From c0ed23cc563d1ba7b7df79e5e53a6bcb5a69655c Mon Sep 17 00:00:00 2001 From: Nexoniarz Date: Sun, 13 Sep 2026 20:28:05 +0200 Subject: [PATCH] freeze: one register for everything half-built, and the corners cleared The failure this is against is not scope creep. It is abandonment by promotion: a thing gets built to the point where it works, the next thing becomes more interesting, and the first is left in a corner -- in the tree, not finished, not deleted, and not in any list. Enough of those and the tree stops matching its own documentation. docs/plan/1.0.md now holds the register, and it is the only place anything half-built is listed. Four sections, because the kinds are genuinely different: built and connected to nothing (the text layer, a font from disk, socket activation, wait_vblank); started and stopped part-way (foundations 2-5, hybrid H2, Schcom v0, a released PCI device that nothing brings back); named in a "what it does not do" and still true; and frozen -- scoped, unstarted, not to be started. The rule is that nothing enters except by being finished or deleted, and that adding a row requires closing one. 1.0.md already said "two things in flight, never three". It was right and it was not being followed. Then the corners themselves, because a register written over a tree full of them is just a longer list: - cmd/init is deleted. Nothing had execed /bin/init since initcode.S changed to /bin/schinit; it was built and installed on every image as an unreachable fallback, and an unreachable fallback rots rather than saves. - projects/schwm/ and projects/schde/ are deleted. Both said "to be deleted" in their own first paragraph, which proves only that a directory marked for deletion is not deleted by being marked. README-only, no code lost, and the arguments each made are moved into wayland.md where they are still true. - projects/schcom/schcom and schcom-demo were tracked binaries. Untracked, and .gitignore'd. Three code comments and one document described a boot that has not happened since 7d44eb7. cons_serial_only starts set, but kmain clears it on its first lines now, so the splash does not survive the boot and the kernel log is on screen from the start. That reversal was earned by a Chromebox with no serial port freezing at boot showing nothing at all, and the comments still argued the losing side of it. The site: - a download page. One ISO, what it has actually been run on -- one machine, under test -- Ventoy and dd, what the first boot looks like step by step, and the build if you would rather. The support table names one Chromebox rather than listing hardware, because one Chromebox is the entire bare-metal record and a list would be fiction. - Assets/usb.svg, drawn to the same rule as the other marks: two tones, geometry only, the accent spent in one place, and that place is the activity light because it is the part that tells you something is happening. - Download is in both navs and is the hero's primary action; a closing band ends the front page on it. - the uberblock-slots and drivers-outside-the-kernel figures are gone. Three numbers left, including the fastfetch one, which stays because it is funny. Two things caught by rendering the pages instead of assuming: .hero p was beating .filemeta on specificity and the file line was coming out at body size, and the dd command was one line long enough that the card clipped it at `conv=` -- a destructive command that can scroll half of itself out of view is a good way to take somebody's disk. It is two lines now. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01VQ7Qx8E1q1M1bbCq8QLHS2 --- .gitignore | 8 ++ Makefile | 2 +- cmd/init/main.c | 134 ------------------- docs/README.md | 9 +- docs/Website/Assets/usb.svg | 27 ++++ docs/Website/README.md | 27 +++- docs/Website/docs.html | 1 + docs/Website/download.html | 226 ++++++++++++++++++++++++++++++++ docs/Website/index.html | 25 +++- docs/Website/style.css | 103 ++++++++++++++- docs/architecture/boot.md | 30 +++-- docs/plan/1.0.md | 77 +++++++++++ docs/plan/README.md | 7 +- docs/plan/schapi.md | 2 +- docs/plan/wayland.md | 21 ++- kernel/core/console.c | 15 ++- kernel/drivers/gpu/vga/vga_fb.c | 10 +- lib/initcode.S | 7 +- projects/README.md | 34 +++-- projects/schcom/README.md | 8 ++ projects/schcom/schcom | Bin 45416 -> 0 bytes projects/schcom/schcom-demo | Bin 40928 -> 0 bytes projects/schde/README.md | 34 ----- projects/schwm/README.md | 34 ----- 24 files changed, 585 insertions(+), 256 deletions(-) delete mode 100644 cmd/init/main.c create mode 100644 docs/Website/Assets/usb.svg create mode 100644 docs/Website/download.html delete mode 100644 projects/schcom/schcom delete mode 100644 projects/schcom/schcom-demo delete mode 100644 projects/schde/README.md delete mode 100644 projects/schwm/README.md diff --git a/.gitignore b/.gitignore index 199ca10..b3c0522 100644 --- a/.gitignore +++ b/.gitignore @@ -33,3 +33,11 @@ dist/ .#* *~ \#*\# + +# Build output under projects/ -- these are programs that run on SchOS, and +# their binaries were tracked by accident. +projects/schcom/schcom +projects/schcom/schcom-demo + +# A built ISO dropped next to the site for download.html. 62 MB; not source. +docs/Website/releases/ diff --git a/Makefile b/Makefile index e7be12c..65fd31c 100644 --- a/Makefile +++ b/Makefile @@ -107,7 +107,7 @@ KERN_NOENTRY = $(filter-out $(BUILD)/kernel/arch/x86_64/entry.o, $(KERNOBJS)) # Third-party software does NOT go here — see ports/. # ----------------------------------------------------------------------- ULIB = $(addprefix $(BUILD)/lib/, start.o glue.o usys.o) -CMDS = init schinit schinitd fbinfo shmtest socktest inputtest nettest reflink schattr ping chvt pcitest vgpud +CMDS = schinit schinitd fbinfo shmtest socktest inputtest nettest reflink schattr ping chvt pcitest vgpud CMDBINS = $(addprefix $(BUILD)/cmd/, $(CMDS)) # ----------------------------------------------------------------------- diff --git a/cmd/init/main.c b/cmd/init/main.c deleted file mode 100644 index a5f68e7..0000000 --- a/cmd/init/main.c +++ /dev/null @@ -1,134 +0,0 @@ -// SPDX-License-Identifier: Apache-2.0 -// SchOS — minimal init (PID 1) - -#include "types.h" -#include "user.h" -#include "fcntl.h" - -static void -xputs(int fd, const char *s) -{ - int n = 0; - while (s[n]) n++; - write(fd, (void *)s, n); -} - -int -main(void) -{ - if (getpid() != 1) - exit(); - - // Open console as fd 0; dup to fd 1 (stdout) and fd 2 (stderr). - if (open("console", O_RDWR) < 0) { - mknod("console", 1, 1); - open("console", O_RDWR); - } - dup(0); - dup(0); - - // --- /dev ------------------------------------------------------------ - // Minor numbers under major 1 follow Linux's mem-major layout where it - // costs nothing (null=2, zero=5, full=7, random=8, urandom=9). - // /dev/hole (was minor 4) is gone: it backed an abandoned experiment and - // nothing has referenced it since. - mkdir("/dev"); - mknod("/dev/console", 1, 1); - mknod("/dev/null", 1, 2); - mknod("/dev/keyboard", 1, 3); - mknod("/dev/zero", 1, 5); - mknod("/dev/full", 1, 7); - mknod("/dev/random", 1, 8); - mknod("/dev/urandom", 1, 9); - mknod("/dev/fb0", 2, 0); - mknod("/dev/mouse", 3, 1); - mknod("/dev/tty", 1, 1); // alias for console — must pass TCGETS ioctl - mknod("/dev/ttyS0", 4, 64); - mknod("/dev/stdin", 1, 1); - mknod("/dev/stdout", 1, 1); - mknod("/dev/stderr", 