8587661684
feat(init): spawn service over IPC with per-client allowance refunded on child death; up to 4 death watchers; spawntest on test image
2026-10-27 16:21:08 +01:00
817a38993b
feat(kernel): WatchDeath + SetTimer events as notification signals; adapter wakes at earliest deadline
2026-10-27 16:06:04 +01:00
a4f467f45d
feat(userspace): log server sole console holder; init boots from manifest; adapter loads only init+storage
2026-10-27 15:40:19 +01:00
4049ba5007
feat(kernel): per-process budgets (pages, endpoints, notifs, regions) carved from creator, returned on death; init owns pool
2026-10-27 15:08:58 +01:00
7415bab7f5
feat(kernel): console cap gates DebugPutchar, no uart fast path; spawn start arg in a0; delegate returns slot
2026-10-27 14:54:36 +01:00
33e1c849d1
feat(kernel): ipc cap transfer drops Delegate; one owner per irq line, freed when owner dies
2026-10-27 14:42:28 +01:00
37c482c36c
chore: fix build warnings (unused vars/imports, dead ext4 param/field/fn)
2026-10-27 14:12:50 +01:00
b4390b7b6d
refactor(userspace): split syscall/fat16/ext4 by concern; trim comments
2026-10-27 13:21:53 +01:00
99273db6d3
refactor(adapter): split trap/page_table/smp/entry by concern; trim comments
2026-10-27 13:21:53 +01:00
95dbe7b9d5
refactor(kernel): split ipc/memory/lib/thread by concern; trim comments
2026-10-27 13:21:53 +01:00
b814bd9570
test: on-hardware SMP self-test (revoke/kill must stop every hart) + make test-qemu
2026-10-27 12:54:42 +01:00
909b3f0503
feat(storage): full file protocol over per-client shared region (badged clients, owned handles, clamped lengths, ReadOnly on unwritable disk); init loads via libkos::fs; drop boot shared buffer
2026-10-27 12:42:37 +01:00
694c675a40
feat: devices via caps (BootInfo, init assigns devices, drivers map MMIO/DMA themselves, IRQ-driven input); drop boot device wiring and info pages; headless display fallback
2026-10-27 12:28:39 +01:00
8bc5402a29
fix(userspace): clip display rects, bounds-check session app ids and fat16 handles, reject bad fat16 geometry, session kills via process cap
2026-10-27 12:11:13 +01:00
e8a2963b24
feat(userspace): Call/ReplyRecv IPC, malformed-request replies, servers exit on dead endpoint; drop send/receive compat, select!, async_syscall
2026-10-27 12:06:10 +01:00
2a8a5b0c27
perf(kernel): linked cap tree, translate-based page lookup, next-fit frames/heap, targeted dequeue; host benches
2026-10-27 01:33:26 +01:00
19aeabd86d
feat: SMP stage 2, lock-free fast paths (GetTime, GetProcessId, DebugPutchar, no-op Yield), storm checks yield equivalence
2026-10-27 01:07:57 +01:00
2e665fd66e
feat: SMP stage 1 (4 harts, per-hart guarded stacks, home-core affinity, IPIs, TLB shootdown, one temporary kernel lock), 4-core storm
2026-10-27 00:56:15 +01:00
b87a62475f
feat(adapter): higher-half kernel (linked at PHYS_OFFSET, boot trampoline, no low mappings), fix SBI ecall clobbers
2026-10-27 00:25:20 +01:00
55758ae5b2
feat: kernel W^X (per-section perms, firmware unmapped, stack guard, boot check), halt on kernel faults, region W^X across processes
2026-10-27 00:13:00 +01:00
bafc18a5f8
feat(kernel): SystemControl cap for halt, board device table + boot caps to init, RegionAddress syscall, idle loop services irqs
2026-10-26 23:59:57 +01:00
130a82f50d
feat(kernel): shared ABI decoder+proofs, no-panic gate, syscall storm, kani on shrunk kernel
2026-10-26 23:44:09 +01:00
154599e3af
feat(kernel): derivation-tree revocation, region caps
...
