rajamesandClaude Opus 5.5
c95191736e
Each board has one serial line that is its console, the lowest common denominator beside the screen and keyboard. Its address was QEMU's, fixed: on a Raspberry Pi 5 the aarch64 one is ordinary RAM, written to with every character printed; on a Milk-V Mars the riscv64 one is the board's UART with registers four bytes apart, and the UEFI loader would have waited for ever on its first character; on a PC with no COM1 the prompt read 0xFF for ever. console_attach_serial(), called before anything prints: COM1 if a UART answers there; elsewhere the UART that /chosen/stdout-path names (fdt_uart.c: alias, path, the address carried up through each bus's ranges, reg-shift and reg-io-width); with no devicetree, QEMU's. A devicetree that names no usable console gives no line, not a guess. A line that does not come ready is given up, not waited on. The riscv64 loader's early serial does the same. tools/fdt_uart_test.c, run by hand, 19 checks plain and under the sanitizers: devicetrees shaped like the Mars's, the Pi 5's and QEMU's, what must find nothing, and a block cut short at every length; and the Pi 5's real devicetree gives a PL011 at 0x107d001000. Acceptance, one QEMU at a time: v3 on amd64 (logs/20261013-204436), aarch64 (-204803) and riscv64 (-205001) to the zuse prompt, no unknown word, hash 0x08873e0f44b7cb2a on all three as before. -204600 is an aarch64 run during which the disk images were restored under it by a script that had been refused a second QEMU; it passed and is kept, and was run again. v4 on the three ISAs (-211149, -211444, -211857): POST 538 of 538, dict_hash 0xe5a57f3eb72ff94a, the typed session unchanged. The Mars image through OpenSBI and U-Boot in QEMU took its line and its interrupt controller from the devicetree and passed POST. Found on the way and not mended (boards/BRINGUP.md 1.6): built with EMERGENCY_CONSOLE_ENABLED=1 the prompt floods with empty lines. -210442 shows it; the commit before this does the same. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
251 lines
9.7 KiB
C
251 lines
9.7 KiB
C
/*
|
||
StarForth — Steady-State Virtual Machine Runtime
|
||
|
||
Copyright (c) 2023–2025 Robert A. James
|
||
All rights reserved.
|
||
|
||
This file is part of the StarForth project.
|
||
|
||
Licensed under the StarForth License, Version 1.0 (the "License");
|
||
you may not use this file except in compliance with the License.
|
||
|
||
You may obtain a copy of the License at:
|
||
https://github.com/star.4th@proton.me/StarForth/LICENSE.txt
|
||
|
||
This software is provided "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
||
express or implied, including but not limited to the warranties of
|
||
merchantability, fitness for a particular purpose, and noninfringement.
|
||
|
||
See the License for the specific language governing permissions and
|
||
limitations under the License.
|
||
|
||
StarForth — Steady-State Virtual Machine Runtime
|
||
Copyright (c) 2023–2025 Robert A. James
|
||
All rights reserved.
|
||
|
||
This file is part of the StarForth project.
|
||
|
||
Licensed under the StarForth License, Version 1.0 (the "License");
|
||
you may not use this file except in compliance with the License.
|
||
|
||
You may obtain a copy of the License at:
|
||
https://github.com/star.4th@proton.me/StarForth/LICENSE.txt
|
||
|
||
This software is provided "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
||
express or implied, including but not limited to the warranties of
|
||
merchantability, fitness for a particular purpose, and noninfringement.
|
||
|
||
See the License for the specific language governing permissions and
|
||
limitations under the License.
|
||
|
||
*/
|
||
|
||
/**
|
||
* console.h - Serial console + framebuffer VT100 interface for StarKernel
|
||
*
|
||
* Output policy:
|
||
* - Serial UART is always active (initialized by console_init).
|
||
* - When console_fb_init() has been called and the framebuffer is ready,
|
||
* every character is also rendered through the VT100 terminal on screen.
|
||
* - Both outputs are always live simultaneously; serial cannot be disabled.
|
||
*/
|
||
|
||
#ifndef STARKERNEL_CONSOLE_H
|
||
#define STARKERNEL_CONSOLE_H
|
||
|
||
#include <stdint.h>
|
||
#include <stddef.h>
|
||
#include "uefi.h"
|
||
|
||
/**
|
||
* Find the board's serial console line. Must be called before anything is
|
||
* printed: until it is, and on a board with no such line, the console writes
|
||
* to no serial port and reads from none. On a PC it is COM1, if a UART
|
||
* answers there; elsewhere it is the UART the devicetree names as the
|
||
* console (/chosen/stdout-path); with no devicetree at all it is the UART of
|
||
* QEMU's virt machine. @p dtb may be NULL.