1, 1); - - // Raw input events, Linux evdev layout (kernel/dev/input.c). /dev/keyboard - // and /dev/mouse above are the old character-and-position interfaces; these - // are the ones with key releases and button transitions, which is what a - // game or a compositor needs and what neither of the others can express. - mkdir("/dev/input"); - mknod("/dev/input/event0", 7, 0); // keyboard - mknod("/dev/input/event1", 7, 1); // mouse - - // Raw disks (kernel/dev/disk.c). The minor is the IDE device number, so - // hda is the disk the kernel booted from and hdb the second channel. - // - // Created unconditionally, the same way every other node here is: opening - // one when no drive is there fails with ENXIO, which is a better answer - // than a missing path, and it keeps this list a description of the device - // model rather than of one machine's hardware. - mknod("/dev/hda", 8, 0); - mknod("/dev/hdb", 8, 1); - - // --- /proc ----------------------------------------------------------- - // Synthesised status files, major 6 (kernel/dev/procfs.c). These are the - // global ones; per-process /proc// directories need synthetic - // directory support in the filesystem, which does not exist yet. - mkdir("/proc"); - // /proc is a real filesystem now (kernel/fs/procfs.c), not a handful of - // mknod'd nodes: it has to be, because one directory per process cannot be - // created ahead of time. This is what makes `ps` and `top` work. - // - // The mknod'd global files below still happen, and are simply covered by - // the mount. They stay for now so that a failure here degrades to the old - // behaviour instead of losing /proc entirely. - mount("proc", "/proc", "proc", 0, 0); - mknod("/proc/meminfo", 6, 1); - mknod("/proc/uptime", 6, 2); - mknod("/proc/cpuinfo", 6, 3); - mknod("/proc/version", 6, 4); - mknod("/proc/stat", 6, 5); - mknod("/proc/loadavg", 6, 6); - mknod("/proc/mounts", 6, 7); - mknod("/proc/filesystems", 6, 8); - // SchFS volume summary: label, generation, free space, and the root - // directory. Reachable whether or not the volume is mounted into the - // namespace, which is what makes it useful when a mount has gone wrong. - mknod("/proc/schfs", 6, 9); - // Write 'c' here to crash on purpose. The crash screen is otherwise the - // one part of the system that can only be seen by having a bug. - mknod("/proc/sysrq", 6, 10); - mknod("/proc/pci", 6, 11); - mknod("/proc/interrupts", 6, 12); - - mkdir("/sys"); - - xputs(1, "\nSchOS\n"); - xputs(1, "init: ready (pid 1)\n"); - - for (;;) { - int pid = fork(); - if (pid < 0) { - xputs(2, "init: fork failed\n"); - sleep(100); - continue; - } - - if (pid == 0) { - // exec.c auto-enables Linux ABI for musl binaries (entry >= 0x100000). - // Do NOT call enable_linux_abi() here — it corrupts SchOS syscall - // numbers for the exec/exit calls that follow in this child. - // Standardized naming: the shell is always /bin/sh, regardless of - // which implementation backs it (dash, mksh, etc.) — no ksh/mksh/ - // other aliases are looked up here. - static char *argv[] = { "sh", 0 }; - exec("/bin/sh", argv); - xputs(2, "init: no shell found; place sh in /bin\n"); - exit(); - } - - // Reap zombies; the loop restarts the shell when it exits. - int wpid; - while ((wpid = wait()) >= 0 && wpid != pid) - ; - // Brief pause before restarting so we don't spin on a missing shell. - sleep(100); - } - return 0; -} diff --git a/docs/README.md b/docs/README.md index 8e122b9..c80ca3e 100644 --- a/docs/README.md +++ b/docs/README.md @@ -55,7 +55,12 @@ files are. **Trying to build it** → [guides/building.md](guides/building.md). -**Wondering what happens next** → [plan/1.0.md](plan/1.0.md), which