Updates kernel submodule to 573b3e6 which replaces context::current()
with context::try_current() in excp_handler(), preventing the cascading
'not inside of context' panic when a page fault occurs during BSP's
start() before context::init() runs.
The cmake-generated Qt6ShaderToolsConfig.cmake has an empty
extra_cmake_include list. Patch it to include Qt6ShaderToolsMacros.cmake
so qt_internal_add_shaders is defined for downstream modules.
cmake install skipped this 405-line macros file that defines
qt_internal_add_shaders/qt6_add_shaders. Without it, downstream Qt
modules (qtdeclarative) fail at cmake configure with 'Unknown CMake
command qt_internal_add_shaders'.
Qt6's standalone module build installs qsb binary but does not generate
the Qt6ShaderToolsTools cmake wrapper package needed by cross-compile.
Added a Python generator that creates the minimal cmake package files
(Config, Targets, Targets-release, Precheck, AdditionalTargetInfo,
VersionlessTargets, Dependencies, ConfigVersion) following the exact
pattern of Qt6WaylandScannerTools from qtbase host build.
qtshadertools cross-compile needs the host 'qsb' (Qt Shader Builder)
tool, just like qtbase needs host moc/rcc/uic. Added a native cmake
build step that compiles qsb from the 6.11.1 source and installs it
into the shared qt-host-build prefix.
Also regenerate pam-redbear Cargo.lock (stale after 0.2.5 version bump
of redbear-login-protocol path dependency).
The host tool build profile was hardcoded with '6.11.0', causing the
stale 6.11.0 host tools to be reused for the 6.11.1 cross-compile.
Updated profile string to invalidate old host build.
Upstream Qt 6.11.1 already absorbed the #if QT_CONFIG(opengl) guards in
qwaylandclientbufferintegration_p.h that our redox.patch was adding.
Removed the now-obsolete hunks per patch governance (rebase, drop if
upstream absorbed it).
All remaining hunks apply cleanly against 6.11.1 with minor line offsets.
- Cookbook Cargo.toml: 0.1.0 → 0.2.5
- All 61 in-house crate Cargo.toml versions: 0.2.4 → 0.2.5
- os-release.in: fix URLs from github.com to gitea.redbearos.org
- sync-versions.sh --check passes with zero drift
The OS version is derived from the git branch name at build time.
Building on branch 0.2.5 produces os-release with VERSION_ID=0.2.5.
Records actual recipes bumped in 0.2.5 on 2026-07-02:
- Qt stack 6.8.2/6.11.0a1 -> 6.11.1 (all 6 sub-recipes, verified BLAKE3)
- Wayland/DRM/Input/expat/seatd to upstream latest stable (8 recipes,
verified BLAKE3)
- KWin 6.3.4 -> 6.7.2 + kdecoration 6.3.4 -> 6.7.2 + konsole 24.08.3
-> 26.04.3
Plus structural fix: created the missing qtshadertools recipe so the
qtdeclarative dependency chain resolves.
Documents what was deliberately NOT done:
- KF6 6.10 -> 6.27.0: 38 frameworks, 17-minor delta. Per AGENTS.md
patch-governance, no commit that bumps recipe.toml URLs without
first rebasing the local patches can be made honestly. Rebase
work (17-minor * 38 recipes) is multi-day and recorded as open.
- Mesa 24.0.8 -> 26.1.4: blocked on redox-os/mesa fork rebase
(0.3.0 work).
Includes the rebase order for the next session that plans to run
'make all CONFIG_NAME=redbear-full' against the bumped pins.
Bump KDE Plasma + KDE utility recipes to upstream latest stable
on 2026-07-02.
Versions resolved against download.kde.org/stable/plasma/ and
invent.kde.org/plasma/* tags:
kwin 6.3.4 -> 6.7.2 (invent.kde.org/plasma/kwin)
kdecoration 6.3.4 -> 6.7.2 (invent.kde.org/plasma/kdecoration)
konsole 24.08.3 -> 26.04.3 (invent.kde.org/utilities/konsole;
note: KDE utility versioning switched
from YY.MM calendars to v26.04-style)
BLAKE3 hashes verified against the actual downloaded upstream tarballs.
State of source trees on disk (NOT touched by this commit):
- local/recipes/kde/kwin/source/ : still KWin 6.6.5 (prior
session imported 6.6.5 source; new tarball at v6.7.2 will replace
on next repo fetch).
