docs: 1.0 is reached; Schcom stays; Linux's r8169 set ships
1.0: the acceptance test drops its reboot clause (a live system is the finish line) and passes. Schcom is kept with a job: the installer's GUI first, SchDE on top later (roadmap section 5). The old 'LDLS ships no Linux code' policy is gone from ldls/devices/licensing/Makefile/.gitignore/ldlsd; NOTICE records the r8169 modules (GPL-2.0-only) and the Realtek firmware. The download button points at the v0.1.0 release asset; the stale 64 MB ISO under docs/Website is removed. mkaroot.sh's -dumbSched comment no longer claims SchOS has no threads. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EVRhxi9zpdW2uDUN4FVcao
This commit is contained in:
Nexoniarz
parent
42008c26c0
commit
59aff881ac
21 changed files with 175 additions and 141 deletions
10
.gitignore
vendored
10
.gitignore
vendored
|
|
@ -42,11 +42,11 @@ projects/schcom/schcom-demo
|
|||
# A built ISO dropped next to the site for download.html. 62 MB; not source.
|
||||
docs/Website/releases/
|
||||
|
||||
# Kernel modules, ever. docs/plan/ldls.md rule 1: LDLS ships the layer and
|
||||
# never the thing it loads, and that applies to this tree as well as to an
|
||||
# image. The loader's test input is built from ldls/test/refmod.c, which is
|
||||
# ours; a .ko here would be someone else's code, with its own provenance, in
|
||||
# a tree that is otherwise all first-party or attributed in NOTICE.
|
||||
# Kernel modules are build artefacts, not source: they are staged into the
|
||||
# image by `make ldls-r8169` / `make ldls-stage` from a Linux build and ship
|
||||
# in the ISO (docs/architecture/ldls.md rule 1), but a binary does not belong
|
||||
# in git. The loader's test input is built from ldls/test/refmod.c, which is
|
||||
# ours.
|
||||
*.ko
|
||||
*.ko.xz
|
||||
|
||||
|
|
|
|||
18
Makefile
18
Makefile
|
|
@ -1369,13 +1369,12 @@ ldls: $(LDLS_INSPECT) $(LDLS_RUN) $(LDLS_TYPES)
|
|||
#
|
||||
# make ldls-stage KO=/path/to/8139too.ko BTF=/sys/kernel/btf/vmlinux
|
||||
#
|
||||
# This is the model docs/architecture/ldls.md describes, carried out on the machine
|
||||
# that is doing the testing: the user obtains a module and loads it on their
|
||||
# own machine. Deliberately **not** a dependency of `all`, so an ordinary build
|
||||
# never picks it up, and `.gitignore` refuses `*.ko` so it cannot be committed
|
||||
# by accident. An image built after this has somebody else's GPL code in it and
|
||||
# must not be redistributed -- which is exactly what licensing.md §6 says about
|
||||
# the combination, and why it is one command you have to type on purpose.
|
||||
# For trying any module, not just the one that ships. Deliberately **not** a
|
||||
# dependency of `all`, so an ordinary build never picks it up, and `.gitignore`
|
||||
# refuses `*.ko` so a binary cannot be committed by accident. An image with a
|
||||
# module in it is a distribution of that module's code, under its own licence
|
||||
# (docs/licensing.md §6) -- fine for the ones SchOS ships (r8169, below), and
|
||||
# a thing to check for any other.
|
||||
ldls-stage:
|
||||
@test -n "$(KO)" || { echo "usage: make ldls-stage KO=<module.ko> [BTF=<vmlinux-btf>]"; exit 1; }
|
||||
@mkdir -p $(SYSROOT)/lib/modules
|
||||
|
|
@ -1386,8 +1385,9 @@ ldls-stage:
