X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=blobdiff_plain;f=TODO;h=d1d7140c8d11b6ede8d48eb680c48aaa934b8ff4;hp=fe27d2b2a0b1dbf8e46a0ea22187b58121cb6bf1;hb=60211b35070a20ed0e78a83f39619139d56f7745;hpb=e41e194340f9a8dbd982b5030449281c3c8bf0dd diff --git a/TODO b/TODO index fe27d2b2a..d1d7140c8 100644 --- a/TODO +++ b/TODO @@ -11,31 +11,170 @@ Bugfixes: * properly handle .mount unit state tracking when two mount points are stacked one on top of another on the exact same mount point. -Fedora 19: +* stop importing kernel exported env variables. The utterly broken logic in + the kernel exports every kernel command line option which is not recognized + as a built-in module option as an env variable. Systemd should not pass-on + that nonsense, a kernel command line option is a command line option not an + env variable: + $ cat /proc/252/environ + initrd=\6a9857a393724b7a981ebb5b8495b9ea\3.10.0-2.fc20.x86_64\initrd -* external: maybe it is time to patch procps so that "ps" links to - libsystemd-logind to print a pretty service name, seat name, session - name in its output. Currently it only shows cgroup membership, but - that's sometimes kinda hard to parse for a human. +Fedora 20: -* cgroup attrs: - - update dbus interface docs in wiki +* external: ps should gain colums for slice and machine * localed: - localectl: support new converted x11→console keymaps +* when installing fedora with yum --installroot /var/run is a directory, not a symlink + https://bugzilla.redhat.com/show_bug.cgi?id=975864 + +CGroup Rework Completion: + +* introduce "mainpid" for scopes + +* implement system-wide DefaultCPUAccounting=1 switch (and similar for blockio, memory?) + +* implement per-slice CPUFairScheduling=1 switch + +* handle jointly mounted controllers correctly + +* logind: implement session kill exceptions + +* make BlockIODeviceWeight=, BlockIODeviceBandwidth= runtime settable + +* introduce high-level settings for RT budget, swappiness + +* wiki: document new bus APIs of PID 1 (transient units, Reloading signal) +* review: scope units, slice units, pid1, pam_system, systemctl commands + +* Send SIGHUP and SIGTERM in session scopes + Features: -* see if we can fix https://bugs.freedesktop.org/show_bug.cgi?id=63672 - without dropping the location cache entirely. +* given that logind/machined now let PID 1 do all nasty work we can + probably reduce the capability set they retain substantially. + +* btrfs raid assembly: some .device jobs stay stuck in the queue + +* Fedora: add an rpmlint check that verifies that all unit files in the RPM are listed in %systemd_post macros. + +* Fedora: post FPC ticket to move add %tmpfiles_create to the packaging guidelines + +* make sure gdm doesn't use multi-user-x but the new default X configuration file, and then remove multi-user-x from systemd + +* when parsing calendar timestamps support the UTC timezone (even if we won't support arbitrary timezone specs, support UTC itself certainly makes sense), also support syntaxes such as +0200 + +* when a kernel driver logs in a tight loop we should ratelimit that too. + +* "systemctl disable" of a unit instance removes all symlinks, should + only remove the instance symlink (systemctl disable of a template + unit however should remove them all). + +* journald: optionally, log debug messages to /run but everything else to /var + +* journald: optionally, when messages with a high log priority are logged, sync() immediately. + +* systemctl list-unit-files should list generated files (and probably with a new state "generated" for them, or so) + +* do we really need both hasprefix() and startswith()? + +* journald: when we drop syslog messages because the syslog socket is + full, make sure to write how many messages are lost as first thing + to syslog when it works again. + +* prohibit Restart= set with Type=oneshot + +* man: the documentation of Restart= currently is very misleading and suggests the tools from ExecStartPre= might get restarted. + +* load .d/*.conf dropins for device units + +* move systemctl set-log-level to systemd-analyze? + +* move "systemctl dump" to systemd-analyze + +* add a fixed dbus path for "my own unit", "my own session", ... to PID1, logind, ... + +* service_coldplug() appears to reinstall the wrong stop timeout watch? -* truncate whitespace