- local/recipes/kde/kdecoration/source/ : still 6.3.4
- local/recipes/kde/konsole/source/ : still 24.08 (RELEASE_SERVICE_
VERSION_MAJOR/MINOR).
Per AGENTS.md fork-adaptation policy, patches in local/patches/
{kwin,kdecoration,konsole}/ must be re-applied against the new
upstream source trees after fetch; rebase is open work for the
next session. Disabling patches or wrapping with feature flags is
NOT acceptable per the in-tree stub/workaround zero-tolerance rule.
This commit does NOT bump any KF6 framework recipe (38 recipes,
17-minor upstream delta). That work is multi-day patch rebase and
remains open.
Bump the lower-delta graphics-stack lane to real upstream latest stable
on 2026-07-02. Per AGENTS.md fork-adaptation policy, the local patches
in local/patches/{libdrm,libwayland,libevdev,libinput}/ must be re-applied
against the new source trees before the next build; rebase is open work.
Versions resolved against authoritative upstream registries (real latest stable):
libwayland 1.24.0 -> 1.25.0 (gitlab.freedesktop.org/wayland/wayland)
wayland-protocols 1.38 -> 1.49 (gitlab.freedesktop.org/wayland/wayland-protocols)
libdrm 2.4.125 -> 2.4.134 (gitlab.freedesktop.org/mesa/libdrm)
libxkbcommon 1.7.0 -> 1.9.2 (github.com/xkbcommon/libxkbcommon mirror)
libevdev 1.13.2 -> 1.13.6 (freedesktop.org/software/libevdev)
libinput 1.30.2 -> 1.31.3 (gitlab.freedesktop.org/libinput/libinput)
seatd-redox 0.9.1 -> 0.9.3 (git.sr.ht/~kennylevinsen/seatd)
expat 2.5.0 -> 2.8.2 (github.com/libexpat/libexpat)
BLAKE3 hashes verified against the actual downloaded upstream tarballs.
Not changed (already at or near upstream latest):
- dbus 1.16.2 (== upstream latest)
- xkeyboard-config (no standalone recipe; consumed via libxkbcommon)
- linux-input-headers (Red Bear original, not upstream)
Patches NOT yet rebased: see local/patches/{libdrm,libwayland,libevdev,
libinput}/. The dependency surfaces they patch (libdrm 2.4.134 has new
DRM modifier code, libwayland 1.25.0 has new server-decoder helpers,
libinput 1.31 has new touchpad gesture tables) will need review before
re-fetch.
Bump the entire Qt 6 recipe surface to 6.11.1 as resolved from
download.qt.io on 2026-07-02. Per AGENTS.md fork-adaptation policy,
patches in local/patches/qtbase/* and local/patches/qtdeclarative/*
must be re-applied against the 6.11.1 source tree after this commit;
rebase is open work for the next session.
Verified BLAKE3 hashes for the 6.11.1 tarballs:
qtbase c3b83023dc54f1173831bbc80abca1901418ef517875bf8071a4895d3c4a3162
qtdeclarative 10f2d0662047ceb0ef221b725b59e7fec5c9092a4c10d5acc7daefea5f11b962
qtwayland 154b80972e472b10330c82d3b171a915959a5d06139289d5b898c16c58de4de8
qtsvg 49b947e1a96bf0a29a1ee84c231a518a1413d9f3ec360617e405400e510508b2
qtshadertools 24dcd88b9e752a380067182687032b2139d2f6220d64e4193428434970102ae2
qtsensors 52ad8a724bc34f724feef197cf29f1cb535831ddd0fbad6e9dfedaa01eef1379
Also:
- qtbase: bumped from 6.8.2 -> 6.11.1. The 6.11.0 source tree had been
imported under local/recipes/qt/qtbase/source/ by a prior session
without updating the recipe.toml. This commit aligns the recipe
with both the imported source and the resolved upstream latest.
- qtshadertools: NEW recipe created. The recipes/qt/qtshadertools
symlink was dangling (target directory did not exist), making
qtdeclarative's qtshadertools dependency unresolvable. Now wired
up following the qt6-sensors recipe pattern. Source tar URL is
the resolved 6.11.1 upstream; per-repo fetch will populate
source/ on next build.
What's NOT done (deliberately):
- Patch rebase in local/patches/qtbase/P*.patch and
local/patches/qtdeclarative/P1-skip-tools-crosscompile.patch.
These must be re-applied against the 6.11.1 source tree before
the next build. AGENTS.md policy: rebase, do not remove.
- KF6 6.10 -> 6.27.0 bump (38 framework recipes, 17-minor delta).