is the finish line, then [plan/README.md](plan/README.md) for the order. +**Wondering what happens next** → [plan/1.0.md](plan/1.0.md), which is the finish +line *and* the register of everything half-built. Then +[plan/README.md](plan/README.md) for the order. + +**About to start something new** → the register, first. Nothing new starts while +it has rows in it, and a row leaves only by being finished or deleted. **Something is broken and it feels familiar** → [history/](history/). There is a good chance it is in there. @@ -84,7 +89,7 @@ good chance it is in there. | | | |---|---| | [README](plan/README.md) | The three current decisions and the order they happen in. | -| [1.0.md](plan/1.0.md) | **The finish line.** What 1.0 contains, and what is explicitly deferred past it. | +| [1.0.md](plan/1.0.md) | **The finish line, and the freeze.** What 1.0 contains, what is deferred, and the register of every half-built thing in the tree. | | [foundations.md](plan/foundations.md) | The kernel work everything else is blocked on. | | [wayland.md](plan/wayland.md) | The GUI stack, and what gets deleted to make room. | | [hybrid.md](plan/hybrid.md) | Moving drivers out of the kernel. | diff --git a/docs/Website/Assets/usb.svg b/docs/Website/Assets/usb.svg new file mode 100644 index 0000000..1b35037 --- /dev/null +++ b/docs/Website/Assets/usb.svg @@ -0,0 +1,27 @@ + + A USB stick + The only medium SchOS is meant to arrive on. There is no installer and + no optical drive in the plan -- the stick is the machine's second disk, and on + a Chromebox it is the only one that matters. + + Drawn to the same rule as the other marks here: two tones, geometry only, and + the accent spent in exactly one place. That place is the activity light, + because it is the part of a USB stick that tells you something is happening -- + the same job the accent does everywhere else in this palette. + + The connector's two contact windows are real holes (fill-rule evenodd) rather + than shapes filled with the background colour, so the mark survives being put + on a surface that is not the page background. + + + + + + + + + + + diff --git a/docs/Website/README.md b/docs/Website/README.md index 9c83e51..cd286b3 100644 --- a/docs/Website/README.md +++ b/docs/Website/README.md @@ -1,19 +1,42 @@ # The website -Five files and a folder of marks. No build step, no dependencies, no +Six files and a folder of marks. No build step, no dependencies, no framework — open `index.html` and it works, including from a `file://` URL. ``` index.html the front page +download.html get it, write it to a stick, or build it docs.html the documentation reader style.css the palette, transcribed from include/theme.h app.js the front page: the boot log types, sections fade in docs.js the Markdown renderer and the doc index Assets/ shift.svg (logo), uberblock.svg (SchFS), guru.svg (PSOGM), - pride.svg + usb.svg (downloads), pride.svg ``` +## 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: + +```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 `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 +scroll half of itself out of view is a good way to take somebody's disk. + ## The docs reader `docs.html` renders the Markdown in `docs/` in the site's own theme, with the diff --git a/docs/Website/docs.html b/docs/Website/docs.html index 676be54..b97210a 100644 --- a/docs/Website/docs.html +++ b/docs/Website/docs.html @@ -21,6 +21,7 @@ diff --git a/docs/Website/download.html b/docs/Website/download.html new file mode 100644 index 0000000..aedd9c6 --- /dev/null +++ b/docs/Website/download.html @@ -0,0 +1,226 @@ + + + + + +Download — SchOS + + + + + + +
+ +
+ +
+ +
+
+
+

Download

+

Put it on a stick

+

+ SchOS ships as one bootable ISO. It boots on UEFI, runs entirely from + RAM, and touches nothing on the machine unless you tell it to — there is + no installer yet, and that is the point of trying it this way. +

+ +

+ schos.iso · about 62 MB · UEFI x86-64 · nothing is written + to your disks +

+
+ +
+

Before you do

+

This has booted on exactly one real machine

+

+ Everything below QEMU is honest reporting rather than a compatibility + list. SchOS drives real hardware, and it has been run on a very small + number of it. +

+ + + + +
QEMU / OVMFboots
Asus Chromebox Guado CN62under test
Anything elseunknown
+

+ If it does not boot, the kernel log is on the screen from its first line + — that is deliberate, and it is there because of this machine. A photo of + where it stops is a genuinely useful bug report. +

+
+
+
+ +
+
+

Writing it

+

Two ways onto a stick

+

+ Ventoy first, because it is the one that does not cost you the stick. An + ISO copied onto a Ventoy drive sits next to every other ISO on it, and the + drive stays an ordinary drive. +

+ +
+
+ recommended +

Ventoy

+

+ Copy the file. That is the whole procedure — Ventoy finds it and offers + it in its menu at boot, and the stick remains usable for everything + else. +

+
$ cp schos.iso /run/media/you/Ventoy/
+

+ This is how SchOS is actually tested on hardware. No optical drive has + been involved at any point. +

+
+ +
+ destructive +

dd

+

+ Writes the image over the whole device. Check the device name + twice — there is no confirmation and no undo, and the wrong letter + takes a disk with it. +

+
$ lsblk
+$ sudo dd if=schos.iso of=/dev/sdX \
+       bs=4M status=progress conv=fsync
+

+ of= is the whole device, not a partition — sdX + and never sdX1. +

+
+
+
+
+ +
+
+

First boot

+

What actually happens

+

+ Written out because a system you have never seen before gives you no way to + tell "working" from "stuck", and the difference is worth knowing in advance. +

+ +
+
+ 1 +

The loader

+

+ A splash, and F8 for boot options. It is a PE/COFF loader built + with clang and lld-link; there is no BIOS path, so the firmware has to + be in UEFI mode. +

+
+
+ 2 +

The kernel log

+

+ On screen, from the first line that can draw. Cores, the PCI bus, the + network card, the filesystem. If it stops, it stops somewhere you can + read. +

+
+
+ 3 +

A shell

+

+ schinit starts one on the console and a second on + tty2. mksh is /bin/sh, with + zsh beside it and the real toybox userland behind both. +

+
+
+ 4 +

Six terminals

+

+ AltGr+F1…F6 switches between them — the one + chord a Linux host does not take for itself — or chvt N + from a shell. +

+
+
+ +

+ Nothing is written to your disks and nothing persists. The root + filesystem is in RAM; the machine is exactly as you left it after a reboot, + including the parts of it you were hoping to keep. There is no installer — + 1.0 is where that starts. +

+
+
+ +
+
+

Or build it

+

Clone it, build it, boot it

+

+ The toolchain is one file. On NixOS shell.nix brings the whole + thing in; elsewhere it is clang, lld, GNU make and a host compiler — see + the build guide. +

+ +
$ git clone <the repository> schos && cd schos
+$ nix-shell            # NixOS only; elsewhere see the build guide
+$ make iso             # produces schos.iso
+$ make qemu            # or just run it, without a stick
+ +

+ make qemu is the one to reach for first. Everything on this + page works there, it takes seconds rather than a reboot, and a serial + console beats a photograph of a screen. +

+
+
+ +
+ +
+
+ SchOS — kernel SchIFT, filesystem SchFS. + Apache-2.0, with the xv6-derived core under MIT. + + + Built by, and for, people of every orientation and gender identity. + The flag stays up all year. + + + On AI. + Parts of SchOS were written with AI assistance, and that is not the measure + of anything. A project is good when the people building it understand what + was built, and bad when nobody does — whoever, or whatever, typed it. Vibe + your way through and you get something no one can fix. Stay in the loop and + you get a system you can explain line by line. The bug writeups in + history/ are where that gets + checked, including the wrong turns. + +
+
+ + + + diff --git a/docs/Website/index.html b/docs/Website/index.html index 6c80330..7b8e43e 100644 --- a/docs/Website/index.html +++ b/docs/Website/index.html @@ -25,6 +25,7 @@ What it has Hybrid Numbers + Download Docs @@ -42,8 +43,8 @@ and runs unmodified musl and toybox on top.

@@ -350,9 +351,7 @@
46,961lines in kernel/ and include/
207syscalls, Linux numbering
-
128uberblock ring slots
0patches carried against fastfetch
-
1driver already running outside the kernel
@@ -407,6 +406,24 @@ +
+
+ +
+

Try it

+

One ISO, one stick, nothing written to your disks

+

+ It runs from RAM and leaves the machine as it found it. Or skip the + stick entirely — make qemu is the same system in a window. +

+ +
+
+
+