of var names the same way as values in parse_env_file_internal() +* transient units: allow creating auxiliary units with the same call + +* how to reset dynamically changed attributes sanely? + +* when reloading configuration, apply new cgroup configuration + +* journald: make sure ratelimit is actually really per-service with the new cgroup changes + +* libsystemd-logind: sd_session_is_active() and friends: verify + validity of session name before appending it to a path + +* gparted needs to disable auto-activation of mount units somehow, or + maybe we should stop doing auto-activation of this after boot + entirely. https://bugzilla.gnome.org/show_bug.cgi?id=701676 + Maybe take a BSD lock at the disk device node and teach udev to + check for that and suppress event handling. + +* when a service changes state make reflect that in the + RUNNING/LISTENING states of its socket + +* when recursively showing the cgroup hierarchy, optionally also show + the hierarchies of child processes + +* document logic of auto/noauto and fail/nofail in fstab in systemd.mount or systemd-fstab-generator man page + +* we should properly escape ssh hostnames we add into dbus server strings + +* something pulls in pcre as so dep into our daemons such as hostnamed. + +* cgroup-agent: downgrade error messages + +* document systemd-journal-flush.service properly + +* change systemd-journal-flush into a service that stays around during + boot, and causes the journal to be moved back to /run on shutdown, + so that we don't keep /var busy. This needs to happen synchronously, + hence doing this via signals is not going to work. + +* allow implementation of InaccessibleDirectories=/ plus + ReadOnlyDirectories=... for whitelisting files for a service. + +* libsystemd-bus: + - default policy (allow uid == 0 and our own uid) + - enforce alignment of pointers passed in + - when kdbus doesn't take our message without memfds, try again with memfds + - implement translator service + - port systemd to new library + - implement busname unit type in systemd + - move to gvariant + - merge busctl into systemctl or so? + - synthesize sd_bus_message objects from kernel messages + - properly implement name registry ioctls for kdbus + - get rid of object hash table, use decision tree everyhwere instead? + - implement monitor logic + - object vtable logic + - longer term: + * priority queues + * priority inheritance + +* in the final killing spree, detect processes from the root directory, and + complain loudly if they have argv[0][0] == '@' set. + https://bugzilla.redhat.com/show_bug.cgi?id=961044 + +* add an option to nspawn that uses seccomp to make socket(AF_NETLINK, + SOCK_RAW, NETLINK_AUDIT) fail the the appropriate error code that + makes the audit userspace to think auditing is not available in the + kernel. + +* Introduce a way how we can kill the main process of a service with KillSignal, but all processes with SIGKILL later on + https://bugzilla.redhat.com/show_bug.cgi?id=952634 + +* maybe add a warning to the unit file parses where the access mode of unit files is nonsensical. + +* investigate endianness issues of UUID vs. GUID * dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we should be able to safely try another attempt when the bus call LoadUnit() is invoked. -* for instanced unit drop-ins we should look in foo@bar.service.d/ as well as foo@.service.d/ - * if pam_systemd is invoked by su from a process that is outside of a any session we should probably just become a NOP, since that's usually not a real user session but just some system code that just @@ -49,33 +188,16 @@ Features: * timedatctl, localectl: possibly make some commands work without the daemon, for chroot situations... -* logind: consider making suspend inhibitor locks non-session specific, but keep suspend-key inhibitor locks session specific - * logind: add Suspend() bus calls which take timestamps to fix double suspend issues when somebody hits suspend and closes laptop quickly. * we need dynamic units -* nss-myhostname should also resolve 'localhost' so that /etc/hosts becomes optional. - -* add s.th. like "systemctl set-log-level debug" - -* sd-login: allow enumerating machines and add inotify iface - -* cgroup-util: verify syntax of cgroup controllers - * cgtop: make cgtop useful in a container -* make sure cg_pid_get_path() works properly for co-mounted controllers - -* nspawn: ensure syntax of --uuid= argument is correct - -* explicitly disallow changing the cgroup path of units in the - name=systemd hierarchy, unless it is outside of /system - -* add 'set -e' to scripts in test/ -* make test in test/ work with separate output dir - -* suppress log output on shutdown when "quiet" is used +* test/: + - add 'set -e' to scripts in test/ + - make stuff in test/ work with separate output dir + - remove all the duplicated code in test/ * systemctl delete x.snapshot leaves no trace in logs (at least at default level). @@ -83,13 +205,6 @@ Features: so that the coredump is properly written to the user's own journal file. -* move /usr/lib/modules/$(uname -r)/modules.devname parsing from udevd to - kmod static-nodes - call kmod as an early service, and drop CAP_MKNOD from udevd.service - -* systemd-delta needs to be made aware of *.d/*.conf drop-in files for - units. - * seems that when we follow symlinks to units we prefer the symlink destination path over /etc and /usr. We shouldn't do that. Instead /etc should always override /run+/usr and also any symlink @@ -100,20 +215,10 @@ Features: * when isolating, try to figure out a way how we implicitly can order all units we stop before the isolating unit... -* teach udev + logind's uaccess to somehow handle the "dead" device nodes from: - /lib/modules/$(uname -r)/modules.devname - and apply ACLs to them if they have TAG=="uaccess" in udev rules. - -* matching against units is currently broken in journalctl. We really - need another AND level in the expressions, - i.e. sd_journal_add_conjunction(). - * add ConditionArchitecture= or so * teach ConditionKernelCommandLine= globs or regexes (in order to match foobar={no,0,off}) -* we should log capabilities too - * Support SO_REUSEPORT with socket activation: - Let systemd maintain a pool of servers. - Use for seamless upgrades, by running the new server before stopping the @@ -129,18 +234,12 @@ Features: * man: remove .include documentation, and instead push people to use .d/*.conf -* disallow .include from included files - -* safe_atod() is too naive, as it is vulnerable to locale parameters, should be locale independent. - * think about requeuing jobs when daemon-reload is issued? usecase: the initrd issues a reload after fstab from the host is accessible and we might want to requeue the mounts local-fs acquired through that automatically. -* neither pkexec nor sudo initialize environ[] from the PAM environment? - -* rework specifier logic so that we can distuingish OOM errors from other errors +* rework specifier logic so that we can distinguish OOM errors from other errors * systemd-inhibit: make taking delay locks useful: support sending SIGINT or SIGTERM on PrepareForSleep() @@ -148,32 +247,22 @@ Features: * remove any syslog support from log.c -- we probably can't do this before split-off udev is gone for good -* fedora: connect the timer units of a service to the service via Also= in [Install] - -* fedora: F20: go timer units all the way, leave cron.daily for cron +* documentation: recommend to connect the timer units of a service to the service via Also= in [Install] -* add a tool that lists active timer units plus their next elapstion and the time the units ran last +* add a tool that lists active timer units plus their next elapse and the time the units ran last * man: document the very specific env the shutdown drop-in tools live in * shutdown logging: store to EFI var, and store to USB stick? -* man: extend runlevel(8) to mention that runlevels suck, and are dead. Maybe add runlevel(7) with a note about that too +* man: extend runlevel(8) to mention that runlevels suck, and are dead. Maybe add runlevel(7) with a note about that too * systemctl: maybe add "systemctl add-wants" or so... -* man: add a link to socket activation blog from systemd.socket(5) - * man: add more examples to man pages * man: maybe sort directives in man pages, and take sections from --help and apply them to man too -* add "# export SYSTEMD_PAGER=" to bash login - -* /usr/bin/service should actually show the new command line - -* fedora: suggest auto-restart on failure, but not on sucess and not on coredump. also, ask people to think about changing the start