Deferred — multi-day patch-rebase work, out of scope for one session.
See local/docs/0.2.5-GRAPHICS-FREEZE-PLAN.md §2.2.
- KWin 6.6.5 -> 6.7.2 + wayland-protocols/libdrm/libwayland/...
bumps (remaining graphics recipes). Deferred.
- No build was attempted. recipe.toml pin is now consistent with
resolved upstream latest-stable; no source tree replacement has
happened.
Plan-only artifact for the 0.2.5 graphics path. No recipe.toml changes,
no build attempted. Documents:
- Per-recipe upstream-latest-stable versions resolved via download.qt.io,
download.kde.org, gitlab.freedesktop.org, and git ls-remote on
invent.kde.org projects (no human guessing).
- Pre-build actions required (re-pull, re-blake3, validate-patches).
- Patch surface to re-evaluate across 17-minor KF6 bump and 1-minor
KWin bump.
- Mesa fork situation (24.0.8 vs 26.1.4): freeze at current fork,
rebase is 0.3.0 work.
- Freeze-when-green criteria for cutting 0.2.5.
- Out-of-scope items (Plasma workspace, real libepoxy, etc.).
Decisions in this plan are reversible; no recipe.toml field has been
modified. All work-in-progress from prior session is preserved as
uncommitted files in the working tree.
The build-redbear.sh script auto-stashes working tree changes
in nested relibc and base source trees before running the
build. These changes were captured by the failed kernel
build attempt that hit the json-target-spec / kernel rust
toolchain mismatch (fixed in 0.2.5 by creating the local
0.2.5 branch).
Captured changes:
- local/recipes/kde/* : KDE Frameworks 6 source CMakeLists
whitespace changes from the autostash (preserved)
- local/recipes/qt/qtbase/* : qtypes.h whitespace from the
autostash (preserved)
- local/sources/kernel/Cargo.lock : dependency lock from
the kernel relibc rebuild attempt
- local/sources/kernel/src/lib.rs : touched (mtime) by the
build script's touch + make prefix command
This is a bookkeeping commit — the actual code changes
for the threading plan are on the 4 submodule branches
(kernel, relibc, base, libredox) and will be pushed
separately.
0.2.5 branch was created from 0.2.4 (HEAD cd3950072e) to
continue Phase 0 of the multi-threading plan work in a
clean branch.
When a git-sourced recipe's source/ directory exists but has no
embedded .git (e.g., a cleanup pass removed .git directories from
build-cache sources, or the dir was extracted from an archive without
.git), the cookbook previously hard-bailed with
'{:?} is not a git repository, but recipe indicated git source'
This required manual intervention: the operator had to find the
broken source/ dir, rm -rf it, and re-run the build. With many
local recipes that use git URLs and embedded .git directories as
build caches (e.g., local/recipes/dev/ninja-build, local/recipes/kde/*),
this happened easily.
Fix: detect the missing .git, wipe the source dir, and re-clone from
the recipe's git URL. The fresh-clone logic is extracted to a new
reclone_git_source() helper used by both the initial-fetch path and
the self-heal path. After the self-heal, the source/ has a valid
.git and the rest of the fetch flow continues normally.
Tested by: deleting local/recipes/dev/ninja-build/source/.git (the
exact regression that triggered this fix) and running
./local/scripts/build-redbear.sh --upstream redbear-mini
which now self-heals instead of hard-failing.
Prepend an UPDATE block at the top of the plan document recording:
- The 8 commits that landed Phase 0c (futex sharding, per-CPU run
queues, vruntime, work stealing, load balancing, cache-affine,
initial placement, NUMA topology, proc scheme handles, fadt fix)
- The upstream-redox kernel audit finding (upstream has none of
these features; local fork is sole implementation)
- A plan-vs-actual state table showing which claimed 'missing'
features are now present
- The kernel-side Phase 0c is complete; remaining work is
relibc-side (Phase 0e) and futex-REQUEUE/PI/robust (Phase 1)
The detailed §1–§9 analysis is preserved unchanged as historical
record. The status column 'Missing' in §1 should be re-read as
'now present in local kernel fork, pending relibc userspace wiring.'
cargo check now exits 0 with 0 errors in the local kernel fork.
