Commit Graph

176 Commits

Author SHA1 Message Date
648e1d641a app: separate interface from implementation
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 23:07:28 +09:00
ffaa12b9d8 sys: wrap log methods
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 22:52:09 +09:00
e0f321b2c4 sys: rename from linux
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 18:47:48 +09:00
2c9c7fee5b linux: wrap fsu lookup error
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 17:39:53 +09:00
d0400f3c81 fmsg: PrintBaseError skip empty message
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 17:01:26 +09:00
e9b0f9faef fmsg: export logBaseError function
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-18 13:02:51 +09:00
90cb01b274 system: move out of internal
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-17 19:00:43 +09:00
b1e1d5627e system: wrap console output functions
This eliminates all fmsg imports from internal/system.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-17 18:17:55 +09:00
3ae2ab652e system/wayland: sync file at caller specified address
Storing this in sys is incredibly ugly: sys should be stateless and Ops must keep track of their state.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-17 13:24:17 +09:00
db71fbe22b system/tmpfiles: fail gracefully in API misuse
Panicking here leaves garbage behind. Not ideal if this package is going to be exported.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-17 12:17:01 +09:00
82a072f641 system/tmpfiles: implement private tmpfiles
These are only available within the mount namespace and should significantly reduce attack surface.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-17 00:07:52 +09:00
468696f611 internal: beforeExit before reachable fatal calls
These are the only two calls to log.Fatal* reachable during suspended output. Call fmsg.BeforeExit here to catch that.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-16 19:03:34 +09:00
29c38caac8 app/shim/manager: return error on bad fsu path
This results in a graceful failure that does not leave garbage behind.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-16 18:59:45 +09:00
e599b5583d fmsg: implement suspend in writer
This removes the requirement to call fmsg.Exit on every exit path, and enables direct use of the "log" package. However, fmsg.BeforeExit is still encouraged when possible to catch exit on suspended output.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-16 18:51:53 +09:00
33a4ab11c2 internal: move shim and init into app
This structure makes more sense, as both processes are part of an app's lifecycle.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-16 16:28:46 +09:00
c667b13a00 system: separate link Op implementation
This Op would still be useful after replacing the Tmpfiles interface, so isolate it here.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-16 12:15:26 +09:00
268a90f1a5 app: improve WAYLAND_DISPLAY correctness
This now has identical behaviour as wayland C library.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 14:45:09 +09:00
ddb2f9c11b app: remove wayland socket hard link
This Op was not doing anything useful.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 10:54:00 +09:00
f955b15b84 system: remove write mode tmpfiles
This interface is ugly and bug-prone. This change removes its write mode which has been obsoleted by CopyBind.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 03:22:20 +09:00
0340c67995 app: port passwd and group files to copy
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 03:19:06 +09:00
ea8d1c07df priv/shim: move /sbin/init setup to app
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 03:06:10 +09:00
a0062d8275 fmsg: resume on exit
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-15 02:22:09 +09:00
1f74b636d3 state/join: use Join method when available
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-14 14:11:02 +09:00
e431ab3c24 app: check username length against LOGIN_NAME_MAX
This limit is arbitrary, but it's good to enforce it anyway.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-14 12:44:55 +09:00
fe7d208cf7 helper: use generic extra files interface
This replaces the pipes object and integrates context into helper process lifecycle.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-13 23:34:15 +09:00
e14923ae53 helper/proc: move package out of internal
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-08 13:03:45 +09:00
a48386bd56 system/dbus: dump messages on early fault
In the current app implementation this gets dumped in the wait method after resuming output. Wait is never called in an early fault condition, so any error messages get lost.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-07 13:20:56 +09:00
2e52191404 system/dbus: dump method prints msgbuf
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-02-07 13:16:54 +09:00
163f15e93f helper/seccomp: separate seccomp package
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-25 12:59:11 +09:00
23e1152baa app/share: clean BaseError message
This removes trailing '\n' in the PulseAudio warning.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-22 11:54:16 +09:00
8c51012ef5 dbus: enable syscall filter
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-22 11:49:23 +09:00
9a239fa1a5 helper/bwrap: integrate seccomp into helper interface
This makes API usage much cleaner, and encapsulates all bwrap arguments in argsWt.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-22 01:52:57 +09:00
82029948e6 proc: append to ExtraFiles slice pointer
This is useful for initialising extra files before command.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-21 12:51:39 +09:00
dfcdc5ce20 state: store config in separate gob stream
This enables early serialisation of config.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-21 12:10:58 +09:00
20a3d4c458 proc/priv/shim: resolve and load seccomp rules
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-20 23:52:56 +09:00
3df344828f proc/priv/shim: seccomp bpf filter via libseccomp
Rulesets adapted from Flatpak for compatibility.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-20 23:39:47 +09:00
27f5922d5c fst: include syscall filter configuration
This value is passed through to shim.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-20 21:12:39 +09:00
3c55fc8e86 proc/priv/shim: do not log bwrap args
This message is very long and does not serve much real purpose. Remove it to de-clutter verbose messages.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-20 19:51:28 +09:00
eb0ef2d115 helper/bwrap: generic extra file interface
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-20 00:20:04 +09:00
2f70506865 helper/bwrap: move sync to helper state
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-19 18:38:13 +09:00
cae567c109 proc/priv/shim: remove unnecessary state
These values are only used during process creation.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-19 18:09:07 +09:00
b31d055e20 proc/priv/init: early init check
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-18 12:33:33 +09:00
7baca66a56 proc: remove duplicate compile-time fortify reference
This is no longer needed since shim and init are now part of the main program.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-18 11:59:33 +09:00
27d2914286 proc/priv/init: merge init into main program
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-18 11:47:01 +09:00
ea8f228af3 proc/priv/shim: merge shim into main program
Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-17 23:43:32 +09:00
16db3dabe2 internal: do PR_SET_PDEATHSIG once
This prctl affects the entire process, doing it on every OS thread is pointless.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-17 23:08:46 +09:00
124743ffd3 app: expose single run method
App is no longer just a simple [exec.Cmd] wrapper, so exposing these steps separately no longer makes sense and actually hinders proper error handling, cleanup and cancellation. This change removes the five-second wait when the shim dies before receiving the payload, and provides caller the ability to gracefully stop execution of the confined process.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-15 23:39:51 +09:00
562f5ed797 fst: hide sockets exposed via Filesystem
This is mostly useful for permissive defaults.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-15 10:13:18 +09:00
6acd0d4e88 linux/std: handle fsu exit status 1
Printing "exit status 1" is confusing. This handles the ExitError and returns EACCES instead.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-01-01 21:34:57 +09:00
c4d6651cae update reverse-DNS style identifiers
Signed-off-by: Ophestra <cat@gensokyo.uk>
2024-12-31 16:16:38 +09:00