Commit Graph

1739 Commits

Author SHA1 Message Date
cat 6d14bb814f container/fhs: add constant for /dev/shm/
This is mounted for the default read-only /dev/ when programs want to use shm_open(3). Defining it here is less error-prone and saves the extra append at runtime.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-14 01:03:26 +09:00
cat be0e387ab0 internal/info: relocate from internal
This is cleaner and makes more sense. The longer LDFLAGS was never a valid concern since it is always inserted by a script.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-13 07:29:46 +09:00
cat abeb67964f treewide: document linkname uses
These provide justification for each use of linkname. Poorly thought out uses of linkname are removed.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-13 07:14:16 +09:00
cat bf5d10743f treewide: import internal/system
For #24.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-13 01:22:47 +09:00
cat 4e7aab07d5 internal/system: relocate from system
These packages are highly specific to hakurei and are difficult to use safely from other pieces of code.

Their exported symbols are made available until v0.4.0 where they will be removed for #24.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-13 01:17:47 +09:00
cat 15a66a2b31 treewide: import internal/helper
For #24.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-12 23:19:34 +09:00
cat f347d44c22 internal/helper: relocate from helper
This package is ugly and is pending removal only kept alive by xdg-dbus-proxy.

Its exported symbols are made available until v0.4.0 where it will be removed for #24.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-12 23:16:13 +09:00
cat b5630f6883 test: move package sandbox internal
This should never be used outside vm tests.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-12 23:03:22 +09:00
cat 17ffdb2dcf release: 0.3.1
Signed-off-by: Ophestra <cat@gensokyo.uk>
v0.3.1
2025-11-12 00:53:14 +09:00
cat ac34635890 container: set FD_CLOEXEC on all open files
While fd created from this side always has the FD_CLOEXEC flag, the same is not true for files left open by the parent. This change prevents those files from leaking into the container.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-12 00:18:29 +09:00
cat 9dec9dbc4b container/init: close setup pipe early
This prevents leaking the setup pipe.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-11 07:31:58 +09:00
cat 2f74adc8bd container/init: close initial process files on termination
This closes them during the adopt wait delay. This also keeps them alive.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-10 20:35:59 +09:00
cat d7e0104ae4 treewide: reject impossible user-supplied fd
These are all trusted user input, however this check reduces the likelihood of hard to debug errors.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-10 20:31:26 +09:00
cat bb92e3ada9 cmd/hakurei: expose current instance identifier
This writes the 16-byte instance identifier to file descriptor specified by --identifier-fd if set, and closes the file.

This enables safely obtaining the new instance's identifier.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-10 07:52:35 +09:00
cat fad419c2a2 internal/outcome: handle group lookup message
This results in slightly less messy error reporting.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-09 02:15:48 +09:00
cat b1a1e73238 nix: update names to reflect new terminology
These are terminology from way early days. Update them now to be less confusing.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-09 01:48:31 +09:00
cat 38e9128a8c container/std/seccomp: remove ineffectual typecast
This is no longer necessary since the return type changed.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-07 05:45:51 +09:00
cat 7ee702a44e container/seccomp/presets: add fields to literals
This keeps composites analysis happy.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-07 05:11:57 +09:00
cat 3d188ef884 std: separate seccomp constants
This avoids inadvertently using PNRs as syscall numbers.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-07 04:30:06 +09:00
cat 34ccda84b2 release: 0.3.0
Signed-off-by: Ophestra <cat@gensokyo.uk>
v0.3.0
2025-11-06 01:37:15 +09:00
cat 042013bb04 container/std: syscall JSON adapter
This provides cross-platform JSON adapter for syscall number.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-06 00:57:53 +09:00
cat 5c2b63a7f1 container: add 386 constants
While it is unlikely a use case for hakurei on i686 exists, it does not hurt to have this support.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 20:21:14 +09:00
cat 9fd97e71d0 treewide: fit test untyped int literals in 32-bit
This enables hakurei test suite to run on 32-bit targets.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 20:13:19 +09:00
cat fba201c995 container/std: relocate rule types
This enables its use in hst for #15.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 06:00:39 +09:00
cat 7f27a6dc51 container/seccomp: use native types
This prepares NativeRule for relocation to std for #15.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 05:48:59 +09:00
cat b65aba9446 container/seccomp: alias libseccomp types
This enables tests to refer to these types and check its size.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 05:21:43 +09:00
cat becaf8b6d7 std: relocate seccomp lookup tables
This should enable resolving NativeRule in hst.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 04:48:05 +09:00
cat 54c0d6bf48 container/seccomp/pnr: define pseudo syscalls
This eliminates the cgo dependency from syscall lookup.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 04:32:41 +09:00
cat c1399f5030 std: rename from comp
Seccomp lookup tables are going to be relocated here, and PNR constants.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 02:47:43 +09:00
cat 9ac63aac0c hst/grp_pwd: add extra test cases
Does not change coverage but this helps me crosscheck with my phone.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-05 01:42:42 +09:00
cat cb9ebf0e15 hst/grp_pwd: specify new uid format
This leaves slots available for additional uid ranges in Rosa OS.

