d945483915 added a per-thread IOPL guard using std::sync::Once in a
#[cfg(target_os = "redox")] thread_local! (pci.rs:20) but did not import Once,
so redox-driver-sys fails to compile for x86_64-unknown-redox with E0425
"cannot find type Once" — breaking redox-drm (the /scheme/drm/card0 provider)
and driver-manager in the redbear-full build. Add the cfg-gated import.
Found while driving the redbear-full desktop build.
Operator confirmed "SDDM is the only greeter". Disable the two competing
display entry points that would contend with SDDM for the single-owner
GPU/card0: 20_greeter (redbear-greeterd, an alternative greeter) and 20_display
(redbear-session-launch, a boot-time auto-KDE-session that bypasses the
greeter). Gated reversibly via condition_path_exists on a sentinel
(/etc/redbear/enable-legacy-greeter) the image does not install — definitions
preserved, re-enable with a single touch. 21_sddm remains ungated. Post-login
session launch is SDDMs job via SessionDir=/usr/share/wayland-sessions.
redbear-info's integration check list still labeled the PCI spawner as
'pcid-spawner' and pointed at /usr/bin/pcid-spawner (deleted). Updated
to 'driver-manager' with /usr/bin/driver-manager and a note about
/scheme/driver-manager/bound. Plan v3.2 marks P2-1 done (pcid AER +
pciehp producers validated in QEMU).
SDDM greeter renders on redbear-compositor (per /etc/sddm.conf CompositorCommand);
KWin is the post-login session compositor. Runtime chain (driver-manager →
redox-drm → card0 → sddm → redbear-compositor → greeter) is coherent, BUT three
display entry points (20_display redbear-session-launch, 20_greeter
redbear-greeterd, 21_sddm) all fire oneshot_async and would contend for the
single-owner GPU/card0. For a clean SDDM path the first two must be disabled
(operator sign-off required).
Phase status for making redbear-full compile + run to the SDDM Wayland greeter:
build-system llvm fix, Mesa EGL gate persisted, VirGL priority (all done);
Rust components (redox-drm, redbear-compositor) compile-verified; remaining =
full C/C++ cook campaign (mesa/qt6/kf6/kwin/sddm) + Qt6 null+8 runtime validation
+ boot. Records upstream cherry-pick candidates and component homes.
driver-manager had a priority tie for QEMU virtio-gpu (vendor 0x1AF4, class
0x03): redox-drm (generic class @60) vs virtio-gpud (@60). If virtio-gpud won,
it registered display.virtio-gpu (VT scanout only) and never provided
/scheme/drm/card0, starving KWin/the compositor. redox-drm had priority-61
entries for Intel and AMD but none for VirtIO. Add a priority-61 redox-drm entry
for vendor 0x1AF4 class 0x03 (subclass omitted — QEMU virtio-gpu is subclass
0x80) so the VirGL path deterministically binds redox-drm and gets a real DRM
card0. virtio-gpud remains as fallback.
The meson gate that enables EGL on Redox — adding "redox" to
system_has_kms_drm (meson.build:159, so with_dri→with_egl/gbm) and to the
_GNU_SOURCE platform list (:1208) — existed only as an in-place edit in the
untracked extracted source/ tree. Any clean re-extract lost it and reverted to
"Feature egl cannot be enabled". Capture it as local/patches/mesa/08-*.patch and
wire it into recipe.toml so EGL/GBM/llvmpipe build reproducibly. Source tree
reverted to pristine so the patch is the sole, tracked change.
The relibc-consumer invalidation (correctly) force-deletes every pkgar+target
when relibc rebuilds or its pkgar is missing, to prevent stale-ABI skew. But it
swept up heavy, ABI-inert toolchain packages (llvm21, clang21, lld21, rust)
identically to a 200KB daemon — forcing a 1-2h LLVM re-cook on EVERY base/relibc
bump, which cripples Mesa/KDE iteration.
These are build-time-only STATIC archives (BUILD_SHARED_LIBS=off; no runtime .so
of their own). Their libc/relibc symbol references are UNRESOLVED in the .a and
resolve at the CONSUMER final link — mesa links libLLVM*.a and resolves against
the freshly-built relibc there. So a relibc protocol change is absorbed by
relinking mesa (NOT on the preserve-list; stays invalidated), and recompiling
all of LLVM from scratch is unnecessary.