|
||
*/
|
||
void console_attach_serial(const void *dtb);
|
||
|
||
/**
|
||
* Initialize serial console (UART).
|
||
* Must be called once during early kernel boot, after console_attach_serial().
|
||
*/
|
||
void console_init(void);
|
||
|
||
/** Print one line saying which serial line the console found, or that there is none. */
|
||
void console_describe_serial(void);
|
||
|
||
/**
|
||
* Initialize the framebuffer VT100 terminal.
|
||
* Call after UEFI boot services have been exited and the GOP framebuffer
|
||
* address is known (from BootInfo). Safe to call with info==NULL (no-op).
|
||
* fmt: FB_PIXEL_BGRX32 is correct for most QEMU / real hardware GOP.
|
||
*/
|
||
#include "framebuffer.h"
|
||
void console_fb_init(const FramebufferInfo *info, FbPixelFormat fmt);
|
||
|
||
/**
|
||
* FABRIC-0.md item 4.4j: switch the framebuffer console's glyph backend from
|
||
* font_8x16.c to TTF-TEXT's rasterizer. Thin wrapper over
|
||
* vt100_enable_ttf() -- see that function's doc comment for the full
|
||
* contract (lazy font load, cell-geometry/cols/rows recompute, screen
|
||
* clear, one-shot). No-op if the framebuffer console was never
|
||
* initialized (console_fb_init() not called, or it no-op'd on a NULL
|
||
* framebuffer).
|
||
*/
|
||
void console_fb_enable_ttf(void);
|
||
|
||
/**
|
||
* FABRIC-0.md item 4.4q: thin wrappers over vt100_scroll_back()/
|
||
* vt100_scroll_fwd() -- see those functions' doc comments for the full
|
||
* contract. No-op if the framebuffer console was never initialized.
|
||
*/
|
||
void console_fb_scroll_back(uint32_t n);
|
||
void console_fb_scroll_fwd(uint32_t n);
|
||
|
||
/**
|
||
* FABRIC-0.md item 4.4y-revised: thin wrapper over vt100_toggle_graphics()
|
||
* -- see that function's doc comment for the full contract (the
|
||
* Alt+TAB graphics/text state machine). No-op if the framebuffer console
|
||
* was never initialized.
|
||
*/
|
||
void console_fb_toggle_graphics(void);
|
||
|
||
/**
|
||
* Thin wrapper over vt100_draw_cursor() -- see that function's doc
|
||
* comment for the full contract (a static block cursor at the terminal's
|
||
* current position). No-op if the framebuffer console was never
|
||
* initialized.
|
||
*/
|
||
void console_fb_draw_cursor(void);
|
||
|
||
/**
|
||
* Thin wrapper over vt100_erase_cursor(). No-op if the framebuffer
|
||
* console was never initialized.
|
||
*/
|
||
void console_fb_erase_cursor(void);
|
||
|
||
/**
|
||
* Write a single character to serial console
|
||
*/
|
||
void console_putc(char c);
|
||
|
||
/**
|
||
* Write a null-terminated string to serial console
|
||
*/
|
||
void console_puts(const char *s);
|
||
|
||
/**
|
||
* Write a string with newline to serial console
|
||
*/
|
||
void console_println(const char *s);
|
||
|
||
/**
|
||
* console_ensure_line_start - Make sure the *next real output* starts at
|
||
* the beginning of a fresh output line, closing off whatever is currently
|
||
* mid-line (e.g. a dangling prompt). No-op if already at line start.
|
||
*
|
||
* FABRIC-2.md §I.9 fix, 2026-09-05: the newline is DEFERRED, not emitted
|
||
* immediately -- it only actually reaches the console on the next real
|
||
* console_putc() call, and is silently dropped (never emitted at all) if
|
||
* console_cancel_deferred_line_start() is called first instead. Before this
|
||
* fix, an immediate, unconditional newline here meant any caller invoking
|
||
* this function "just in case" (sk_repl_idle() being the one real caller)
|
||
* would visibly snap a bare, unfinished prompt line to a fresh blank line
|
||
* even when nothing was actually about to be printed -- indistinguishable
|
||
* from Enter having already been pressed at that prompt. Deferring means a
|
||
* caller that turns out to have nothing to print can cancel cleanly, with
|
||
* zero visible effect, while a caller that does print gets the correct
|
||
* "close the old line first" behavior for free, still counted by
|
||
* console_tx_count() as real output (unlike the old always-silent inner
|
||
* form) since it is realized through the normal console_putc() path.