limit logic. Also point people to RestartPreventExitStatus=, SuccessExitStatus= - * write UI tool that pops up emergency messages from the journal as notification * think about window-manager-run-as-user-service problem: exit 0 → activate shutdown.target; exit != 0 → restart service @@ -185,8 +274,6 @@ Features: * use "log level" rather than "log priority" everywhere -* ensure sd_journal_seek_monotonic actually works properly. - * timedate: have global on/off switches for auto-time (NTP), and auto-timezone that connman can subscribe to. * Honour "-" prefix for InaccessibleDirectories= and ReadOnlyDirectories= to @@ -206,7 +293,7 @@ Features: - write man page for efi boot generator - honor language efi variables for default language selection (if there are any?) - honor timezone efi variables for default timezone selection (if there are any?) - - introduce bootctl (backed by systemd-bootd) to control temporary and persistent default boot goal plus efi variables + - change bootctl to be backed by systemd-bootd to control temporary and persistent default boot goal plus efi variables * maybe do not install getty@tty1.service symlink in /etc but in /usr? @@ -256,7 +343,6 @@ Features: - check if we can make journalctl by default use --follow mode inside of less if called without args? - maybe add API to send pairs of iovecs via sd_journal_send - journal: when writing journal auto-rotate if time jumps backwards - - gatewayd: should run under its own UID - journal: add a setgid "systemd-journal" utility to invoke from libsystemd-journal, which passes fds via STDOUT and does PK access - journactl: support negative filtering, i.e. FOOBAR!="waldo", and !FOOBAR for events without FOOBAR. @@ -265,18 +351,15 @@ Features: - journal-send.c, log.c: when the log socket is clogged, and we drop, count this and write a message about this when it gets unclogged again. - journal: find a way to allow dropping history early, based on priority, other rules - journal: When used on NFS, check payload hashes - - journal: When used on NFS make sure wake up sd_journal_wait() every 2s, to handle missing inotify - - Introduce journalctl -b to show journal messages of a previous boot - journald: check whether it is OK if the client can still modify delivered journal entries - journal live copy, based on libneon (client) and libmicrohttpd (server) - journald: add kernel cmdline option to disable ratelimiting for debug purposes - refuse taking lower-case variable names in sd_journal_send() and friends. - journald: we currently rotate only after MaxUse+MaxFilesize has been reached. - journal: deal nicely with byte-by-byte copied files, especially regards header - - journalctl: show multiline log messages sanely, expand tabs, and show all valid utf8 messages + - journalctl: expand tabs - journal: store euid in journal if it differs from uid - journal: sanely deal with entries which are larger than the individual file size, but where the components would fit - - journalctl: make journalctl smarter, and actually check groups that have access to /var/log/journal before printing message about recomming group membership for journal access - Replace utmp, wtmp, btmp, and lastlog completely with journal - Port upower to use the journal for historical power information used in future calculations @@ -305,9 +388,6 @@ Features: - systemctl: "Journal has been rotated since unit was started." message is misleading - support "systemctl stop foobar@.service" to stop all units matching a certain template - Something is wrong with symlink handling of "autovt@.service" in "systemctl list-unit-files" - - add configuration/switches to use - freeze (http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git) and - standby (https://bugs.freedesktop.org/show_bug.cgi?id=57793) as suspend modes * introduce ntp.service (or suchlike) as symlink that is used to arbitrate between various NTP implementations @@ -330,10 +410,6 @@ Features: mode, it will never touch the RTC if the no reliable time source is active or the user did not request anything like it. -* hwdb: - - implement conditional properties (dmi matches) - - hwdb --filter=ID_DRIVE_* - * if booted in "quiet" mode, and an error happens, turn on status output again, so that the emergency mode isn't totally surprising. Also, terminate plymouth. @@ -380,8 +456,6 @@ Features: * currently x-systemd.timeout is lost in the initrd, since crypttab is copied into dracut, but fstab isn't -* WorkingDirectory: support env var replacements like in ExecStart= so that people can use $HOME - * refuse boot if /etc/machine-id is not useful (or set taint?) * nspawn: @@ -398,11 +472,10 @@ Features: - nspawn: make it work for dwalsh and shared /usr containers -- tmpfs mounts as command line parameters, selinux exec context * cryptsetup: - - cryptsetup-generator: warn if the password files are world-readable - cryptsetup-generator: allow specification of passwords in crypttab itself - move cryptsetup key caching into kernel keyctl? https://bugs.freedesktop.org/show_bug.cgi?id=54982 - - when key file cannot be found, read it from kbd in cryptsetup + - support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator * move debug shell to tty6 and make sure this doesn't break the gettys on tty6 @@ -443,8 +516,6 @@ Features: when done. That means clients don't get a successful method reply, but much rather a disconnect on success. -* remember which condition failed for services, not just the fact that something failed - * use opterr = 0 for all getopt tools * properly handle loop back mounts via fstab, especially regards to fsck/passno @@ -471,14 +542,8 @@ Features: * introduce Type=pid-file -* maybe allow services with ExecStop= set, but no ExecStart=? - * change Requires=basic.target to RequisiteOverride=basic.target -* support rd.luks.allow-discards= kernel cmdline params in cryptsetup generator - -* drop accountsservice's StandardOutput=syslog and Type=dbus fields - * when breaking cycles drop sysv services first, then services from /run, then from /etc, then from /usr * move passno parsing to fstab generator @@ -498,8 +563,6 @@ Features: * ExecOnFailure=/usr/bin/foo -* ConditionSecurity= should learn about IMA and SMACK - * udev: - remove src/udev/udev-builtin-firmware.c (CONFIG_FW_LOADER_USER_HELPER=n) - move to LGPL @@ -509,8 +572,6 @@ Features: * cleanup syslog 'priority' vs. 'level' wording -* dbus upstream still refers to dbus.target and shouldn't - * when a service has the same env var set twice we actually store it twice and return that in systemctl show -p... We should only show the last setting * support container_ttys= @@ -525,8 +586,6 @@ Features: * default unix qlen is too small (10). bump sysctl? add sockopt? -* dbus: in fedora, make /var/lib/dbus/machine-id a symlink to /etc/machine-id - * save coredump in Windows/Mozilla minidump format * support crash reporting operation modes (https://live.gnome.org/GnomeOS/Design/Whiteboards/ProblemReporting) @@ -609,8 +668,6 @@ Features: * support systemd.mask= on the kernel command line. -* reuse mkdtemp namespace dirs in /tmp? - * recreate systemd's D-Bus private socket file on SIGUSR2 * Support --test based on current system state @@ -641,6 +698,12 @@ Features: works with ^C - add documentation to systemd.daemon +* bootchart: + - plot per-process IO utilization + - group processes based on service association (cgroups) + - document initcall_debug + - kernel cmdline "bootchart" option for simplicity? + External: * dbus: @@ -666,12 +729,23 @@ External: * kernel: add device_type = "fb", "fbcon" to class "graphics" -* bootchart: - - plot per-process IO utilization - - group processes based on service association (cgroups) - - document initcall_debug - - put bootcharts in the journal - - kernel cmdline "bootchart" option for simplicity? +* drop accountsservice's StandardOutput=syslog and Type=dbus fields + +* dbus upstream still refers to dbus.target and shouldn't + +* dbus: in fedora, make /var/lib/dbus/machine-id a symlink to /etc/machine-id + +* add "# export SYSTEMD_PAGER=" to bash login + +* /usr/bin/service should actually show the new command line + +* fedora: suggest auto-restart on failure, but not on success and not on coredump. also, ask people to think about changing the start limit logic. Also point people to RestartPreventExitStatus=, SuccessExitStatus= + +* fedora: F20: go timer units all the way, leave cron.daily for cron + +* neither pkexec nor sudo initialize environ[] from the PAM environment? + +* fedora: update policy to declare access mode and ownership of unit files to root:root 0644, and add an rpmlint check for it Regularly: