An HCD programs a bus-master engine: it needs a descriptor ring that is physically
contiguous, at a physical address it knows, uncacheable, and pinned. mmap gives none
of those. Add dma_alloc(len, flags) -> vaddr (rax), paddr (rdx) and dma_free(vaddr,
len): grant contiguous, zeroed, pinned, strong-uncacheable memory in a per-process DMA
arena (PML4[228]) and hand back both addresses.
Pieces: pmm.allocContiguous(count, max_phys) finds a run of contiguous free frames
below a cap (dma_below_4g for 32-bit engines); mapUserDmaInto maps them uncacheable
(PCD|PWT) but WITHOUT device_grant, so unlike an MMIO grant these frames are real RAM
and freeSubtree returns them on teardown — a driver that dies leaks nothing. dma_free
is bounded to the DMA arena so it can never unmap the caller's stack/heap/MMIO.
dma_write_combining is accepted but falls back to coherent (WC needs PAT programming).
Runtime: runtime.dma.alloc/free (a two-return-value stub, like replyWait). New `dma`
kernel test drives the mechanism directly — contiguity, the below-4G cap, coherent
mapping, and reclaim-on-teardown (no leak). The thin syscall wrappers follow the tested
mmap/mmio_map shape and land their first real use with the first DMA driver. Suite
38/38 plus host tests.
The tree had zero memory barriers — correct-by-accident on x86 (TSO + strong-
uncacheable MMIO), but a landmine for the first DMA driver and for ARM, which is the
win condition. Add /lib/mmio: typed volatile register access (read/write) plus mb /
rmb / wmb, lowered per-architecture (mfence/lfence/sfence on x86_64, dsb sy/ld/st on
aarch64) so the ordering rules are a named primitive, not scattered `asm volatile`.
`volatile` is not a barrier — it says nothing about ordinary stores (a DMA descriptor
in WB RAM) relative to a volatile doorbell write; wmb() between them is the fix.
Prove it on the one existing caller: hpet now does its register access through
mmio.read/write. It needs no barriers itself (pure MMIO, no DMA, UC grant on x86) —
the point is the typed, arch-portable access every driver should use; the barriers are
there for the DMA drivers to come.
New `mmio` module injected into addUserBinary; host test asserts the barriers assemble
and a register round-trips. Suite 37/37 plus host tests.
ipc_call and ipc_reply_wait grow a `send_cap` argument (r9) and a `received_cap`
return (r8): an endpoint travels alongside a message, installed into the receiver's
handle table. The transfer is a share, not a move — the endpoint's refcount is bumped
and the sender keeps its handle. If the receiver's table is full the call fails
-ENOSPC and the message is NOT delivered (a half-delivered capability is worse than a
failed send); a bad handle fails -EBADF. Both directions carry a cap: a client's call
hands one to the server (seen in the server's replyWait), and the server's reply hands
one back (seen in the client's call return).
This is the "open" primitive the driver model was blocked on: a bus driver mints a
per-device endpoint and hands it to a class driver, giving it a private channel to one
device without the 8-slot global name registry.
Kernel: shareCapability in ipc-synchronous.zig at both copy points; new
setSystemCallResult3 (r8, saved/restored by the syscall stub); Task gains
ipc_send_cap / ipc_received_cap. Runtime: callCap + Reply, replyWait gains send_cap
and Received.cap; plain call/replyWait delegate with no_cap. New abi.no_cap.
New ipc-cap test (two kernel tasks exercise both directions, each verifying the
endpoint it received is the same object shared, refcount bumped to 2). No class driver
consumes callCap yet — it lands with the first one. Suite 37/37 plus host tests.
drivers.md gains a "How a driver gets started" section — the doc explained what a
running driver does but never who starts it. It lays out the three-level supervision
hierarchy (kernel spawns init; init spawns the services; the device-manager discovers,
matches, and spawns the drivers) and names the two user-space policies that "configure"
drivers today: init's service list and the device-manager's match table.
driver-model.md's "what exists today" adds system_spawn and the supervision model, and
the drivers.md restart bullet is refreshed: a spawning supervisor now exists, a
restarting one still doesn't.
Retire the kernel's spawn-every-initial-ramdisk-binary loop, resolving the
service/driver split into a real three-level supervision hierarchy:
kernel -> spawns init (PID 1) only, and publishes the initial-ramdisk
init -> the service supervisor: spawns the system services (vfs, device-manager)
device-manager -> spawns the drivers it matches (hpet)
The kernel now only hands the initial-ramdisk image to the process layer
(publishInitialRamdisk) so user space can system_spawn from it; it launches nothing
bundled itself. init gains a boot-services list ("vfs", "device-manager") and spawns
them best-effort before settling into its heartbeat — policy lives in user space,
where a microkernel keeps it. Drivers are absent from that list on purpose: the
device manager owns them. Test-only binaries (vfs-test, bus) no longer run at boot;
their kernel self-tests still spawn them directly.
This removes increment 2's transitional double-spawn: a real boot now brings hpet up
exactly once (verified — kernel -> init -> vfs/device-manager -> hpet, zero "claim
failed"). The automated suite is unaffected: test builds run their case and halt
before the normal boot path, so each already spawns its own binaries. Suite 36/36
plus host tests; normal boot verified by hand under QEMU.
Add a `system_spawn(name)` system call: the kernel loads a binary bundled in the
initial-ramdisk, by name, as a fresh ring-3 process. It's the mechanism a user-space
supervisor needs — discovery and policy stay in user space, the kernel only spawns.
The kernel already holds the initial-ramdisk image from the boot handoff; it now
stashes it (process.setInitialRamdisk) so the handler can resolve names, bounds-checks
the name into the user half like debug_write, and returns -1 for an unknown name or a
load failure. Ungated for now (any process may spawn any bundled binary); a spawn
capability belongs here once the model grows one.
The device manager stops logging "would spawn it" and calls runtime.system.spawn on
its matched driver. On QEMU it discovers the HPET, matches `hpet`, and spawns it — and
the driver comes all the way up (claims the timer, maps its MMIO, binds and services
its IRQ, prints "hpet: ok"). The device-manager test now keys on that final marker:
since only the manager is spawned, `hpet: ok` appearing proves the whole
discover -> match -> system_spawn -> driver-up chain end to end.
Transitional: the kernel still auto-spawns the whole initial-ramdisk at boot, so a
real boot briefly double-spawns hpet (the second claim fails harmlessly). Increment 3
removes that redundancy so the manager is the sole owner of driver spawning. Suite
36/36 plus host tests.
The reorg moved user binaries under /system (init -> /system/services/init,
drivers -> /system/drivers/<name>, vfs-test -> /system/services/vfs/vfs-test), but
many comments and log strings still named the old /sbin/ home. Retarget them all:
kernel/loader/test comments and the two boot log lines, plus vision.md and the
driver-model.md proposed tree (also dropped the stale `d` suffixes and rt->runtime
there). The initial-ramdisk spawn log no longer fakes a /sbin/ prefix, since those
binaries live in different homes (services vs drivers).
Left the FSH design doc's /sbin and /lib rows alone — whether /sbin stays a
directory at all is a design call for its owner, not a stale-comment fix.
The `system` module (formerly `danos`) had become a grab-bag: it held the
loader<->kernel handoff *and* the kernel<->user ABI *and* the device wire types, in
one module three different audiences imported. Usage proved the seam — the
bootloader never touched the syscall/device ABI, and user space never touched the
boot handoff — so split it by audience, one module per contract:
system/boot-handoff.zig loader <-> kernel: BootInformation, Framebuffer,
MemoryMap, the VM layout + physicalToVirtual, kernel_abi
system/abi.zig kernel <-> user, core: SystemCall, mmap prot flags,
page_size, notify_badge_bit, ServiceId
system/devices/device-abi.zig kernel <-> user, devices: DeviceDescriptor,
DeviceClass, ResourceDescriptor, ResourceKind, ...
device-abi is the devices sub-project's public interface, exposed as its own module
the way vfs exposes vfs-protocol — importable by user space, unlike the
kernel-internal device model it also feeds. That collapses a real duplication:
DeviceClass and ResourceKind were defined twice (device-model.zig and the contract,
kept "in sync by hand"); device-model now re-exports them from device-abi, so the
enum a driver matches on and the one the kernel classifies with are one type.
Each import now declares which contract it speaks: the bootloader imports only
boot-handoff; a driver only abi + device-abi (via the runtime); the kernel all
three. This also retires the `system` / `runtime.system` name overlap. page_size
lands in abi (it's part of the mmap contract user space aligns to); the bootloader
keeps its own local 4 KiB constant so it depends on nothing but the handoff.
All 21 importers rewired, docs updated to keep /system mapping to source. Build,
host tests, and the QEMU suite (36/36) all green.
The device manager is the ring-3 process that turns the device tree into a running
system — the udev-analog. It is mechanism-vs-policy done right: the kernel
enumerates the hardware and enforces the claim capability; this decides which
driver serves which device, using no special privilege (the same device_enumerate
any process could call).
This first increment does the discovery + matching half: system/services/
device-manager enumerates /system/devices, matches each device to a driver by
class (a small static policy table), and logs the decision — finding the HPET
(a timer) and deciding `hpet` serves it. It does not spawn yet: spawning needs a
`system_spawn` system call (the kernel spawns every initial-ramdisk binary in a
loop today), which is the next increment. New `device-manager` test; suite 36/36
plus host tests.
The serial capture is a dev/host artifact, so it lands in the qemu-test scratch
area rather than inside zig-out (which we mount as the boot volume). /var/log/system
stays reserved for the kernel's own logging system later.
The shared kernel<->user ABI contract (BootInformation, the SystemCall numbers,
DeviceDescriptor, page_size, ...) is now the `system` module at
system/system.zig, following the convention that a directory's root file takes
the directory's name.
One overlap to note: the runtime's syscall wrappers are already `runtime.system`,
so the single file that uses both the contract and those wrappers
(library/runtime/heap.zig) aliases the wrappers locally as `system_calls`. The
two are distinct (top-level `system` vs `runtime.system`); everywhere else the
contract is just `system`.
Also: the QEMU run's serial capture now lands in the FHS log location,
zig-out/var/log/system/serial0-<timestamp>.log — a stand-in for the kernel's own
logging system, which will eventually write there itself.
Suite 35/35 plus host tests green.
`zig build` now installs into a FHS-shaped zig-out that *is* the danos filesystem
and the boot volume — no more zig-out/bin or a separate esp/:
zig-out/EFI/BOOT/BOOTX64.efi (firmware entry; UEFI fixes this path)
zig-out/boot/initial-ramdisk.img
zig-out/system/kernel (the kernel binary)
zig-out/system/services/init vfs
zig-out/system/drivers/hpet bus
Binaries land at their addressed, leaf-collapsed paths per the sub-project
resolution rule (system/services/init/init.zig -> system/services/init); vfs, hpet,
and bus are installed to their FHS homes too, so the image is complete even though
at boot they arrive inside the initial-ramdisk.
The bootloader (boot/efi.zig) now loads each artifact from its FHS path
(system\kernel, system\services\init, boot\initial-ramdisk.img); run-x86-64 mounts
zig-out directly; the QEMU test harness assembles its ESP from the FHS zig-out.
Also renames system/kernel/main.zig -> kernel.zig so the kernel follows the
name/name.zig convention (kernel/ = ring-0 code, services/ = ring-3 OS services).
Documents the resolution rule in the repository-layout section (README + coding
standard). Suite 35/35 plus host tests green.
Follow-up to the monorepo re-org. Suite 35/35 plus host tests green.
POSIX compatibility is now its own library, library/posix/ (unistd, stdio),
layered strictly over the runtime — it calls the runtime's IPC/heap, never
system calls directly. The runtime is now POSIX-free (the danos-native
application ABI). The VFS wire protocol is danos-native throughout
(Stat -> FileStatus, .stat -> .status, O_CREAT -> create); the POSIX layer
maps the POSIX spellings at the boundary. The coding standard's ABI-name
exception is scoped to one place: a file is allowed POSIX spellings only if it
lives under library/posix/ — everywhere else, danos naming with no exception.
Naming fixes, all mechanical:
- initrd -> initial-ramdisk: the source file, the module, the tool
(make-initial-ramdisk.py), the artifact (initial-ramdisk.img, including the
bootloader's load path), and the identifiers.
- system/kernel/device-service.zig -> devices-broker.zig: it is ring-0 kernel
code (the trusted device table + claim capability), not a ring-3 service. The
future user-space device *manager* (policy) will live in system/services/.
- Dropped the daemon `d` suffix: hpetd -> hpet, busd -> bus. A driver lives in
system/drivers/, so the folder already says what it is; encoding the role in
the name too is redundant. The coding standard drops that exception.
- system/devices/aml/interp.zig -> interpreter.zig (the type was already
Interpreter).
The source layout now mirrors the runtime filesystem hierarchy
(docs/danos-file-system-hierarchy-FSH.md): what lives under system/ in the
source is what a running danos represents under /system. Each service and
driver is a sub-project directory that is its own Zig module — cross-project
references go by module name, never by a path into another project's files.
Moves (all git mv, history preserved):
- src/ -> system/ (danos internals; the self-representation)
root.zig -> danos.zig (the kernel<->user contract module)
kernel/arch/ -> kernel/architecture/ (arch -> architecture)
device/ -> devices/ (what /system/devices reflects)
boot/ -> /boot (the loaders, top level)
- sbin/ -> split by role:
init, vfs -> system/services/<name>/<name>.zig
hpetd, busd -> system/drivers/<name>/<name>.zig
vfs-test -> system/services/vfs/vfs-test.zig (inside the vfs project)
- lib/ -> library/runtime/ (room for other libraries beside runtime)
The VFS wire protocol becomes its own module, system/services/vfs/protocol.zig
("vfs-protocol"): the vfs sub-project exposes its interface, and the runtime's
file layer imports it by name. First instance of the "protocol module" pattern
(docs/driver-model.md); usb/block will expose theirs the same way.
Also: fix a naming-standard violation in the protocol — Op -> Operation (and
req -> request, _pad -> _padding). Docs updated: /system/services added to the
FHS doc, a repository-layout section added to the docs index, and stale source
paths swept across comments and docs.
Runtime boot paths are unchanged (the bootloader still loads /sbin/init);
aligning the runtime filesystem to the FHS is a separate follow-up. Suite 35/35
plus host tests green.
Two driver-model milestones plus a tree-wide naming pass. Suite 35/35
(QEMU) + host tests green.
M11 — IRQ-as-IPC. A ring-3 driver now sleeps until its device interrupts
it. New src/kernel/irq.zig: per-GSI endpoint bindings, comptime per-vector
trampolines, dispatch = mask GSI -> LAPIC EOI -> notifyLocked, all under one
lock region. irq_bind/irq_ack syscalls, gated by the device claim like
mmio_map. interruptDispatch no longer EOIs — each handler owns its EOI,
because a level line must be masked before it is acknowledged (irq_ack is
the unmask). Bindings are keyed on the owning task and released on exit
(a shared endpoint's siblings survive). hpetd rewritten interrupt-driven.
Tests: hpet (rewritten, reads back the I/O APIC routing) and irqfree.
M12 — bus drivers. DeviceDesc gains a parent, making the device table a
tree. dev_register (device_register) lets a process publish children below
a device it claimed; the kernel enforces resource containment (a child's
resources must nest in its parent's), so a descriptor can't fabricate a
window over kernel RAM. Descriptor copied in via copyFromUser (physmap
walk — an unmapped user pointer fails the call instead of faulting the
kernel). Per-parent child cap bounds table exhaustion. sbin/busd.zig is a
worked bus driver. Test: bus.
Naming — per docs/coding-standards.md: non-acronym abbreviations spelled
out (message, descriptor, device_service, scheduler, runtime, physical,
interpreter, ...); acronyms kept (IPC, MMIO, DMA, HCD, ...); files are
kebab-case (ipc-synchronous.zig, device-service.zig, vfs-protocol.zig, ...).
Exceptions: POSIX/C ABI names and Zig idioms (init/len/ptr) kept. Module
collisions resolved by specific naming (config -> parameters, device.zig
alias -> device_model). AML op/Op disambiguated: op = opcode, Op =
operation; per-opcode parse handlers renamed opX -> parseX.
New driver docs: drivers.md, driver-model.md (bus/class/HCD shapes + the
proposed M13–M16 ABI), coding-standards.md.
A user-space process can now touch real hardware directly, capability-gated by
the device tree — the microkernel driver model.
- src/kernel/devsvc.zig: flattens the discovered device tree into an
id-indexed snapshot + a claim table at boot (devsvc.init from main.zig).
- Syscalls 11-13: dev_enumerate (snapshot the table), dev_claim (take
exclusive ownership), mmio_map (map a claimed device's MMIO window into the
caller's AS and return the register base). The claim is the capability:
mmio_map refuses any device the caller doesn't own.
- paging.mapUserDeviceInto: maps device MMIO strong-uncacheable (PCD|PWT) and
marks each leaf with a device_grant PTE bit; freeSubtree skips pmm.free on
those leaves, so tearing down a driver never returns MMIO frames to the RAM
pool (the teardown hazard). MMIO grants live in a distinct arena, PML4[226]
(Task.dev_map_next), so device pages widen no kernel mapping.
- lib/dev.zig: user enumerate/claim/mmioMap wrappers; shared DeviceDesc/ResDesc
in danos (root.zig). sbin/hpetd.zig: finds the HPET, claims it, maps its
registers, enables the counter (an MMIO write) and reads it (0xF0) — proving
read+write passthrough to real hardware.
- Tests: `hpet` (driver reads the counter advancing from ring 3) and `iopass`
(device-granted frame survives address-space teardown). Suite 33/33.
irq_bind/irq_ack (IRQ-as-message) are stubbed (-1) pending; notifyFromIsr (M7)
is the hook they'll use.
The payoff milestone: files are served by a user-space process, reached over
IPC — the kernel never sees a path or an fd.
- lib/vfs_proto.zig: the VFS wire protocol (Op, Request/Reply fixed header +
inline payload, Stat), shared by client and server; one message <= MSG_MAX.
- lib/ipc.zig: replyWait() — the server-side dual-return stub (length in rax,
badge in rdx via a "+{rdx}" read-write operand), deferred from M7.
- sbin/vfs.zig: the real VFS server — an in-heap ramfs (open creates a node)
with an IPC_ReplyWait dispatch loop serving open/read/write/stat/close.
Registers its endpoint under the well-known vfs id at startup.
- lib/unistd.zig: POSIX-style client API in rt — a per-process fd table +
open/close/read/write/lseek/stat, each an IPC_Call to the VFS. The kernel
knows nothing of fds; the table lives here.
- lib/stdio.zig: C stdio over unistd — FILE + fopen/fclose/fread/fwrite/
fseek/ftell/rewind/feof/ferror/fputs/fputc/fgetc (unbuffered for now).
- sbin/vfstest.zig: a client that opens/writes/seeks/reads a file and only
heartbeats "vfstest: ok" if the round trip matched. Packed in the initrd.
- New `vfs` test drives it end to end. initrd test relaxed to generic
liveness. Suite 31/31.
Ship more than one user binary: the bootloader now ferries an initrd bundle
(the VFS server + future drivers) alongside sbin/init, and the kernel unpacks
and spawns each program.
- src/user/proto/initrd.zig: the container format (Header{magic,count} +
Entry{name[32],offset,len} + blobs) with a validating Reader, shared by the
kernel and the packer.
- tools/mkinitrd.py: host-side packer (Python — trivial format, and sidesteps
the reworked Zig 0.16 std fs/args API). build.zig runs it on the built user
binaries via addSystemCommand and installs initrd.img to zig-out/bin + the
ESP root.
- BootInfo gains initrd_base/initrd_len; efi.zig loadInit refactored into a
shared loadFile, and loadInitrd ferries \initrd.img into surviving
LoaderData like init.
- main.zig startInitrdBinaries(): parse the image, spawn every entry (kernel-
spawns-all for now; init takes over via sys_spawn later).
- sbin/vfs.zig: a heartbeat stub (the real VFS server is M9), packed into the
initrd to prove the pipeline.
- New `initrd` test: parse + spawn + confirm the vfs stub reaches ring 3 and
heartbeats. Harness ships initrd.img on the ESP. Suite 30/30.
The microkernel message backbone the VFS server and drivers will ride on.
- src/kernel/ipc_sync.zig: Endpoint (sender FIFO threaded via Task.next +
a recv WaitQueue for servers + a small notification ring). call() (client
sends, wakes a server, blocks) and replyWait() (server replies to the held
caller, then receives the next). Reply routing keys on Task.ipc_client —
synchronous IPC owes one reply at a time. Payloads copy frame-to-frame
through the physmap (copyAcross on arch.translate, added in M5); an unmapped
page fails the copy instead of #PF-ing. MSG_MAX 256.
- Bootstrap naming: an integer name registry (danos.ServiceId, vfs=1) with
create_endpoint / ipc_register / ipc_lookup — any process finds a server
without threading a handle through spawn.
- notifyFromIsr(): ISR-safe async wake (badge with the high bit set), the
hook M10's IRQ-as-message needs. Unused/untested until then.
- Task gains handles[16] (opaque *Endpoint, to avoid a sched<->ipc import
cycle) + ipc_client/send/reply/status fields; sched gains
blockCurrentLocked/readyLocked; arch gains setSyscallResult2 (rdx badge).
- Syscalls 6..10 wired in process.zig; exit() now drops the caller's endpoint
refs. lib/ipc.zig: user-side createEndpoint/register/lookup/call
(server-side replyWait lands with the first server in M9).
- New `ipc-call` test: two kernel tasks ping-pong 100 calls, every reply
request+1. Suite 29/29.
Add lib/ — the shared user-space runtime every user binary links against
(init now, servers/drivers later): syscall wrappers, the heap, IPC stub,
and the process start shim.
- lib/heap.zig: the kernel first-fit free-list ported to user space, grown
via the mmap syscall instead of pmm+mapPage. Dual API over one global free
list: extern "C" malloc/free/calloc/realloc (C ABI for future C code) and a
std.mem.Allocator adapter (with in-place resize) for Zig std containers.
- lib/syscall.zig + sys.zig: raw syscall0..5 (arg3 in r10) and typed
yield/write/sleep/exit/mmap/munmap over danos.Syscall.
- lib/start.zig: naked _start -> rt_start -> root.main() (SysV realign via
call), panic -> exit(127).
- lib/user.ld: the user link script, moved from sbin/linker.ld (shared by all
user binaries).
- build.zig: register the `rt` module; add an addUserBinary() helper that is
the one recipe for every user binary (freestanding, .large, use_lld,
user.ld, image_base), replacing the bespoke init block.
- sbin/init.zig: migrated onto rt; drops its hand-rolled syscall2/shims. Now
proves the heap (alloc -> write from a heap pointer -> free) before the
heartbeat loop. Serial shows "init: heap ok". Suite 28/28.
Start the user-space driver track (VFS + IPC + heap). This lays the
process/syscall foundation the runtime heap will grow on.
- Rename usermode.zig -> process.zig; drop the retired hello/ping blob and
its `user` test (subsumed by the real /sbin/init exerciser). Keep the
isolation-proof pf blob and the user-pf test.
- Add danos.Syscall as the single source of truth for syscall numbers, shared
by the kernel dispatcher and (later) the user runtime lib. Dispatch on the
enum. New calls: 1=yield, 4=mmap, 5=munmap. Widen debug_write's bounds
check to the whole user low half so heap buffers are writable.
- mmap grants zeroed RW+NX pages from a per-process bump arena
(Task.heap_next, PML4[224] above image+stack); munmap frees the frames.
Add paging.translateIn / unmapInto (+ arch.translate / unmapUserPageInto)
as the primitives munmap and future cross-AS copies need.
- New `usermem` test: grant three pages into a fresh AS, translate them,
release via the munmap path, tear down, and assert no frames leak.
Suite 28/28 (user -> usermem).
The generic kernel imports cpu.zig as @import("arch") but still spoke
x86: readCr3/readCr2, pml4 and rip/rsp parameters, lapicHz/tscHz,
ioapicEntry*, IST stacks, APIC ids. Rename the public interface so the
same names work for x86_64, aarch64, and riscv64:
- readCr3 -> activePageTable; pml4 params -> root; vectorName ->
exceptionName; postCode -> checkpoint; lapicHz/tscHz ->
timerClockHz/clockHz; ioapicEntry* -> irqRoute*; ist_stack_size/
setApIstStack -> fault_stack_size/setFaultStack; startSecondary
takes a hw_id (APIC id here; MPIDR/hart id elsewhere)
- New trap-frame accessors (instructionPointer, stackPointer,
fromUser, faultAddress, syscallNumber/syscallArg/setSyscallResult)
so the generic syscall dispatcher and fault printer never name an
x86 register; readCr2 folds into faultAddress (null unless #PF)
- Generic-kernel identifiers follow: Task.pml4 -> aspace, rsp -> sp,
user_rip/user_rsp -> user_ip/user_sp, PerCpu.apic_id -> hw_id
PlatformConfig fields and the ACPI apic_id stay as-is: they describe
hardware actually discovered on this platform, and another arch would
define its own. The user-mode tests now assert fromUser instead of
the exact CS selector; the user-pf expectation follows the fault
printer's RIP -> IP label. All 28 QEMU tests pass.
init is now a real scheduled ring-3 process: it prints a heartbeat and
sleeps, forever. The kernel spawns it at boot via spawnProcess (its own
address space, preemption on) and the boot context drops to idle — the
system's steady state is "kernel idle, init alive in ring 3", beating
~1 Hz. Adds the sleep(ms) syscall. The init/process tests are reshaped
around the heartbeat (repeated syscalls, still-alive, two concurrent
processes on distinct address spaces). Pins init to LLD and folds the
.large code model's .ltext/.lrodata/.ldata into the linker script so its
code segment is R+X. Removes the now-dead borrowed-thread ELF loader
(runInitElf); spawnProcess is the one path. Docs updated. Suite 28/28.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Per-process address spaces (AddressSpace = a PML4 with an empty user
half and the shared kernel half copied in; create/destroy in paging.zig)
with CR3 switched on context switch and TSS.rsp0/kernel_rsp published per
switch. The GS base now points at an arch per-CPU block and every ring
transition observes the swapgs discipline, so a ring-3 `mov %ax,%gs` can
no longer poison per-CPU access. syscall/sysret is the primary user entry
(int 0x80 kept as a test path); one handler, installed once at boot,
serves both and dispatches on whether the caller is a scheduled process
or a borrowed test thread. spawnProcess loads an ELF into a fresh address
space and schedules it; exit frees the address space after switching to
the kernel tables. New `process` test: init runs twice as a real process
(create/exit/recreate) on its own page tables, coexisting with a kernel
task under preemption. Suite 28/28.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Per-core MSR setup (EFER.SCE, STAR, LSTAR, SFMASK) enables syscall; the
GDT layout was already chosen so sysret lands on CS 0x23 / SS 0x1B. A new
syscall_entry stub swaps in the kernel GS, switches to the task's kernel
stack via the per-CPU block, builds a CpuState frame identical to the
interrupt path's, and reuses interruptDispatch (vector 128) — then
sysretq back. enter_user now also publishes kernel_rsp so the borrowed
path's syscalls land on a good stack. /sbin/init uses the `syscall`
instruction. int 0x80 stays for the test blobs. Suite 27/27.
(Noted in the stub: sysretq #GPs in ring 0 on a non-canonical return
RIP — a hardening item once untrusted user code exists.)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The GS base now points at an arch-owned ArchPerCpu (percpu.zig) holding
the kernel RSP and a scratch slot (for the coming syscall stub, at fixed
%gs offsets) plus the scheduler pointer. isr_common conditionally
swapgs's on entry/exit when the interrupted frame was ring 3, and
enter_user swapgs's before dropping to ring 3 — so kernel code always
sees the kernel GS base and a ring-3 `mov %ax,%gs` can no longer poison
cpuLocal(). No swapgs on ring-0 interrupts (the common case). Suite
27/27, including int 0x80 from ring 3 and faults.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Task gains kstack_top and pml4 (0 = kernel task); PerCpu tracks the
loaded CR3. A shared switchTo() publishes the incoming task's kernel
stack (TSS.rsp0) and address space (CR3, only when it changes — every
write is a full TLB flush), used by both schedule() and exit(). APs
adopt the kernel page tables explicitly rather than the caller's live
CR3. All tasks are kernel tasks today, so this is a no-op beyond the
register switch. Suite 27/27.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
paging.init now maps only the physmap, the framebuffer/LAPIC windows,
and the kernel's own segments; the entire low canonical half is left to
user space. Every higher-half PML4 entry is pre-created so a per-process
address space can share the kernel half by copying PML4[256..512), with
an assert against late top-half entries and a 4 GiB guard on
pre-switch table frames. The AP trampoline's low identity page is now
created transiently by arm() and unmapped by disarm(); startAp asserts
the page-table root is 32-bit addressable. Docs (paging.md) updated.
Suite 27/27; 4-core normal boot reaches /sbin/init.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The kernel now links at 0xFFFF_FFFF_8000_0000 (code_model .kernel) and
loads low via the linker script's AT() clauses (.text at 1 MiB). A new
asm _start installs a 64 KiB kernel-owned .bss stack — the loader stack
is a low address that goes away with the identity map — and calls the
Zig entry, which reaches boot_info through the physmap. The user-pf test
blob reads the LAPIC through the physmap window (still present|user, ec
0x5). The low identity map still coexists in the kernel's tables as the
safety net. Suite 27/27.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
paging.init now builds the physmap (physToVirt(phys)) alongside the low
identity map, so both addressing modes resolve during the transition.
tableAt (the page-table walk hinge), the pmm bitmap, the framebuffer,
LAPIC/IOAPIC/HPET/PM-timer/SPCR MMIO, the ACPI table walk, AML
OperationRegions, the ACPI power registers, the user-ELF frame fills,
and the AP trampoline arm/disarm all reach physical memory through the
physmap. Hal.mapMmio now maps into the physmap and returns the virtual
address, so the device layer never learns the layout. boot_info and its
pointees are converted at kmain entry. The kernel still links and runs
low; identity is the safety net until it's removed. Suite 27/27.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Add the higher-half layout constants (physToVirt/virtToPhys, physmap
and kernel bases) to the danos handoff module, and KernelSegment.phys
so the loader records where it actually placed each segment.
The loader now builds its own page tables before ExitBootServices —
identity + a physmap of low RAM (2 MiB pages) plus 4 KiB mappings for
any higher-half kernel segment — and switches CR3, then calls the entry
with boot_info in RDI, interrupts off. The kernel still links and runs
low (identity), oblivious; this proves the CR3 dance in isolation.
Suite 27/27.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The self-hosted linker ignores parts of linker.ld (PHDRS, /DISCARD/,
section order). The higher-half move needs the script authoritative.
Suite 27/27 on the LLD build.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ring 3 works: user GDT descriptors (sysret-ready layout), TSS.rsp0,
U/S-bit user mappings (W^X preserved), an int 0x80 syscall gate with a
mutable trap frame, and a setjmp-style enter/exit path. /sbin/init is a
real freestanding Zig binary built from sbin/, shipped on the ESP,
loaded by the bootloader (BootInfo.init_base/len), validated and mapped
by an in-kernel user-ELF loader, and run at CPL 3 — syscalls: exit,
ping, write. Tests: user, user-pf (U/S isolation proof, error code
0x5), init. Suite 27/27.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
max_cpus, max_tasks, kernel/IST stack sizes, and timer_hz move from scattered constants into one config module. root.zig goes back to being just the boot contract. Values unchanged; any can become a -D build option later.
Hoist max_cpus into danos (root.zig), raise it to 128, and stop reserving a static [max_cpus][16 KiB] IST array. Only the BSP's IST stack is static (needed before the allocator); each AP's is heap-allocated when it comes online. Shrinks the kernel from ~1.1 MB to ~139 KiB in ReleaseSmall.
fault-ap-df pins a #DF to an AP (its own IST must catch it); affinity checks a pinned task never migrates. onException now names the core, so an AP fault is attributed and shown contained. Both teeth-checked.
spawnOn(entry, priority, cpu) routes to a per-core pinned queue, merged with the global queue at O(1) selection. Falls back to unpinned for an offline/invalid core.
Add a discovery case (asserts stable ACPI/MADT/FADT/AML facts) and a wx case (audits R+X code vs NX data/rodata/heap/stack via arch.pageExecutable). Wire the existing poweroff/reboot cases into the harness. Suite: 18 -> 22.
Adds paging.isExecutable and arch.trampolinePage; the smp case now asserts the frame is zeroed and non-executable after bring-up. Teeth-checked against a no-op disarm.
Arm the low frame (copy blob, make executable) only while a core climbs, then zero it and restore RW+NX; the frame stays reserved so cores can be re-woken. startSecondary is one re-runnable attempt; boot retries a non-responding core 3x. Validated the retry path by forcing a first-attempt failure.
Daniel Samson