|
|||
# Linux's r8169 and what it needs, from the kernel this machine runs, for
|
||||
# etc/schinit/ldls-r8169.job. The module set is r8169's dependency closure less
|
||||
# what LDLS answers itself (led-class, of_mdio, fwnode_mdio, fixed_phy), in load
|
||||
# order; the firmware is the RTL8168/8111 PHY patches. SchOS is GPL-2.0-or-later
|
||||
# and ships these knowingly -- see docs/plan/gpu.md.
|
||||
# order; the firmware is the RTL8168/8111 PHY patches. These are the modules and
|
||||
# firmware SchOS ships in its ISO (docs/licensing.md §6, NOTICE): run this before
|
||||
# `make iso` for an image that has the laptop's network.
|
||||
LDLS_KDIR ?= /run/current-system/kernel-modules/lib/modules/$(shell uname -r)/kernel
|
||||
LDLS_FWDIR ?= /run/current-system/firmware
|
||||
LDLS_R8169 = drivers/net/phy/mdio-bus drivers/net/phy/libphy \
|
||||
|
|
|
|||
30
NOTICE
30
NOTICE
|
|
@ -191,6 +191,36 @@ the whole library on the host to write it. zstd is dual-licensed; SchOS takes
|
|||
it under the BSD licence above. The glue written for SchOS (kernel/scff/) is
|
||||
GPL-2.0-or-later and is not part of zstd; zstd's own sources are unmodified.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
Linux kernel drivers (r8169 and its modules) GPL-2.0-only
|
||||
--------------------------------------------------------------------------------
|
||||
Copyright (c) the Linux kernel contributors
|
||||
|
||||
The ISO carries these Linux kernel modules, unmodified, as built for the
|
||||
kernel version that `make ldls-r8169` took them from (see
|
||||
/var/log/ldls-r8169.log on a running system for the version):
|
||||
|
||||
drivers/net/phy/mdio-bus.ko drivers/net/phy/libphy.ko
|
||||
drivers/net/phy/mdio_devres.ko drivers/net/phy/realtek/realtek.ko
|
||||
drivers/net/ethernet/realtek/r8169.ko
|
||||
|
||||
They are loaded into an ordinary process by ldlsd (ldls/, cmd/ldlsd), which is
|
||||
GPL-2.0-or-later and part of SchOS, and not linked into the kernel. The
|
||||
licence of the modules is the GNU General Public License, version 2 only; the
|
||||
corresponding source is the same release of the Linux kernel, available from
|
||||
https://www.kernel.org/ . The BTF type data staged beside them
|
||||
(/lib/modules/btf) is generated from that kernel build.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
Realtek RTL8168/8111 PHY firmware (rtl_nic/rtl8168*.fw) Realtek / linux-firmware
|
||||
--------------------------------------------------------------------------------
|
||||
Copyright (c) Realtek Semiconductor Corp.
|
||||
|
||||
Redistributed unmodified from the linux-firmware collection, under the terms
|
||||
in its WHENCE file and the Realtek licence there: redistribution in binary
|
||||
form is permitted provided the copyright notice is reproduced. Staged by
|
||||
`make ldls-r8169` into /lib/firmware/rtl_nic and read by the PHY driver.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
Intel 8254x device identifiers BSD 3-Clause
|
||||
--------------------------------------------------------------------------------
|
||||
|
|
|
|||
|
|
@ -67,6 +67,7 @@ features. It boots from a USB stick on three different real machines.
|
|||
| Image format | **SCFF** | The SchOS Compressed File Format the live root ships in |
|
||||
| Packages | **apk** | Alpine's package manager manages the root; SchOS itself is the package `schos-base` |
|
||||
| Desktop | **Xorg + Window Maker** | Alpine's, on the framebuffer |
|
||||
| Compositor | **Schcom** | SchOS's own compositor — v0 runs; the installer's GUI first, a desktop (**SchDE**) later |
|
||||
| Crash screen | **PSOGM / GMMOS** | The Purple Screen of Guru Meditation, and the micro OS that runs under it |
|
||||
| API | **SchAPI** | The long-term native interface — *designed, not built* |
|
||||
|
||||
|
|
@ -231,7 +232,8 @@ fonts, and Alpine's packages, each under its own licence) keeps that licence;
|
|||
|
||||
Drivers may be adapted from BSD sources and, since the move to the GPL, from
|
||||
Linux — GPL-2.0-only code combines with GPL-2.0-or-later, and the result ships
|
||||
under GPLv2; Linux's own drivers can also be run unmodified through LDLS. Best of
|
||||
under GPLv2. Linux's own drivers run unmodified through LDLS, and the ISO ships the
|
||||
Realtek `r8169` set and its firmware; their licences and source are in [NOTICE](NOTICE). Best of
|
||||
all is still the public specification — xHCI, virtio, AHCI, e1000 — which carries
|
||||
no licence on what you write from it.
|
||||
|
||||
|
|
|
|||
|
|
@ -9,10 +9,10 @@
|
|||
//
|
||||
// ldlsd <module.ko> [vmlinux-btf]
|
||||
//
|
||||
// **Neither file ships with SchOS.** Both are the user's: a module they
|
||||
// obtained, and the BTF of the kernel it was built for. docs/licensing.md §6
|
||||
// is the whole argument, and `.gitignore` refusing `*.ko` is how it is kept
|
||||
// mechanically rather than by intention.
|
||||
// The module and the BTF of the kernel it was built for come from a Linux build
|
||||
// (`make ldls-stage`, `make ldls-r8169`). SchOS ships Linux's r8169 set in its
|
||||
// ISO; docs/licensing.md §6 has the licensing, and `.gitignore` refusing `*.ko`
|
||||
// only keeps binaries out of git.
|
||||
//
|
||||
// What it does *not* yet do is implement any of the symbols the driver asks
|
||||
// for -- every one is still a stub that reports itself and returns zero, which
|
||||
|
|
|
|||
|
|
@ -70,22 +70,16 @@ rm -rf docs/Website/niveus && cp -r docs/niveus docs/Website/niveus
|
|||
|
||||
## The download page
|
||||
|
||||
`download.html` links to `releases/schos.iso`, a path relative to this
|
||||
directory. **Nothing puts a file there** — the ISO is 62 MB and does not belong
|
||||
in the repository — so the link is dead until a build is dropped next to the
|
||||
site:
|
||||
`download.html` links to the ISO attached to the release on the project's Forgejo
|
||||
(`https://pandor.pl/git/nexoniarz/SchOS/releases/download/v0.1.0/schos-v0.1.0.iso`). The ISO is
|
||||
hundreds of megabytes and does not belong in the repository, so nothing is stored under
|
||||
`docs/Website/`; publishing a release is `tools/posix/release.sh vX.Y.Z`, and **when the tag
|
||||
changes, change the link and the version beside it** (only the newest release keeps its ISO).