- caps record parent (mint/delegate/ipc), CapRef w/ pid+slot gens
- Revoke = delete subtree, keep self; drop/death re-parents children
- CreateRegion/MapRegion; mapping lives only as long as its cap
- drop Permission::Revoke + object-wide revoke
- mint any cap (badge only ep/ntfn), process caps delegable
- kani: descendants == subtree; model-based fuzz vs spec
2026-10-26 20:41:03 +01:00
76f10ff1f1
feat(kernel): seL4-style IPC, notifications, status ABI
...
- badges + Mint, Call/Reply/ReplyRecv w/ gen-checked reply slot
- cap transfer (1/msg, copy, needs Delegate, rechecked at delivery)
- notifications: signal/wait/poll/bind, irq -> notification routing
- WaitAny/WaitForInterrupt removed; a0=status, ErrorCode
- libkos: single ecall primitive, compat wrappers + typed new api
- fix: trap return clobbered user t0 on every trap
- fix: wait queues FIFO, delegate keeps badge, fault cap checked
- docs/ipc.md spec, ipc tests, ipc fuzz, kani ntfn + waitqueue
2026-10-26 20:27:38 +01:00
d43421e33e
feat(kernel): object lifetimes via slot generations
...
- SlotGenerations per table, never wraps, retires exhausted slots
- caps to dead/reused process, endpoint, region rejected
- endpoints owned, die w/ owner, waiters woken w/ error
- DropCap syscall + libkos wrapper
- irq table bounds-checked, routes dropped w/ thread
- self-kill terminates caller, exit_thread full cleanup
- sched run_now: switch to named thread, not queue head
- lifetime tests, fuzz kills/recreates, kani gen proofs
2026-10-26 19:54:23 +01:00
8d2c40cb1e
feat(kernel): kernel-owned memory model
...
- FrameTable: per-frame owner/use, replaces PagePool + ProcessMemory
- Quota per process, page tables charged to owner
- Regions for shared mem + device MMIO, torn down w/ owner
- Platform: create/destroy AS, map/unmap, zero_frame; drop translate_va
- RequestPages/ReleasePages/MapPage(move) checked vs kernel records
- user window 0x1000..32MiB, private pts by construction
- boot loader + device setup go through kernel api
- kernel built in place, boot allocator bounded
- isolation tests, randomized invariant test, FrameTable kani proofs
- fixes: inflated release, unmapped release, dst va unchecked,
heap overflow, unbounded pt alloc, unzeroed frames, double free
2026-10-26 19:40:50 +01:00
e931485ca7
fix: sched round-robin, cap slot OOB panic, clippy, make ci
...
- RunList -> intrusive FIFO, no starvation, no silent drop
- cap table sized to CapSlot range (kani caught OOB panic)
- snake food color test uses const
- kani cfg lint, RUST_MIN_STACK for big kernel struct in tests
2026-10-26 19:08:30 +01:00
1c04ddf909
style: cargo fmt kernel crates
2026-10-26 19:07:54 +01:00
818eaecadd
refactor: replace manual debug_putchar with libkos::println!
...
Init, storage server, and hello app now use println!/print! macros
instead of manual byte-by-byte UART output.
2026-10-19 01:00:33 +02:00
052d7c4a64
refactor: remove old userspace and userspace-rt crates
2026-10-19 00:58:14 +02:00
9dd400656c
refactor: migrate all servers, apps, binaries to libkos
2026-10-19 00:57:27 +02:00
a4315d7665
refactor: migrate server-framework to depend on libkos
2026-10-19 00:55:54 +02:00
edc55d6f2e
feat(libkos): add fs, display, input service client modules with newtypes
2026-10-19 00:55:15 +02:00
25a4c9415b
feat(libkos): add print!/println! macros via debug_putchar
2026-10-19 00:53:52 +02:00
dbec3842fa
feat: create libkos crate with core modules (syscall, alloc, rt)
2026-10-19 00:53:31 +02:00
2792387348
fix: VirtIO keyboard input latency and direction accuracy
...
