systemctl/systemadm, and local privileged access → dbus patches need to be merged + +* configurable jitter for timer events * Support ProcessNeededForShutdown=true to allow stuff like mdmon - to be killed very late after the rootfs is read only + to be killed very late after the rootfs is read only (?) * Integrate "mdadm --wait-clean". Maybe just let SIGTERM to mdmon trigger the needed action -* oneshot services which do not remain: 'exited' instead of 'dead'? - it should be visible in 'systemctl' that they have been run? - * dot output for --test showing the 'initial transaction' * calendar time support in timer, iCalendar semantics for the timer stuff (RFC2445) @@ -71,32 +159,28 @@ Features: - 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/gregkh/tty-2.6.git;a=commitdiff;h=fbc92a3455577ab17615cbcb91826399061bd789 + - 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/gregkh/tty-2.6.git;a=commitdiff;h=fbc92a3455577ab17615cbcb91826399061bd789 * implicitly import "defaults" settings file into all types * port over to LISTEN_FDS/LISTEN_PID: - - uuidd DONE - - dbus DONE - - rsyslog DONE - - rpcbind (/var/run/rpcbind.sock!) DONE - - cups DONE - - avahi-daemon (/var/run/avahi-daemon/socket) DONE - - ssh CLASSIC + - uuidd HAVEPATCH + - rpcbind (/var/run/rpcbind.sock!) HAVEPATCH + - cups HAVEPATCH - postfix, saslauthd - apache/samba - libvirtd (/var/run/libvirt/libvirt-sock-ro) - bluetoothd (/var/run/sdp! @/org/bluez/audio!) - distccd -* fingerprint.target, wireless.target, gps.target +* fingerprint.target, wireless.target, gps.target, netdevice.target * set_put(), hashmap_put() return values check. i.e. == 0 doesn't free()! @@ -112,33 +196,21 @@ Features: * add systemctl switch to dump transaction without executing it -* suspend, resume +* suspend, resume support? * readahead: btrfs/LVM SSD detection -* when processes remain in a service even though the start command failed enter active - -* add seperate man page for [Install] settings - -* only add quotacheck deps to .mount units which mention grpquota/usrquota in the mount flags - -* systemctl condrestart should return 0 if service isn't running +* add separate man page for [Install] settings * allow runtime changing of log level and target -* automatically determine TERM= based on tty name even for /dev/console - http://git.kernel.org/?p=linux/kernel/git/gregkh/tty-2.6.git;a=commitdiff;h=fbc92a3455577ab17615cbcb91826399061bd789 - http://git.kernel.org/?p=linux/kernel/git/gregkh/tty-2.6.git;a=commitdiff;h=b7b8de087384cc1954a8cd075af3f9e5977caa2e - -* global defaults for StandardOuput=xxx - -Fedora: +External: -* chkconfig → systemd enable/daemon-reload glue +* udisks should not use udisks-part-id, instead use blkid. also not probe /dev/loopxxx -* /var/lock, /var/run → what happens if rpm stuff is no %ghosted? +* snd-seq should go, https://bugzilla.redhat.com/show_bug.cgi?id=676095 -External: +* gnome-shell python script/glxinfo/is-accelerated must die * make cryptsetup lower --iter-time @@ -146,8 +218,7 @@ External: * 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 - -* place /etc/inittab with explaining blurb. + (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