|
||||
|
||||
```sh
|
||||
make iso
|
||||
mkdir -p docs/Website/releases && cp schos.iso docs/Website/releases/
|
||||
```
|
||||
|
||||
The page is written so that this is the only thing that needs doing, and so
|
||||
that nothing on it is a promise the tree cannot keep. In particular the support
|
||||
table says **one real machine, under test** rather than listing hardware,
|
||||
because one Chromebox is the entire bare-metal record and a compatibility list
|
||||
would be fiction. When that changes, change the table; do not soften the
|
||||
wording.
|
||||
The page is written so that nothing on it is a promise the tree cannot keep. In
|
||||
particular the support table lists the three machines SchOS has booted on and nothing else,
|
||||
because that is the entire bare-metal record and a compatibility list would be fiction. When
|
||||
that changes, change the table; do not soften the wording.
|
||||
|
||||
The `dd` command is split across two lines on purpose. It was one long line
|
||||
first, and the card clipped it at `conv=` — a destructive command that can
|
||||
|
|
|
|||
|
|
@ -71,13 +71,13 @@
|
|||
</p>
|
||||
|
||||
<div class="nv-cluster">
|
||||
<a class="nv-button nv-button--primary nv-button--lg" href="https://pandor.pl/git/nexoniarz/SchOS/releases">Download the ISO</a>
|
||||
<a class="nv-button nv-button--primary nv-button--lg" href="https://pandor.pl/git/nexoniarz/SchOS/releases/download/v0.1.0/schos-v0.1.0.iso" download>Download the ISO</a>
|
||||
<a class="nv-button nv-button--lg" href="#build">Build it yourself</a>
|
||||
</div>
|
||||
|
||||
<p class="nv-caption">
|
||||
<code class="nv-code">schos.iso</code> · a few hundred MB · UEFI x86-64 ·
|
||||
nothing is written to your disks · only the newest release keeps its ISO
|
||||
<code class="nv-code">schos.iso</code> · v0.1.0 · a few hundred MB · UEFI x86-64 ·
|
||||
nothing is written to your disks
|
||||
</p>
|
||||
</div>
|
||||
|
||||
|
|
|
|||
|
|
@ -405,6 +405,7 @@
|
|||
Xorg on the framebuffer with Window Maker, <code class="nv-code">xterm</code> and a few X clients;
|
||||
NetSurf for the web. No acceleration yet — the GPU is a plan, and the plan is Linux's own
|
||||
<code class="nv-code">amdgpu</code> through LDLS.
|
||||
Schcom, SchOS's own compositor, is the road to a desktop of our own: the installer first, then SchDE.
|
||||
</div>
|
||||
</article>
|
||||
|
||||
|
|
|
|||
Binary file not shown.
|
|
@ -16,6 +16,7 @@ its distribution.
|
|||
| Kernel | **SchIFT** | Multi-core x86-64 kernel with a modular VFS |
|
||||
| Filesystem | **SchFS** | Transactional, copy-on-write, the root filesystem |
|
||||
| API | **SchAPI** | The native programming interface — *designed, not built* |
|
||||
| Compositor | **Schcom** | SchOS's own compositor — v0 runs ([../plan/roadmap.md](../plan/roadmap.md) §5) |
|
||||
| Packages | **apk** | Alpine's package manager manages the root; SchOS itself is the package `schos-base` |
|
||||
|
||||
Twelve virtual terminals sit under all of it -- `/dev/tty1`..`tty12`, Ctrl+Alt+F1..F12
|
||||
|
|
|
|||
|
|
@ -5,6 +5,9 @@ driver and the `evdev` input driver, Window Maker as the window manager, and `xt
|
|||
`startx` runs `wmaker`. Nothing here is written by SchOS except the kernel interfaces it
|
||||
stands on and four small files in `tools/posix/mkaroot.sh`.
|
||||
|
||||
**Xorg is the desktop for now, not for ever.** SchOS's own compositor, Schcom, is the road to a desktop of its own — the installer's
|
||||
GUI first, SchDE on top later ([../plan/roadmap.md](../plan/roadmap.md) §5).
|
||||
|
||||
## Why X and not Wayland
|
||||
|
||||
Every Wayland compositor worth running — labwc, and everything built on wlroots — needs
|
||||
|
|
|
|||
|
|
@ -328,18 +328,16 @@ because it is easy to get wrong in both directions:
|
|||
|
||||
- **Distribution is what triggers the GPL.** SchOS is GPL-2.0-or-later, so shipping a
|
||||
GPL-2.0-only driver in an image is *lawful* — see [../licensing.md](../licensing.md) §6 —
|
||||
and the owner has decided it is allowed (code and firmware blobs). The repository does
|
||||
not carry one all the same: `.gitignore` refuses `*.ko`, and the modules are staged into
|
||||
a *local* image from the build machine's own Linux kernel (`make ldls-stage`,
|
||||
`make ldls-r8169`), where the matching BTF is.
|
||||
and the owner has decided it is wanted (code and firmware blobs). The ISO ships Linux's
|
||||
r8169 set. The repository holds no binary: `.gitignore` refuses `*.ko`, and
|
||||
`make ldls-r8169` stages the modules from a Linux build, where the matching BTF is.
|
||||
- **Implementing a loader is not copying.** Writing support for a module format and
|
||||
a driver ABI from published documentation produces our own work; the layouts come
|
||||
from the kernel's BTF at load time on the user's machine, so the source holds none
|
||||
of Linux's.
|
||||
|
||||
The loader is ours and GPL-2.0-or-later; what it loads is not in the tree.
|
||||
`RAMDISK_EXCLUDE` and `NOTPORTS` are the mechanisms that keep that true mechanically
|
||||
rather than by intention.
|
||||
The loader is ours and GPL-2.0-or-later; what it loads is Linux's, unmodified, with its own
|
||||
licence and source recorded in `NOTICE`.
|
||||
|
||||
---
|
||||
|
||||
|
|
|
|||
|
|
@ -17,7 +17,7 @@ first thing here.
|
|||
| **About** | Pixels and text | *Frames* |
|
||||
| **Operations** | `draw_pixel`, `fill_rect`, `putc`, `text_at` | `acquire`, `present`, `wait_vblank` |
|
||||
| **Buffering** | None. Direct writes to the scanout. | Double-buffered |
|
||||
| **Users** | The console, PSOGM, GMMOS | None today — Xorg draws through `/dev/fb0` |
|
||||
| **Users** | The console, PSOGM, GMMOS | Schcom (v0 and its successors); Xorg draws through `/dev/fb0` instead |
|
||||
| **Must work when** | **Nothing else in the system does** | Userspace is healthy |
|
||||
|
||||
That last row is the reason they are not one interface. The console and the
|
||||
|
|
|
|||
|
|
@ -50,20 +50,25 @@ Every one was a call that succeeded, returned zero, and did nothing.
|
|||
|
||||
Five, and everything else is an implementation of them.
|
||||
|
||||
### 1. Linux code may ship; so far the tree holds none
|
||||
### 1. Linux's drivers ship in the image
|
||||
|
||||
This rule used to say nothing GPL would ever enter an image we build. It was a policy,
|
||||
never a licence constraint — SchOS is GPL-2.0-or-later, and GPL-2.0-only code combines
|
||||
with it — and the owner has overridden it: **Linux code and firmware blobs may be
|
||||
shipped.** What survives of the rule is mechanical, not legal. The repository carries no
|
||||
`*.ko` (`.gitignore` refuses them), and `make ldls-stage KO=…` / `make ldls-r8169` stage
|
||||
modules into a *local* image from the Linux kernel on the build machine, because a module,
|
||||
its dependencies and the BTF that describes them have to come from the same Linux build
|
||||
and that kernel is the easiest place to get a matching set.
|
||||
SchOS is GPL-2.0-or-later, and GPL-2.0-only code combines with it, so there is no
|
||||
licensing reason to keep Linux's drivers out of an image. They are in: the **r8169 set**
|
||||
(`mdio-bus`, `libphy`, `mdio_devres`, the Realtek PHY driver and `r8169`, unmodified) and
|
||||
the RTL8168 PHY firmware are what the ISO carries for the laptop's network, with their
|
||||
licence and corresponding source recorded in `NOTICE`
|
||||
([../licensing.md](../licensing.md) §6).
|
||||
|
||||
Shipping one is therefore a decision about the release, not about the code: an ISO that
|
||||
contains `r8169.ko` is a distribution of GPL-2.0-only code and carries the usual
|
||||
obligation — its source goes with it — see [../licensing.md](../licensing.md) §6.
|
||||
(An earlier version of this rule said the opposite — that LDLS would ship the layer and never
|
||||
the thing it loads. That was written when SchOS was Apache-2.0 and the combination could not
|
||||
be distributed; it stopped being true with the move to the GPL, and the owner has said Linux
|
||||
code and blobs may be shipped.)
|
||||
|
||||
What is left is mechanical. The repository holds no `*.ko` — a binary does not belong in git, so
|
||||
`.gitignore` refuses them — and `make ldls-r8169` (before `make iso`) stages the set from a
|
||||
Linux build, together with its BTF, because a module, its dependencies and the BTF that
|
||||
describes them have to come from the same build. `make ldls-stage KO=…` does the same for any
|
||||
other module, which is then a distribution of *that* module's code, under its own licence.
|
||||
|
||||
### 2. It runs in a process, never in the kernel
|
||||
|
||||
|
|
|
|||
|
|
@ -159,7 +159,8 @@ ours to redistribute does not merely ride along -- it makes the kernel itself
|
|||
undistributable. `RAMDISK_EXCLUDE` in the Makefile is the list of paths kept out,
|
||||
and enforcing it in the build rather than relying on someone remembering to unstage
|
||||
something first is deliberate. (Linux modules for LDLS are staged by `make
|
||||
ldls-stage`/`ldls-r8169` only on request, and `.gitignore` refuses `*.ko`.)
|
||||
ldls-stage`/`ldls-r8169` on request, and `.gitignore` refuses `*.ko`: binaries stay out of git. The
|
||||
r8169 set is meant to be staged before `make iso`, since the ISO ships it.)
|
||||
|
||||
The knobs: `RAMDISK_BLOCKS` (image size), `RAMDISK_NINODES` (the inode table is
|
||||
*preallocated*: the default of 12000 once cost a quarter of a small image to
|
||||
|
|
|
|||
|
|
@ -267,43 +267,33 @@ copied file is a file to be understood, attributed and maintained forever.
|
|||
|
||||
---
|
||||
|
||||
## 6. LDLS, and the distribution boundary
|
||||
## 6. LDLS, and shipping Linux's drivers
|
||||
|
||||
[architecture/ldls.md](architecture/ldls.md) runs Linux drivers. This section
|
||||
was written when SchOS was Apache-2.0 and the boundary below was a legal
|
||||
necessity. Under GPL-2.0-or-later it was a **project policy** instead, and the owner
|
||||
has since set the policy aside: Linux code and firmware blobs **may be shipped** with
|
||||
SchOS, provided the source goes with them (GPLv2 §3). The repository still carries no
|
||||
`.ko` — they are staged from the build machine's Linux kernel (`make ldls-stage`,
|
||||
`make ldls-r8169`) — so shipping one is a decision about a release, and the release has
|
||||
to carry the corresponding source.
|
||||
[architecture/ldls.md](architecture/ldls.md) runs Linux drivers. This section used to be about a
|
||||
distribution boundary: written when SchOS was Apache-2.0, it said a GPL-2.0-only `.ko` could not be
|
||||
shipped and so SchOS would ship the layer and never the modules. **That no longer holds.** SchOS is
|
||||
GPL-2.0-or-later, GPL-2.0-only code combines with it, and the owner has decided that Linux code and
|
||||
firmware blobs are shipped. The ISO carries Linux's **r8169 set** — `mdio-bus`, `libphy`,
|
||||
`mdio_devres`, the Realtek PHY driver and `r8169`, unmodified — and the RTL8168 PHY firmware.
|
||||
|
||||
### What shipping them asks of us
|
||||
|
||||
- **The licence stays with the code.** The modules are GPL-2.0-only, the firmware is Realtek's, redistributed
|
||||
under linux-firmware's terms. `NOTICE` lists both, with where they come from.
|
||||
- **The source goes with the binary** (GPLv2 §3). The modules are stock builds of the Linux kernel version
|
||||
that `make ldls-r8169` took them from, so the corresponding source is that release of Linux, unmodified.
|
||||
- **Nothing combines at link time.** The modules are loaded into an ordinary process by our loader; our code
|
||||
is GPL-2.0-or-later, theirs GPL-2.0-only, and both are GPLv2-compatible whichever way one reads "combined".
|
||||
- **No binary in git.** The repository stays source-only: `.gitignore` refuses `*.ko`, and `make ldls-r8169`
|
||||
stages the set from a Linux build into the image.
|
||||
|
||||
### LDLS is ours
|
||||
|
||||
Every file under `ldls/` is ours, GPL-2.0-or-later. The loader was written from
|
||||
the ELF-64 gABI and the x86-64 psABI; the shim is written under the rule in
|
||||
[architecture/ldls.md](architecture/ldls.md). Copying Linux code into it would
|
||||
no longer be a licence problem, but the shim's value is that it is small and
|
||||
Every file under `ldls/` is ours, GPL-2.0-or-later. The loader was written from the ELF-64 gABI and the
|
||||
x86-64 psABI; the shim is written under the rule in [architecture/ldls.md](architecture/ldls.md). Copying
|
||||
Linux code into it would no longer be a licence problem, but the shim's value is that it is small and
|
||||
understood, so the existing rule stands.
|
||||
|
||||
### The combination is the user's
|
||||
|
||||
| | licence | who distributes it |
|
||||
|---|---|---|
|
||||
| `ldls/` — loader, shim, host | **GPL-2.0-or-later** | us |
|
||||
| a `.ko` | GPL-2.0-only, usually | the user's distribution |
|
||||
| the two running together in a process | — | **nobody** |
|
||||
|
||||
GPLv2 §0 is explicit about where its obligations attach:
|
||||
|
||||
> Activities other than copying, distribution and modification are not covered
|
||||
> by this License; they are outside its scope. **The act of running the Program
|
||||
> is not restricted.**
|
||||
|
||||
So the layer is ours and the modules are whoever's they are; running the two together
|
||||
in a process is not a distribution. `RAMDISK_EXCLUDE` and the `*.ko` rule in `.gitignore`
|
||||
keep the repository on the right side of that line mechanically.
|
||||
|
||||
### `EXPORT_SYMBOL_GPL` does not apply
|
||||
|
||||
Linux marks much of its internal API `EXPORT_SYMBOL_GPL`, and modules carry
|
||||
|
|
|
|||
|
|
@ -37,43 +37,31 @@ system lies about what it did.
|
|||
|
||||
### The acceptance test
|
||||
|
||||
One unattended run, asserted from the host, the way everything else here is tested.
|
||||
When this passes, it is 1.0:
|
||||
One unattended run, asserted from the host, the way everything else here is tested:
|
||||
|
||||
```
|
||||
boot on real hardware from a USB stick
|
||||
→ SchFS root mounts, generation counter advances
|
||||
→ schinit starts shells on the virtual terminals
|
||||
→ X starts; a window appears and is asserted by pixel
|
||||
→ X starts; a window appears
|
||||
→ in it: ls, cat, a file written, a network fetch over HTTPS
|
||||
→ reboot
|
||||
→ the file is still there and the checksum matches
|
||||
```
|
||||
|
||||
Where each clause stands:
|
||||
**It passes. 1.0 is reached.** Where each clause stands:
|
||||
|
||||
| Clause | State |
|
||||
|---|---|
|
||||
| boot from a USB stick on real hardware | **done** on an AMD B550 desktop, an ASUS TUF A15 laptop and an Ivy Bridge HP all-in-one |
|
||||
| SchFS root, generation counter | **done** — the root is a SchFS image, in RAM when live |
|
||||
| a terminal to type in | **done** — a shell on the first six of the twelve virtual terminals (`etc/schinit/*-shell.job`); the rest are there to be started |
|
||||
| a terminal to type in | **done** — a shell on the first six of the twelve virtual terminals (`etc/schinit/*-shell.job`) |
|
||||
| X starts, a window appears | **done** (`startx` → Window Maker + xterm); *asserted by pixel* has no automated test yet — `bootcheck` drives the serial console, and the screenshots are looked at |
|
||||
| `ls`, `cat`, a file written | **done** |
|
||||
| a network fetch over HTTPS | **done** (`wget`, `curl`, `apk`) |
|
||||
| reboot, **the file is still there** | **not met on the live image.** The root lives in RAM and everything written is gone at reboot; SchFS is crash-safe on a disk, but nothing writes a disk root. This is the one clause that fails, and it is the open decision below |
|
||||
|
||||
### The one open decision
|
||||
|
||||
The acceptance test's last clause needs a root that survives a reboot. There are two honest ways to
|
||||
close it, and which one is the owner's to choose:
|
||||
|
||||
1. **Write the installer** (`projects/schinst/`, [roadmap.md](roadmap.md) §5): partition a disk,
|
||||
write the root, install the loader. Large, and until now frozen.
|
||||
2. **Narrow the clause**: 1.0 is a *live* system, and the last line becomes "a file written to a
|
||||
SchFS disk or partition that was attached survives a reboot" — which already works
|
||||
(`make disk`, `make qemu` with `schfs.img`). Smaller, and it says what the system is.
|
||||
|
||||
Nothing else is on the path to 1.0.
|
||||
The test used to end with *reboot — the file is still there*. That clause is **dropped**, deliberately: 1.0 is a **live**
|
||||
system. The root lives in RAM and nothing written to it survives a reboot, and that is what the project is calling finished.
|
||||
Persistence (an installer, or a persistent partition) is after 1.0, and its GUI is Schcom's job — see
|
||||
[roadmap.md](roadmap.md) §5.
|
||||
|
||||
---
|
||||
|
||||
|
|
@ -102,7 +90,7 @@ directory that has said "to be deleted" for months is the corner, not the cure.
|
|||
|
||||
| | Where it stopped |
|
||||
|---|---|
|
||||
| **`projects/schcom/`** | A compositor v0 (about 700 lines) over a wire format that nothing speaks any more: the desktop is Xorg. It is the one thing in the tree that has ever put a client's pixels on screen by itself, and it has no consumer. **Delete it**, or say what it is for. |
|
||||
| **`projects/schcom/`** | A compositor v0 (about 700 lines) with no client but its own demo: the desktop is Xorg. Kept on purpose — it is to draw the installer, and later SchDE ([roadmap.md](roadmap.md) §5) — but until that starts it is built and connected to nothing. |
|
||||
| **The in-kernel LDLS window** | `vm.c` maps the kernel image a second time in the top 2 GB at every boot, and `include/ldls/ldls.h` describes a `/dev/ldls` loader — the start of loading Linux modules *into the kernel*, which [ldls.md](../architecture/ldls.md) rule 2 says is not how it is done. Nothing calls either. **Delete it**, or decide. |
|
||||
| **`display_ops`** | `acquire` and `wait_vblank` are in the vtable, implemented by one driver at most, with no consumer: Xorg draws through `/dev/fb0`. Kept for the GPU work ([gpu.md](gpu.md)); that is the reason it stays. |
|
||||
| **Socket activation** | `schinitd` parses `[Socket name]` and remembers the path. Nothing holds the socket. This is the entire reason for copying launchd, and it is the half that was not built. |
|
||||
|
|
@ -141,7 +129,7 @@ Neither cancelled nor in progress. **Do not open these.**
|
|||
|
||||
| | |
|
||||
|---|---|
|
||||
| `projects/schinst/` | the installer, the installation environment, `.sip` (see the open decision above) |
|
||||
| `projects/schinst/` | the installer, the installation environment, `.sip` — drawn by Schcom |
|
||||
| `projects/schfs-linux/` | SchFS as a filesystem for Linux |
|
||||
| [gpu.md](gpu.md) | amdgpu and nouveau through LDLS — measured, with the work it needs written down |
|
||||
| [finch.md](finch.md) | Finch inside SchOS (the language itself is finished and lives in its own repository) |
|
||||
|
|
@ -164,7 +152,7 @@ In order. **Do not start a gate before the one before it closes**, and closing m
|
|||
| **1 — The foundations** | descriptor passing, a shareable memory object (`memfd`), `AF_UNIX` in the filesystem, dynamic linking, the event descriptors | **closed** — [kernel.md](../architecture/kernel.md), [networking.md](../architecture/networking.md) |
|
||||
| **2 — Init as a role** | `schinit`, the supervisor, the heartbeat, PID 1 re-exec'd in place | **closed** — [init.md](../architecture/init.md) |
|
||||
| **3 — The desktop** | Xorg on the framebuffer, Window Maker, a terminal | **closed** — [desktop.md](../architecture/desktop.md) |
|
||||
| **4 — A root that survives a reboot** | the open decision above | **open** |
|
||||
| **4 — A root that survives a reboot** | dropped from 1.0 by decision; it is the installer, after 1.0 | **not in 1.0** |
|
||||
|
||||
---
|
||||
|
||||
|
|
|
|||
|
|
@ -40,21 +40,21 @@ Plans leave this directory when they are done. These did, and what replaced them
|
|||
|
||||
And one that was **decided against**: `wayland.md`. Wayland needs DRM/KMS, libinput, udev and libseat, and DRM/KMS
|
||||
alone is a kernel graphics subsystem. The desktop is X. The reasoning moved to [desktop.md](../architecture/desktop.md).
|
||||
`SchWM` and `SchDE`, which were to be the window manager and the desktop on top of it, are gone from the tree: Window
|
||||
Maker is the window manager, and a panel or launcher is an ordinary client.
|
||||
`SchWM` and `SchDE` were removed from the tree as READMEs with no code: Schcom is its own window manager, and SchDE returns as
|
||||
ordinary clients of Schcom when there is a Schcom to build them on. Meanwhile the window manager is Window Maker.
|
||||
|
||||
---
|
||||
|
||||
## The order
|
||||
|
||||
Set by [1.0.md](1.0.md), which is the authority. In short:
|
||||
Set by [1.0.md](1.0.md), which is the authority. **1.0 is reached** (a live system; persistence is not part of it).
|
||||
What is left before anything new starts:
|
||||
|
||||
1. **Close the one open gate**: decide what makes a root survive a reboot — the installer, or a narrower
|
||||
definition of 1.0 (see *The one open decision* there).
|
||||
2. **Empty the register.** Delete `projects/schcom/` and the unused in-kernel LDLS window, or say what they are for; build
|
||||
socket activation or remove the key.
|
||||
3. **Then**, and only then, whatever [roadmap.md](roadmap.md) says is next. GPU acceleration is the thing the laptop
|
||||
wants most ([gpu.md](gpu.md)), and Finch inside SchOS is the thing the author wants next ([finch.md](finch.md)).
|
||||
1. **Empty the register**: decide the unused in-kernel LDLS window (delete it, or use it), build socket activation
|
||||
or remove the key, and give `projects/schcom/` its first client.
|
||||
2. **Then**, whatever [roadmap.md](roadmap.md) says is next. The two things the owner wants most are the installer
|
||||
drawn by Schcom — SchOS's own compositor, and the road to SchDE ([roadmap.md](roadmap.md) §5) — and GPU
|
||||
acceleration on the laptop ([gpu.md](gpu.md)). Finch inside SchOS ([finch.md](finch.md)) comes after.
|
||||
|
||||
## What this plan does not address
|
||||
|
||||
|
|
@ -62,4 +62,4 @@ Stated here so it is not mistaken for covered:
|
|||
|
||||
- **No IOMMU**, so a driver in a process is isolated against its own bugs and not against a hostile driver.
|
||||
- **No page cache**, so a general writable `MAP_SHARED` of a file is refused; `memfd` is the narrow answer.
|
||||
- **No installer**, so nothing a live system writes survives a reboot.
|
||||
- **No installer**, so nothing a live system writes survives a reboot — by decision, for 1.0.
|
||||
|
|
|
|||
|
|
@ -54,7 +54,24 @@ what is present ([boot.md](../architecture/boot.md)).
|
|||
- **udev's job** — a uevent socket and `/sys/devices` — so X and others are told about devices instead of being
|
||||
given a list.
|
||||
|
||||
## 5. The installer and `.sip`
|
||||
## 5. Schcom, the installer, and SchDE
|
||||
|
||||
**Schcom** (`projects/schcom/`) is SchOS's own compositor. v0 is 715 lines that composite clients over a wire format of its
|
||||
own, and everything it was blocked on in the kernel now exists: Unix sockets, descriptor passing, `memfd`, evdev, framebuffer
|
||||
`mmap`. The plan is in three steps, each useful on its own:
|
||||
|
||||
1. **The installer's GUI is Schcom's first real client.** An installer wants a window, a list, a progress bar and a handful of
|
||||
input events — small enough to be the program that forces Schcom's protocol to be real, and independent of Xorg, which is not in
|
||||
the installation environment. It needs a wire format for clients that is designed (Wayland-shaped: surfaces, buffers shared by
|
||||
descriptor, input events), not the v0 one.
|
||||
2. **If that works, Schcom becomes useful in its own right**: window management as a module of the compositor (a compositor *is*
|
||||
its window manager), multiple clients with focus and stacking, damage tracking.
|
||||
3. **SchDE on top of it**: a panel, a launcher, a wallpaper and settings, as ordinary clients — only once there is a Schcom worth
|
||||
putting them on. Until then the desktop is Xorg with Window Maker ([desktop.md](../architecture/desktop.md)).
|
||||
|
||||
Nothing here starts before [1.0.md](1.0.md)'s register is empty.
|
||||
|
||||
### The installer and `.sip`
|
||||
|
||||
Scoped in [`projects/schinst/README.md`](../../projects/schinst/README.md), which holds the detail. In short:
|
||||
|
||||
|
|
@ -79,7 +96,7 @@ Open questions, all wanting a decision before code:
|
|||
now, or not?
|
||||
- **Can the kernel mount a `.sip` read-only directly?** If it can, the installation environment's root can simply *be* the `.sip`.
|
||||
|
||||
The GUI for this is an X client, like everything else on the desktop.
|
||||
The GUI for this is a Schcom client (above).
|
||||
|
||||
## 6. SchFS beyond SchOS
|
||||
|
||||
|
|
|
|||
|
|
@ -27,14 +27,14 @@ Concretely, code in here:
|
|||
|
||||
| Directory | What it is |
|
||||
|------------|-----------------------------------------------------------------|
|
||||
| `schcom/` | SchOS Compositor v0. Superseded by Xorg; awaiting deletion (see Status). |
|
||||
| `schcom/` | SchOS Compositor. v0 runs; the plan is for it to draw the installer, and later a desktop (SchDE) on top. |
|
||||
| `schinst/` | The installer, the installation environment, and `.sip` tooling. |
|
||||
| `schfs-linux/` | SchFS as a filesystem for Linux. FUSE first, a kernel module later. |
|
||||
|
||||
`schwm/` and `schde/` used to be here and have been **deleted**: both held a README and no code, and both
|
||||
had said "to be deleted" for long enough to prove that a directory marked for deletion is not deleted by being
|
||||
marked. The desktop is Xorg with Window Maker ([`docs/architecture/desktop.md`](../docs/architecture/desktop.md)),
|
||||
so the window manager is someone else's and a panel or launcher is an ordinary client.
|
||||
`schwm/` and `schde/` used to be here and were **deleted**: both held a README and no code. The window manager
|
||||
is Schcom's own business now (a compositor is its window manager), and **SchDE**, a panel, launcher and the rest,
|
||||
comes back as ordinary clients of Schcom once there is something for them to run on. Until then the desktop is
|
||||
Xorg with Window Maker ([`docs/architecture/desktop.md`](../docs/architecture/desktop.md)).
|
||||
|
||||
**`schfs-linux/` is written to stand on its own.** It does not mention SchOS
|
||||
anywhere, and that is deliberate rather than an oversight: it is intended to be
|
||||
|
|
@ -51,9 +51,11 @@ by anything above it.
|
|||
|
||||
## Status
|
||||
|
||||
`schcom/` is a compositor v0, 715 lines that run, over a wire format of its own. **It is superseded**: the desktop is
|
||||
Xorg on the framebuffer, not a compositor of ours, and nothing speaks Schcom's protocol. It is on the register in
|
||||
[`docs/plan/1.0.md`](../docs/plan/1.0.md) as *built and not connected*: delete it, or say what it is for.
|
||||
`schcom/` is a compositor v0, 715 lines that run, over a wire format of its own. It is **kept, with a job**: the
|
||||
installer's GUI ([`docs/plan/roadmap.md`](../docs/plan/roadmap.md) §5) is its first real client, and if it grows
|
||||
into something actually useful it becomes SchOS's own desktop, SchDE on top. Everything it was blocked on in the kernel
|
||||
exists (Unix sockets, descriptor passing, `memfd`, evdev, framebuffer `mmap`); what it lacks is a protocol designed for
|
||||
real clients and the clients themselves.
|
||||
|
||||
`schinst/` and `schfs-linux/` are scoped and unstarted, and are **frozen** -- see
|
||||
[`docs/plan/1.0.md`](../docs/plan/1.0.md).
|
||||
|
|
|
|||
|
|
@ -98,11 +98,13 @@ grep -v '^export \(PATH\|HOME\|TERM\)=' "$P/etc/profile" > "$P/etc/profile.d/sch
|
|||
rm -f "$P/etc/profile"
|
||||
|
||||
# How startx starts the X server for root: Alpine's xserverrc, plus
|
||||
# -dumbSched. Xorg reads input on a thread of its own, SchOS has no threads
|
||||
# yet, and the server crashed on its way out the first time a client
|
||||
# disconnected, in pthread_join on a thread that was never made. -dumbSched is
|
||||
# Xorg's own switch for input on the main loop. The file is root's own rather
|
||||
# than /etc/X11/xinit/xserverrc, which belongs to the xinit package.
|
||||
# -dumbSched. Xorg reads input on a thread of its own; when SchOS had no threads
|
||||
# the server crashed on its way out the first time a client disconnected, in
|
||||
# pthread_join on a thread that was never made. -dumbSched is Xorg's own switch
|
||||
# for input on the main loop. SchOS has threads now (clone, futex), so this may
|
||||
# no longer be needed -- it has simply not been tried without it. The file is
|
||||
# root's own rather than /etc/X11/xinit/xserverrc, which belongs to the xinit
|
||||
# package.
|
||||
mkdir -p "$P/root"
|
||||
cat > "$P/root/.xserverrc" <<'XEOF'
|
||||
#!/bin/sh
|
||||
|
|
|
|||
Loading…
Reference in a new issue