Add TOOLCHAIN_PRESERVE={llvm21,clang21,lld21,rust} skipped in both invalidation
branches, with escape hatch REDBEAR_FORCE_TOOLCHAIN_RECOOK=1 for a genuine
libc-ABI change. Trades a rare theoretical risk for a large repeated speedup;
the churn during driver/desktop work is Redox scheme/syscall shims the compilers
never call, so the default is safe.
Record the two concrete prerequisites for compositor input wiring, found by
reading the compositor source: (1) an event-loop restructure — it uses a
blocking accept loop, drains keys only reactively, and ClientState stores no
UnixStream, so no input source can push wl_keyboard to a client; (2) the
desktop cannot run without DRM/Mesa (no /scheme/drm/card0 on mini), so the
wiring is unvalidatable until the Mesa EGL/DRI port lands. Defines the exact
wiring for when those clear (read /scheme/evdev, set_vt_mode, real XKB keymap
fd, keycode = evdev - 8).
P0 green (acpid compiles). Fresh redbear-mini boot confirms: no console
regression (login + shell command execution intact) and the raw tap is live
(evdevd reads consumer_raw, evdev scheme functional). Boot-log errors observed
are pre-existing and unrelated to the input stack.
- AER listener path moves from /scheme/acpi/aer (no producer) to
/scheme/pci/aer (pcid's new producer, submodule bump 5a43628d).
- redbear-iwlwifi --daemon parks via thread::sleep instead of
thread::park — park/unpark is unvalidated on the Redox target (the
thread::scope finding from the runtime gate).
Record what landed (S1 raw evdev tap, S4-core VT graphics-mode op, S5 seatd
local package; all compile-checked, boot-validation P0-gated) and correct the
Defect-3 framing: fbcond is a byte-queue + CR/LF + special-key escapes, not a
second canonical line discipline (ptyd is the only one), and the console login
works — so the old "collapse two disciplines" console surgery is deferred as
unnecessary and risky. Grab arbitration + compositor wiring remain gated on the
Mesa desktop bring-up.
Flip [source] from git=jackpot51/seatd to path="source" so the already-tracked
in-tree source/ (52 files, incl. libseat/seatd-launch) is the authoritative
implementation — a first-class local Red Bear package like evdevd and
redbear-compositor, per Local Fork Supremacy and the operator decision
(INPUT-STACK-LINUX-ALIGNMENT-PLAN.md §4b). We own it and integrate upstream
into our tree rather than fetch+patch.
Does not affect redbear-mini/bare (they do not pull seatd). Making it compile
for Redox + wiring the compositor seat path remain part of the Mesa-gated
desktop bring-up.
Point evdevd at /scheme/input/consumer_raw (the new raw device tap added in
base bd5e3db3) instead of /scheme/input/consumer. The console consumer handed
evdevd keymapped, VT-gated events — it only received input while its own VT
was active, and the character had already been keymapped, both wrong for a
libinput/xkbcommon (Wayland) feed. The raw tap delivers the pre-keymap event
stream for every device event regardless of the active VT, matching the Linux
evdev contract.
evdevd only ever used scancode+pressed (never character), so no translation
change is needed on its side.
Bumps the base submodule pointer to include the consumer_raw tap. Part of the
Linux-aligned input-stack consolidation
(local/docs/INPUT-STACK-LINUX-ALIGNMENT-PLAN.md). Compile-checked for
x86_64-unknown-redox.
Phase 6.2 foundation. Port the Mini-MLD layer from Linux 7.1
(drivers/net/wireless/intel/iwlwifi/mld/), the MLD analog of Mini-MVM:
- linux_mld.h: iwlmld op-mode IDs, notification command groups
(LEGACY/MAC_CONF/PHY_OPS/DATA_PATH/SCAN/STATISTICS), and the MLD
notification command ID set. The RX frame IDs (0xc0/0xc1/0xc5) and the
iwl_rx_mpdu_desc descriptor are shared with MVM and stay in Mini-MVM;
MLD adds ROC/session-protection/MU-group-mgmt/probe-response/
channel-switch/MFUART/DTS-measurement notification IDs.
- linux_mld.c: rb_iwl_mld_handle_notification() — real MLD notification
dispatch. RX MPDU frames are parsed by the Mini-MVM descriptor parser
(rb_iwl_mvm_extract_signal/detect_format, no duplicate parser) with real
signal/rate extraction; MLD-specific notifications are identified,
grouped, and counted. Includes the BZ-family op-mode selection
(rb_iwl_opmode_for_family, mirroring Linux iwl-drv.c).
- linux_port.c: include linux_mld.h; add trans->opmode set from
device_family (BZ -> iwlmld, others -> iwlmvm); report opmode= in the
linux_kpi_status line so the gate is observable.
- build.rs: compile linux_mld.c.
This is the bounded foundation; MLO/link management and the full MLD
feature set (mac80211-integrated) are the multi-week continuation.
All 8 host tests pass.
Cross-referenced the console-input path against Linux/CachyOS
(local/reference/linux-7.1/drivers/{input,tty}/). Linux applies the
keymap once (vt/keyboard.c) and runs one line discipline (n_tty.c)
shared by the console tty and ptys; RedBear duplicates the keymap
(inputd + fbcond) and CR/LF (fbcond + ptyd) and runs two disciplines
bridged by getty — which is where typed characters are lost (echoed but
not executed). Plan: inputd owns the keymap and emits the full byte
stream (incl. escape sequences); fbcond becomes display-only passthrough
(no keymap, no CR/LF, no cooking); ptyd is the sole line discipline.
Staged, each stage independently bootable; validation needs the build
green (acpi-rs currently blocks it).
00_base.target, 10_evdevd, 10_smolnetd, 30_redox-drm retargeted to
driver-manager services; stale dormant-era comments updated. Eliminates
'init: unit 00_pcid-spawner.service not found' from the boot log.
- AGENTS.md + docs/README.md: pcid-spawner is retired and removed
(2026-07-24); driver-manager owns the path unconditionally; services
list and networking description updated.
- Migration plan v3.1 section: fallback framing replaced with the full
retirement (crate, service files, build wiring removed; source in git
history).
- Assessment: pcid-spawner row updated to retired; the never-delete
constraint noted as superseded by the operator's retirement decision.
Repairs the broken initialize_namespace that 0c11c2b5 captured (unterminated
comment) and lands the connect_op_regions() function. Base compiles again;
_REG opregion connect is now fully wired (acpi-rs function + acpid call).
NOTE: aml/mod.rs still has a broken initialize_namespace from 0c11c2b5
(parallel session captured an intermediate edit via git add -A; their
repeated git reset on base is reverting my concurrent fix before it can
be committed). acpi.rs connect_op_regions() call is in; the acpi-rs
connect_op_regions() function + initialize_namespace repair still needs
to land once base edits stop being reset.
The stash loop ended with '[ any_dirty = 0 ] && echo ...' as the
function's last command — when a fork IS dirty (any_dirty=1), the
compound returns 1 and set -e kills the build silently right after the
stash phase. Only visible with REDBEAR_ALLOW_DIRTY=1 (the dirty gate
refuses earlier otherwise). Now an explicit if + return 0.
Redesign grounded in the Linux 7.1 quirk-consumption cross-reference
(9 families, one three-layer invariant: match -> store -> consume):
- New open domain channel: [[<domain>_quirk]] TOML tables matched by
(vendor, device) with wildcards, accumulated as string flags by
quirks::lookup_driver_quirks(domain, vid, did). The table name IS the
domain key — new driver domains need zero registry code (Linux Type-C
model: driver-owned fixup tables like HDA's per-codec lists).
- lookup_audio_quirks(vid, did) as the first domain convenience API for
ihdad's future integration.
- quirks.d/15-audio.toml converted from [[pci_quirk]] to [[audio_quirk]]
— the audio_* flags (force_eapd, single_cmd, position_fix_lpib,
mirroring Linux sound/pci/hda) are now carried as data instead of
warned-and-dropped on every probe cycle.
- Docs: three-layer model in QUIRKS-SYSTEM.md + consumption contract
(bind-time / core-runtime / driver-runtime); QUIRKS-IMPROVEMENT-PLAN.md
v2.0 with the redesign assessment and cutover-era reality (stale
pcid-spawner broker references removed).
- 75/75 redox-driver-sys tests pass (new domain loader test).
- base-initfs recipe: pcid-spawner removed from initfs BINS and the
lib/pcid.d initfs staging dropped (dead pcid-spawner config).
- acid.toml, redoxer.toml: requires_weak now points at
00_driver-manager.service.
- redbear-mini.toml: dead /etc/pcid.d/* staging removed; driver-manager
service override no longer gated on the retired fallback flag.
- redbear-full.toml: dead legacy-format /etc/pcid.d/ihdgd.toml and
virtio-gpud.toml removed (driver-manager's 30-graphics.toml covers
the same drivers in the current format).
- Comments updated: no fallback exists; driver-manager owns the path
unconditionally. Base submodule bump (0c11c2b5).
The driver-list filter ran `grep -R "^name = \"$bin\"$" $COOKBOOK_SOURCE`
once PER bin (~40 bins). $COOKBOOK_SOURCE is local/sources/base, whose
target/ build tree is ~8.7 GB, so each grep re-scanned gigabytes and the
whole loop scanned hundreds of GB — a very slow build phase.
Collect the workspace's crate names ONCE from Cargo.toml files (skipping
target/ via --exclude-dir), into a bash associative array, then do an
O(1) membership check per bin. Verified equivalent (same bins selected)
and near-instant: the single scan of local/sources/base completes in
~0.02s vs minutes for the per-bin recursive greps.
Two changes to the Redox minimal input backend patch:
1. read_input_line accumulated into a *fresh* String each retry. On the
Redox console pty, a `read_line` can return WouldBlock (or Ok(0))
after only a partial line — the characters just typed — and those
bytes, left in the read_line buffer, were discarded on every retry.
Only a fully-queued line (e.g. a leftover newline) ever came through,
so the shell echoed keystrokes but never executed the command.
Fix: accumulate partial reads into a thread-local buffer that
persists across retries and only yield a line once a newline arrives.
2. Removed all BRUSH-DBG / rb_dbg startup and read tracing (the "debug
mode" output on the console). The functional Redox changes stay:
tokio current-thread runtime and the minimal input backend selection.
Verified: brush-interactive + brush-shell compile. Runtime verification
of the interactive shell still pending a boot test.
- Plan v3.2: runtime gate PASSED (initfs ahcid bind, switchroot, rootfs
e1000d concurrent bind with zero deferred, scheme live, resident
hotplug loop); the four runtime-only bugs it found and their fixes;
cookbook cargo-path-dep staleness fix verified.
- Assessment § 12: gate evidence table + runtime-only bug register.
- PACKAGE-BUILD-QUIRKS.md: new platform quirk — std::thread::scope
hangs on the Redox target (park/unpark unvalidated); spawn+join is
the required pattern; RwLock::write and park/scope remain unvalidated.
- Pointers bumped to v3.2.
Runtime validation of the cutover in QEMU (q35, e1000 + AHCI):
- thread::scope's park/unpark path hangs on Redox (scoped worker
completed all work but the scope join never returned — the boot
stalled inside every concurrent enumeration). The concurrent probe
pool now uses plain thread::spawn + JoinHandle::join (all captures
were already owned/Arc); the counting semaphore is Arc-based so
guards are 'static. bound map is Mutex (RwLock write was unproven
on target).
- initfs driver-manager no longer registers scheme:driver-manager —
the transient initfs manager's registration survived into the
rootfs phase and made the resident manager's registration fail
EEXIST (which then exit(1)'d the rootfs manager).
- config: matchless [[driver]] entries now parse (serde default) —
70-usb-class.toml's USB-class drivers (no [[driver.match]]) broke
config loading entirely on the first gate. Includes a regression
test for load_all with matchless entries.
- graphics: 30-graphics.toml moves from the shared
redbear-device-services.toml to redbear-full.toml (redox-drm is a
full-only driver; mini no longer defers it every hotplug cycle).
vesad removed from drivers.d — it is an init-managed service, not
a spawnable driver.
Gate evidence: initfs 'bound: 0000--00--1f.2 -> ahcid', switchroot,
rootfs 'bound: 0000--00--02.0 -> e1000d' with ZERO deferred,
scheme:driver-manager registered, resident hotplug loop (250ms),
pcid-spawner dormant on both phases.
Record that redbear-mini now boots to the Red Bear login: prompt in QEMU.
Document the root cause (dw_i2c::endpoint::serve register_sync_scheme
never returned the cap fd to init — ptyd-class deadlock on type={scheme}
services) and the fix (deliver cap fd via INIT_NOTIFY). Note follow-ups:
_REG opregion-connect is dead code (initialize_namespace never called),
and driver-manager maps class=0x03 to full-only redox-drm in mini.
The content-hash cache covered recipe-level PKGARs and the recipe's own
source tree, but not Cargo 'path =' dependency sources inside a crate —
the exact pattern the local fork model mandates (redox_syscall, libredox,
redox-driver-*, pcid_interface). Editing a path-dep's source left the
dependent recipe cached and stale. Observed live during the
driver-manager QEMU gate: 'cook driver-manager - cached' while
redox-driver-core sources had changed; the markers never reached the ISO.
Fix:
- New cook::cargo_path_deps module: resolves path deps from Cargo.toml
([dependencies], dev/build deps, target-specific deps, [patch.*]
tables, workspace members) recursively with cycle protection, and
hashes each resolved source tree (sorted relpath+content walk,
excluding .git/target — same policy as SourceContentHash).
- DepHashes gains a cargo_path_deps table (serde default for backward
compatibility; a missing field invalidates once, then re-syncs).
- The rebuild decision now ORs recipe-PKGAR changes with cargo path-dep
tree changes; the post-build write records the new map.
Acceptance: cook -> cached; content edit in redox-driver-core ->
'DEBUG: cargo path-dep source hashes changed' + rebuild; revert ->
rebuild once -> cached. 48 tests pass (43 cookbook lib + 5 new).
read_msix_capability previously read the whole config file unbounded
(pc hang: the config char device never EOF'd at the time) and then
returned None unconditionally. Now it reads at most 256 bytes (the
standard capability chain region) and parses the chain for real:
status bit, cap list walk, MSI-X (0x11) message-control table size.
Includes base submodule bump for the pcid config EOF fix.
Completes the stale-tree cleanup: redox-driver-core/Cargo.toml (left
target-less after the src/ removal, breaking cargo workspace search),
driver-manager/Cargo.toml and driver-manager/src/*.rs (Jul-10 duplicate
sources whose exec.rs predates the live tree's P0-3 removal). Live
trees are */source/.
- base bump (b0e76065): pcid PCI 3.0 fallback no longer panics on
extended config reads (returns 0xFFFFFFFF, drops writes).
- redox-driver-pci PciBus reads only the 64-byte standard header it
consumes (vendor/device/class/subsystem), instead of the whole 4 KiB
config file per device per poll cycle — safe on MCFG-less machines
and cheaper everywhere.
The src/ directory held a July-10 snapshot of the crate from before
dynids, the concurrent path, and the modern-technology helpers — a
duplicate that could only mislead. The canonical tree is
local/recipes/drivers/redox-driver-core/source/src/.
Plan v3.1 records the completed P0/LDR/P2 work program and the
operator-ratified cutover (driver-manager owns the boot path in all
redbear-* configs; pcid-spawner gated as fallback). Assessment § 11
documents the OS-services integration: init ConditionPathExists gate,
live driver-params bridge, udev-shim driver-binding gap (P3), and the
correct absence of D-Bus (bridge-on-demand pattern).
Operator-ratified cutover (the D5/C-phase gate):
- redbear-device-services.toml: driver-manager added to [packages]
(was 'intentionally not included'); /lib/drivers.d/70-wifi.toml
staged (redbear-iwlwifi --daemon); stale pre-cutover comments
replaced with the single-spawner invariant for driver-manager.
- redbear-mini.toml: /etc/init.d override now stages
00_driver-manager.service (--hotplug, gated !disabled) instead of
00_pcid-spawner.service; all requires_weak references switched to
00_driver-manager.service.
- redbear-full.toml, redbear-greeter-services.toml: same requires_weak
switch for iommu, greeter, SDDM.
- base submodule bump (c72d4247): init.d/init.initfs.d service gates —
pcid-spawner services start only when
/etc/driver-manager.d/disabled exists (operator fallback, never
deleted); driver-manager services run by default.
make lint-config: OK — no init service path violations.
init now supports systemd-style ConditionPathExists (with ! negation)
in service files, and the C0 driver-manager service file has correct
polarity. The C0 dormancy is now by design (condition gate) instead of
by parser rejection accident.
vasilito