1362 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
Ophestra	4862e8b74d	nix: configure sharefs as a mount unit ... This should have been configured via fileSystem instead, but that somehow has no effect, it is not even being evaluated, and I am not wasting any more time on nix. This change aims to keep everything working until all nix/systemd tooling is replaced. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 20:57:10 +09:00
Ophestra	ef1ebf12d9	cmd/sharefs: handle mount -t fuse.sharefs ... This should have been handled in a custom option parsing function, but that much extra complexity is unnecessary for this edge case. Honestly I do not know why libfuse does not handle this itself. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 20:49:27 +09:00
Ophestra	775a9f57c9	cmd/sharefs: check option parsing behaviour ... This change makes it possible to check parseOpts behaviour as part of Go tests. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 17:33:12 +09:00
Ophestra	2f8ca83376	cmd/sharefs: containerise filesystem daemon ... This replaces the forking daemonise libfuse function which prevents Go callbacks from calling into the runtime. This also enforces least privilege on the daemon process. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 10:16:35 +09:00
Ophestra	3d720ada92	container: optionally allow orphan ... This is required for the typical daemonise use case. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 09:12:02 +09:00
Ophestra	2e5362e536	cmd/sharefs: opaque setup state ... This allows unrestricted use of the type system and prepares setup code for cross-process initialisation. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 04:14:00 +09:00
Ophestra	6d3bd27220	cmd/sharefs: expand fuse_main ... This change should not change behaviour other than making output more consistent. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 02:30:28 +09:00
Ophestra	a27305cb4a	cmd/sharefs: improve help message ... This improves consistency with the fuse_main help message. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-27 02:20:41 +09:00
Ophestra	0e476c5e5b	cmd/sharefs: allocate sharefs_private early ... This also removes global state used by sharefs_init. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 08:08:41 +09:00
Ophestra	54712e0426	nix: set noatime on sharefs ... Could improve performance, atime is not useful for this filesystem anyway. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 05:34:05 +09:00
Ophestra	b77c1ecfdb	cmd/sharefs/test: check option handling ... This verifies behaviour related to setuid/setgid when starting as root. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 05:28:45 +09:00
Ophestra	dce5839a79	nix: do not restart sharefs ... This avoids disrupting running containers. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 04:12:14 +09:00
Ophestra	d597592e1f	cmd/sharefs: rename fuse-helper to fuse-operations ... This is not really just library wrapper functions, but instead implements the callbacks, so fuse-operations makes more sense. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 03:19:32 +09:00
Ophestra	056f5b12d4	cmd/sharefs: move translate_pathname body to macro wrapper ... This is never called directly anywhere and it is simple enough to be included in the macro. This avoids passing the pointer around and dereferencing errno location, resulting in over 5% increase in throughput on the clang build. No change in the gcc build though. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-26 02:10:59 +09:00
Ophestra	da2bb546ba	cmd/sharefs: remove readlink ... This filesystem does not support symbolic links, so readlink is not useful, and unreachable in this case because of the check in getattr. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-25 06:00:58 +09:00
Ophestra	7bfbd59810	cmd/sharefs: implement shared filesystem ... This is for passing files between applications, similar to android /sdcard. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-25 05:13:02 +09:00
Ophestra	ea815a59e8	nix: disable source fortification in devShell ... This generates warnings when compiling without optimisation. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-21 02:22:28 +09:00
Ophestra	28a8dc67d2	internal/pipewire: raise Core::Sync timeout ... Hopefully relieves spurious failures on a very overloaded system. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-19 00:49:33 +09:00
Ophestra	ec49c63c5f	internal/pipewire: EPOLL_CTL_ADD instead of EPOLL_CTL_MOD ... Implementation is no longer tied down by the limitations of SyscallConn. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-19 00:43:44 +09:00
Ophestra	5a50bf80ee	internal/pipewire: hold socket fd directly ... The interface provided by net is not used here and is a leftover from a previous implementation. This change removes it. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-19 00:28:24 +09:00
Ophestra	ce06b7b663	internal/pipewire: inform conn of blocking intent ... The interface does not expose underlying kernel notification mechanisms. This change removes the need to poll in situations were the next call might block. This is made cumbersome by the SyscallConn interface left over from a previous implementation, it will be replaced in a later commit as the current implementation does not make use of any net.Conn methods other than Close. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-19 00:00:33 +09:00
Ophestra	08bdc68f3a	internal/pipewire: sendmsg/recvmsg errors are fatal ... When returned wrapped as a syscall error, these are impossible to recover from, so wrap them as a fatal error. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-18 23:33:12 +09:00
Ophestra	8cb0b433b2	release: 0.3.3 ... Signed-off-by: Ophestra <cat@gensokyo.uk> v0.3.3	2025-12-15 20:34:45 +09:00
Ophestra	767f1844d2	test: check shim private dir cleanup ... This asserts that no shim private dir was left behind after all containers terminate. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-15 20:30:19 +09:00
Ophestra	54610aaddc	internal/outcome: expose pipewire via pipewire-pulse ... This no longer exposes the pipewire socket to the container, and instead mediates access via pipewire-pulse. This makes insecure parts of the protocol inaccessible as explained in the doc comment in hst. Closes #29. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-15 12:57:06 +09:00