- Driver now drains SYN/repeat/unknown events in a loop instead of
returning None on the first non-key event
- Snake drains all pending input events per tick, applies only the
latest direction change (prevents stale event accumulation)
2026-10-19 00:34:13 +02:00
0a47b5489a
feat: VirtIO keyboard input for QEMU GUI
...
- Add VirtIO input device driver (device ID 18) reading from eventq
- Translate Linux keycodes (WASD, arrows) to InputEvent
- Kernel probes for VirtIO input device, falls back to UART
- Add -device virtio-keyboard-device to run-gui Makefile target
- input-qemu tries VirtIO first, falls back to UART
2026-10-19 00:30:20 +02:00
68849f698b
fix: GPU display at correct resolution, clean up diagnostics
...
- Add GET_DISPLAY_INFO command to query QEMU's preferred resolution
(was hardcoded 640x480, QEMU wanted 1280x800)
- Allocate framebuffer matching actual display dimensions
- Remove debug diagnostic output from display server and snake
2026-10-19 00:24:37 +02:00
f5ac7189a9
fix: VirtIO GPU driver and snake rendering through display server
...
- Fix swapped command IDs: attach_backing was 0x0104 (FLUSH),
resource_flush was 0x0106 (ATTACH_BACKING). Now correct.
- Increase process stack from 4 pages (16 KB) to 16 pages (64 KB)
to prevent stack overflow in display process.
- Restructure display-qemu to avoid MockFramebuffer's 32 KB array
inflating the VirtIO path's stack frame.
- Wire snake app to render through display server via IPC instead
of UART escape codes. Uses existing renderer.rs pixel renderer.
- Init now delegates display endpoint to snake process.
2026-10-19 00:01:29 +02:00
194024bc08
fix(ext4): remove old monolithic ext4.rs superseded by ext4/ directory
2026-10-18 23:30:56 +02:00
9be67e403b
fix(ext4): review fixes — bounds checks, delete safety, blocks_count, multi-extent dirs
2026-10-18 23:28:34 +02:00
73661e925d
test(ext4): comprehensive edge case and integration tests
2026-10-18 23:17:15 +02:00
bc4659eabd
feat(ext4): mkdir and nested directory support
2026-10-18 23:16:35 +02:00
2ede01bdd2
feat(ext4): file create, write, and delete with extent management
2026-10-18 23:15:38 +02:00
bff18df349
feat(ext4): directory entry find/add/remove operations
2026-10-18 23:14:02 +02:00
fe0693b969
feat(ext4): block and inode bitmap allocation/deallocation
2026-10-18 23:12:33 +02:00
381669d9dd
feat(ext4): inode read/write, init_empty_inode, multi-extent read
2026-10-18 23:09:57 +02:00
5d5714914f
refactor(ext4): split into submodules, add io/superblock/bgd with write-back
2026-10-18 23:08:07 +02:00
d85cb5b55d
feat(storage): add mkdir to FileSystem trait with stubs
2026-10-18 23:06:26 +02:00
25824a01fe
fix: remove broken QEMU GPU swap hack, add graceful fallback
...
VirtIO GPU:
- Removed QEMU 11.x command type swap hack (0x0104↔0x0106).
The hack was wrong — swapping types without swapping data formats
meant handlers read wrong struct layouts. Commands returned OK
but did nothing.
- Restored correct VirtIO spec command types.
- Removed broken test pattern (fill + flush that never displayed).
- GPU setup now reports success/failure without crashing.
Display server resilience:
- VirtioFramebuffer::new() → try_new() -> Option<Self>.
Returns None if phys-info page is unmapped or has zero dimensions.
- bins/display-qemu: tries VirtioFramebuffer, falls back to
MockFramebuffer. Display server runs headless instead of crashing.
- No more page faults when GPU is absent.
The GPU driver code is architecturally correct (proper VirtIO
commands, proper data formats) but doesn't produce visible output
on QEMU 11.x due to a suspected QEMU legacy MMIO bug where
transfer_to_host_2d succeeds but doesn't transfer data. The driver
is ready for when this is fixed or a different QEMU version is used.
2026-10-18 22:49:55 +02:00