|
||
*/
|
||
void console_ensure_line_start(void);
|
||
|
||
/**
|
||
* console_cancel_deferred_line_start - Discard a pending deferred newline
|
||
* from console_ensure_line_start() without ever emitting it. No-op if no
|
||
* newline is currently deferred. Callers that speculatively deferred a line
|
||
* break before checking whether they actually have anything to print
|
||
* (sk_repl_idle()'s idle-beat check being the motivating case, FABRIC-2.md
|
||
* §I.9) call this when the check comes back negative, so the console is
|
||
* left exactly as it was -- no stray newline, no phantom blank line.
|
||
*/
|
||
void console_cancel_deferred_line_start(void);
|
||
|
||
/**
|
||
* console_tx_count - Monotonic count of console_putc() calls delivered to
|
||
* either output (serial and/or framebuffer). In use by the REPL to detect
|
||
* that an idle bottom half (heartbeat, USB attach/detach) wrote to the
|
||
* console while the top-level prompt was showing, so it can re-anchor the
|
||
* prompt afterward. Never decreases.
|
||
*/
|
||
uint64_t console_tx_count(void);
|
||
|
||
/**
|
||
* Read a single character from serial console (non-blocking)
|
||
* Returns -1 if no character available
|
||
*/
|
||
int console_getc(void);
|
||
|
||
/**
|
||
* Check if character is available for reading
|
||
*/
|
||
int console_poll(void);
|
||
|
||
/**
|
||
* Set the active VM name shown as [Name] prefix on each output line.
|
||
* Pass NULL to suppress the prefix (kernel-only output before any VM).
|
||
* Copies into internal storage (FABRIC-2.md Phase F, 2026-08-28) -- the
|
||
* caller's own pointer does not need to remain valid afterward.
|
||
*/
|
||
void console_set_vm_name(const char *name);
|
||
const char *console_get_vm_name(void);
|
||
|
||
/**
|
||
* console_save_vm_name - Copy the current active-VM name into the
|
||
* caller's own buffer, for a later console_set_vm_name() restore.
|
||
*
|
||
* console_get_vm_name() alone is NOT safe for save-then-restore: it
|
||
* returns a pointer into the single internal buffer console_set_vm_name()
|
||
* copies into, so an intervening console_set_vm_name() call (the normal
|
||
* "switch, do work, switch back" pattern every BIRTH/VM-EXEC/CONNECT-*
|
||
* call site uses) overwrites the very bytes the saved pointer points at
|
||
* before the restore ever runs -- found live 2026-08-28, the restore
|
||
* silently no-ops. Copies at most cap-1 bytes plus a NUL terminator;
|
||
* writes "" if there was no active name (NULL) to save.
|
||
*
|
||
* @param out Caller-owned buffer.
|
||
* @param cap Its size in bytes.
|
||
*/
|
||
void console_save_vm_name(char *out, size_t cap);
|
||
|
||
/**
|
||
* FABRIC-3.md SXXV follow-up (2026-09-13): the line prefix used to be a
|
||
* bare "[VMName] ", ambiguous between a console-proxy VM and the actual
|
||
* identity VM behind it (e.g. "rajames" the console vs. "rajames~user"
|
||
* the real WIREBIND-restricted identity) -- confirmed live to cause real
|
||
* confusion during testing. Unified to "[user@VMName] " when a user
|
||
* context is available, still "[VMName] " when not (unauthenticated/
|
||
* pre-login, unchanged from today).
|
||
*
|
||
* console.c is a clean HAL module with no dependency on capsule/WIREBIND
|
||
* logic (zuse_session, capsule_wirebind_attached_username()) -- pulling
|
||
* either in directly here would be a real layering violation, not just a
|
||
* style preference. This callback lets repl.c (which already computes
|
||
* exactly this for sk_print_prompt()) supply the "user" half without
|
||
* console.c knowing anything about VMs, sessions, or WIREBIND. Returns
|
||
* NULL/empty for "no user context" (falls back to the bare "[VMName] "
|
||
* form); the returned pointer must remain valid until the next call
|
||
* (matches console_get_vm_name()'s own single-buffer convention).
|
||
*/
|
||
typedef const char *(*console_user_prefix_fn)(void);
|
||
void console_set_user_prefix_provider(console_user_prefix_fn fn);
|
||
|
||
/* Last FORTH word name set by the dispatcher before entry->func(vm).
|
||
* Printed by the #GP fault handler to identify the faulting word. */
|
||
extern volatile const char *g_sk_fault_word;
|
||
|
||
#endif /* STARKERNEL_CONSOLE_H */
|