This breaks all existing installations! Users are required to fix ownership manually.

Closes #18.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-04 08:24:41 +09:00
cat 9a2a7b749f cmd/hakurei/print: handle nil config
There is nothing to print in this case, and such a nil check is missing.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 02:20:18 +09:00
cat ec5cb9400c cmd/hpkg/test: print share directory
This is more useful now that state is tracked here.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 01:51:57 +09:00
cat ae66b3d2fb message: rename NewMsg to New
Should have done this when relocating this from container. Now is a good time to rename it before v0.3.x.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 01:49:27 +09:00
cat 149bc3671a internal/store: remove compat adapter
This is no longer used as everything has been migrated.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 01:26:01 +09:00
cat 24435694a5 hst/config: make identifier omitempty
This is an optional field. Serialise it as such.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 01:23:15 +09:00
cat 1c168babf2 cmd/hakurei/print: use new store interface
This removes the final uses of the compat interfaces.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-03 01:19:16 +09:00
cat 0edcb7c1d3 test: print share directory
This is more useful now that state is tracked here.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 17:00:59 +09:00
cat 0e5ca74b98 cmd/hakurei/print: serialise array for ps
Wanted to do this for a long time, since the key is redundant. This also makes it easier to migrate to the new store interface.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 16:37:08 +09:00
cat 23ae7822bf cmd/hakurei/parse: use new store interface
This greatly reduces overhead. The iterator also significantly cleans up the usage code.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 16:00:41 +09:00
cat 898b5aed3d internal/store: iterator over all entries
This is quite convenient for searching the store or printing active instance information.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 15:54:00 +09:00
cat 7c3c3135d8 internal/outcome: track state in TMPDIR
The SharePath is a more stable path than RunDirPath, since it is available all the time and should remain consistent. This also fits better into the intended use case of XDG_RUNTIME_DIR.

Closes #17.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 12:40:58 +09:00
cat f33aea9ff9 internal/env: cleaner runtime dir fallback
This now places rundir inside the fallback runtime dir, so special case in internal/outcome is avoided.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 12:22:32 +09:00
cat e7fc311d0b internal/outcome/shim: cover reparent and exit request paths
These test cases were missed when making the changes.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 11:58:09 +09:00
cat f5274067f6 internal/outcome/process: nil-safe unlock when failing to lock
This also prints a debug message which might be useful.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 11:47:51 +09:00
cat e7161f8e61 internal/outcome: measure finalise time
This also increases precision of state time output.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 05:17:33 +09:00
cat 6931ad95c3 internal/outcome/shim: EOF as exit request fallback
In some cases the signal might be delivered before the signal handler is installed, and synchronising against such a case is too expensive. Instead, use the pipe being closed as a fallback to the regular exit request. This change also moves installation of the signal handler early.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 04:41:26 +09:00
cat 2ba599b399 internal/outcome/process: use new store interface
This change also spawns shim before committing system state, leaving it blocking on the setup pipe. The internal/outcome/process structure is also entirely reworked to be much more readable and less error-prone, while enabling basic performance measurements. A long-standing bug where segment lock is not held during Commit is also resolved.

Closes #19.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-02 04:25:45 +09:00
cat d3d3417125 internal/outcome/process: relocate start and serve
This is useful for reordering these operations for further cleanup.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-11-01 19:14:59 +09:00
cat 651cdf9ccb internal/outcome: remove guard on main
This is no longer exported. Such a check is pointless.

Signed-off-by: Ophestra <cat@gensokyo.uk>
2025-10-31 22:58:26 +09:00