Ophestra	2e80660169	internal/outcome: look up pipewire-pulse path ... This is for setting up the pipewire-pulse container in shim, for #29. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-15 12:38:39 +09:00
Ophestra	d0a3c6a2f3	internal/outcome: optional shim private dir ... This is a private work directory owned by the specific shim. Useful for sockets owned by this instance of the shim and requires no direct assistance from the priv-side process. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-15 12:32:46 +09:00
Ophestra	0c0e3d6fc2	hst: add direct hardware option ... This is unfortunately the only possible setup to securely expose PipeWire to the container. Further explanation explained in the doc comment and #29. This will be implemented in a future commit. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-15 12:29:32 +09:00
Ophestra	fae910a1ad	container: sync stubbed wait4 loop after notify ... This ensures consistent state observed by wait4 loop when running against stub. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 10:22:48 +09:00
Ophestra	178c8bc28b	internal/pipewire: handle SecurityContext::Create error ... This method can result in an error targeting it, so it is handled here. This change also causes a call to Create to also Core::Sync, as it should have done. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 09:41:28 +09:00
Ophestra	30dcab0734	internal/pipewire: SecurityContext as destructible ... This proxy can be destroyed by sending a Core::Destroy targeting it. This change implements the Destroy method by embedding destructible. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 09:31:50 +09:00
Ophestra	0ea051062b	internal/pipewire: reorder context struct ... This change reorders and groups struct elements. This improves readability since this struct holds a lot of state loosely related to each other. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 09:26:30 +09:00
Ophestra	b0f2ab6fff	internal/pipewire: implement Core::Destroy ... This change also implements pending destructible check on Sync. Destruction method should always be implemented as a wrapper of destructible.destroy. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 09:20:58 +09:00
Ophestra	00a5bdf006	internal/pipewire: do not emit None for spa_dict ... Turns out the PipeWire server does not expect a value of type None here at all. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 09:06:44 +09:00
Ophestra	a27dfdc058	internal/pipewire: implement Core::CreateObject ... Nothing uses this right now, this would have to be called by wrapper methods on Registry that would search the objects Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 08:10:57 +09:00
Ophestra	6d0d9cecd1	internal/pipewire: handle nil spa_dict correctly ... This now marshals into a value of type None when the slice is nil, and correctly unmarshals from type None. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 07:12:00 +09:00
Ophestra	17248d7d61	internal/pipewire: unmarshal nil pointer correctly ... This now calls unmarshalCheckTypeBounds to advance to the next message. Additionally, handling for None value is relocated to a function for reuse by other types. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 06:58:53 +09:00
Ophestra	41e5628c67	internal/pipewire: return correct size for nil spa_dict ... A nil spa_dict results in a None type value being sent over the wire. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 06:07:46 +09:00
Ophestra	ffbec828e1	internal/pipewire: move Core wrapper methods under Core ... These do not belong under Context, and is an early implementation limitation that carried over. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-14 05:37:21 +09:00
Ophestra	de0467a65e	internal/pipewire: treat noAck violation as fatal ... Receiving this event indicates something has gone terribly wrong somehow, and ignoring Core::BoundProps causes inconsistent state anyway. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-13 22:30:06 +09:00
Ophestra	b5999b8814	internal/pipewire: implement Core::RemoveId ... This is emitted by the server when a proxy id is removed for any reason. Currently, the only path for this to be emitted is when a global object is destroyed while some proxy is still bound to it. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-13 21:50:32 +09:00
Ophestra	ebc67bb8ad	nix: update flake lock ... NixOS 25.11 introduces a crash in cage and an intermittent crash in foot. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-12 08:39:55 +09:00
Ophestra	e60ff660f6	internal/pipewire: treat unknown opcode as fatal ... Skipping events can cause local state to diverge from the server. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-11 04:22:03 +09:00
Ophestra	47db461546	internal/pipewire: generic Core::Error handling ... This flushes message buffer before queueing the event expecting the error. Since this is quite useful and relatively complex, it is relocated to a method of Context. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-11 04:07:55 +09:00
Ophestra	0a3fe5f907	internal/pipewire: export Registry::Destroy ... This handles the error returned by Sync. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-11 03:34:33 +09:00
Ophestra	b72d502f1c	internal/outcome: populate instance metadata for PipeWire ... These have similar semantics to equivalent Wayland security-context-v1 fields. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-10 03:01:30 +09:00
Ophestra	f8b3db3f66	internal/pipewire: cleaner error message for unsupported type ... The error string itself is descriptive enough, so use it as the error message directly. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-10 01:51:06 +09:00
Ophestra	0e2fb1788f	internal/pipewire: implement Registry::Destroy ... This requires error handling infrastructure in Core that does not yet exist, so it is not exported for now. It has been manually tested via linkname against PipeWire. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-10 01:25:30 +09:00
Ophestra	d8417e2927	internal/pipewire: implement Registry::GlobalRemove ... This is emitted by PipeWire when a global object disappears, because PipeWire insists that all clients that had called Core::GetRegistry must constantly sync its local registry state with the remote. Signed-off-by: Ophestra <cat@gensokyo.uk>	2025-12-10 01:00:03 +09:00
Ophestra	ccc0d98bd7	release: 0.3.2 ... Signed-off-by: Ophestra <cat@gensokyo.uk> v0.3.2	2025-12-09 08:12:52 +09:00