The local kernel fork just gained 3 commits:
- add Context::set_sched_policy and set_sched_other_prio
- acpi/fadt: fix pre-existing usize/u32 type mismatch on x86_64
- add SchedPolicy/Name/Priority proc scheme handles
After these, the local kernel fork:
- Has 13 of the 18 P5-P9 plan patches re-applied (3 obsolete by
refactor, 1 misnamed, 1 already in fork from P4)
- cargo check exits 0 with 0 errors
- Provides userspace API for pthread_setname_np, sched_setscheduler,
and setpriority via /proc/<tid>/{name, sched-policy, priority}
- Fixes a long-standing pre-existing ACPI FADT type mismatch
Phase 0c (patch recovery) is functionally complete for the kernel
side. Remaining P5-P9 work is in relibc fork (P3/P5/P7/P9 threading
patches) which is a separate, more self-contained effort.
The local kernel fork just gained two commits:
- P6-futex-sharding: replaces single global Mutex<L1, FutexList> with
64-shard hash table, eliminating contention between futex operations
on different addresses.
- RUN_QUEUE_COUNT: pre-flight constant at context::RUN_QUEUE_COUNT = 40
that downstream patches (P8-percpu-sched, P8-percpu-wiring) need but
none of the existing P5–P9 patches define.
Phase 0c, plan orders #1 and pre-flight.
The relibc fork at local/sources/relibc just gained a one-line
correctness fix for pthread_cond_signal (was calling broadcast() which
wakes all waiters; POSIX requires wake exactly one). The fix was
committed to the fork as 6caad3a5 and pushed to the submodule/relibc
branch on the canonical RedBear-OS repo.
This parent-repo commit updates the gitlink so a fresh clone picks
up the new relibc fork HEAD.
First execution of the multi-threading plan (Phase 0a).
The working tree had accumulated git-tracking drift across the
local/sources, local/recipes/*/source, and local/reference trees.
Restored:
- local/sources/libredox: add missing 160000 gitlink at
d01da350 (submodule/libredox). The .gitmodules entry already
declared this fork; the parent tree entry was missing so a
fresh clone of the parent would not pull the libredox source.
- .gitignore: mark the four local/recipes/*/source build-cache
trees (uutils-tar, ninja-build, sddm, sddm/source-pristine)
and the two local/reference/* entries (linux-7.1, seL4) as
ignored. These are build caches and external references, not
durable Red Bear code. The durable code for the four recipes
is recipe.toml + the corresponding patch (redox.patch).
- Note in .gitignore: do not extend local/recipes/**/source to
a blanket rule, because ~150 Red Bear fork recipes do keep
their durable source under local/recipes/<name>/source/.
Removed six broken 160000 gitlinks:
- local/recipes/archives/uutils-tar/source (e4c2affa...): on-disk
working tree was a self-clone of RedBear-OS; gitlink pointed to
a non-existent commit in the parent object database.
- local/recipes/dev/ninja-build/source (d829f42b...): gitlink
was a dangling commit on a diverged branch that has since been
rewritten; the on-disk HEAD is upstream v1.13.1 (79feac0) which
the recipe re-fetches via recipe.toml anyway. The 6.4MB
embedded .git directory was also removed.
- local/recipes/kde/sddm/source (63780fcd...): build cache for
sddm 0.21.0 re-fetched via recipe.toml. The 11MB embedded
.git directory was also removed.
- local/recipes/kde/sddm/source-pristine (63780fcd...): empty
placeholder, build cache. Removed.
- local/reference/linux-7.1 (ab9de95c...): external Linux
reference tree, gitignored by size. The on-disk directory
is preserved per AGENTS.md 'NEVER delete the reference tree'.
- local/reference/seL4 (a0b4f2d2...): empty placeholder,
gitignored.
Removed untracked pollution at repo root:
- kernel (empty 0-byte file)
- qqmljsgrammar.cpp, qqmljsgrammar_p.h, qqmljsparser.cpp,
qqmljsparser_p.h (393KB total: build artifacts that escaped
a qtdeclarative build into the working tree root; they belong
inside the recipe source tree, not at the parent level)
Added:
- local/docs/MULTITHREADING-COMPREHENSIVE-ASSESSMENT-AND-PLAN.md:
comprehensive multi-threading audit and implementation plan
covering kernel scheduler, kernel futex, syscall ABI, relibc
pthreads, and userspace threading correctness. Will drive
the next implementation cycle after the git tracking work
is wrapped.
After this commit:
- 9 submodule entries in HEAD, all of local/sources/* forks.
- All previously-existing 8 fork submodules unchanged.
- libredox is now durable across clones (was previously lost).
- No untracked files at root.
- No dangling or self-referencing gitlinks.
vasilito
Red Bear Build System