Files
LithosAnanake/include/starkernel/vm/stadium_blocks.h
T
Robert Allan JamesandClaude Sonnet 5 c7c9332321 Stadium: real block-patron admission + MIGRATE dispatch (FABRIC-3.md §B)
stadium_admit()'s mass==1 refusal looked like a hard blocker for 1024-byte
blocks, but stadium_word_dispatch()'s real candidate construction proves
Stadium cells carry pure identity/heat/bookkeeping, never the resident's
actual content -- a block patron follows the same shape (identity=LBN,
payload unused), so this was real, scoped work, not a case for stubbing.

New stadium_blocks.h/.c mirror stadium_words.c's admission/cooling shape,
keyed by (quota_slot, lbn) in a fixed-capacity open-addressing hash table
(tombstone deletion) instead of a dense array, since LBN space isn't
densely bounded like word_id. Wired into block_word_block()/buffer()/
update() (block_words.c), __STARKERNEL__-guarded. stadium_dispatch()'s
MIGRATE case now calls blk_flush(lbn) for real instead of printing
"(stub)". Three new Kconfig constants (STADIUM_BLOCK_HEAT_QUANTUM/
STADIUM_BLOCK_COOL_RATE_Q48/STADIUM_BLOCK_TRACK_CAP_MULT) mirror the
word-patron ones, same three-layer wiring.

VM-COOL/DELIVER/EXPIRE stay explicit punch-list items -- VM-COOL
deferred pending the still-iterating Tripod/Zuse/messaging vision,
DELIVER/EXPIRE are their own future subsystem integrations per
FABRIC.md's own "open, not resolved" notes.

Verified clean compile (zero warnings) and clean boot to REPL with
conservation intact (resident_sum + reservoir == Q48_ONE) on all three
architectures (amd64/aarch64/riscv64); BLOCK/BUFFER touches exercised
live from the REPL with no crash; a 22,000-distinct-block flood loop
against an artificially shrunk Stadium ran clean under heavy admission
load. A live MIGRATE console fire was not directly observed this
session (root-caused to a pre-existing reservoir-floor/density-eviction
interaction unrelated to this change, documented in FABRIC-3.md) --
flagged as an honest follow-up, not silently claimed.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CXjAPTEKrgY2Mrk25KoLDn
2026-08-25 23:28:59 -04:00

125 lines
5.7 KiB
C
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
/*
StarForth — Steady-State Virtual Machine Runtime
Copyright (c) 20232025 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.
*/
/**
* stadium_blocks.h - Block patrons on the Stadium (FABRIC-3.md §B/§D,
* MIGRATE punch-list item)
*
* The block-specific layer on top of the generic L0 engine (stadium.h), same
* relationship stadium_words.h/.c already has: nothing in stadium.c/.h knows
* a block patron exists -- it only ever sees cell_index, VMUuid, and
* StadiumPatronHeader. This file is where "block" becomes a concrete
* meaning: an (owning quota slot, LBN) -> cell_index map, the starter-grant
* admission rule (mirrors stadium_word_dispatch()'s Option B exactly), and
* the reservoir-quantum touch/cool that feeds and drains a resident block's
* Stadium heat.
*
* Unlike words, LBN is not densely bounded (block_subsystem.c's unified LBN
* space spans RAM/RAMDRIVE/DISK/USB and can be large), so the map here is a
* fixed-capacity open-addressing hash table sized off stadium_cell_count()
* at init, not a dense per-VM array -- see stadium_blocks.c for the layout.
* A block's actual 1024 content bytes are never copied into a Stadium cell;
* they stay exactly where block_subsystem.c already keeps them. The cell
* only ever carries identity (the LBN) and heat/bookkeeping, same as a word
* patron's cell never carries the word's own dictionary entry.
*/
#ifndef STARKERNEL_VM_STADIUM_BLOCKS_H
#define STARKERNEL_VM_STADIUM_BLOCKS_H
#ifdef __STARKERNEL__
#include <stdint.h>
#include "starkernel/vm_uuid.h"
/*
* stadium_blocks_init - Allocates and zeroes the (quota slot, LBN) ->
* cell_index hash table (capacity computed from stadium_cell_count() *
* STADIUM_BLOCK_TRACK_CAP_MULT at call time, kmalloc'd). Must be called
* after stadium_boot_init() -- so stadium_cell_count() is non-zero -- and
* before any block ever dispatches; the real boot site is immediately after
* the existing stadium_words_init() call (kernel_main.c), same M7/M7.1
* ordering. Not safe to call twice -- guarded internally as a no-op if
* already initialized, same convention as stadium_words_init().
*/
void stadium_blocks_init(void);
/*
* stadium_block_dispatch - The per-touch entry point, called from
* block_word_block()/block_word_buffer()/block_word_update()
* (src/word_source/block_words.c) -- mirroring stadium_word_dispatch()'s
* call pattern and cooling/admission logic 1:1, keyed by LBN instead of
* word_id.
*
* If (vm_id, lbn) is already resident: applies the same redirected Loop #3
* cooling stadium_word_dispatch() applies (fraction of the cell's own
* current heat, scaled by elapsed_ticks since this block's own last touch --
* STADIUM_BLOCK_COOL_RATE_Q48), crediting the cooled amount back to vm_id's
* reservoir, then pulls STADIUM_BLOCK_HEAT_QUANTUM from the reservoir into
* the cell -- clamped to the reservoir's actual balance AND to the same
* Q48_ONE / 3 floor stadium_word_dispatch() enforces, so block-touch
* admission alone can never starve other reservoir consumers sharing the
* same VM.
*
* If not resident (or the table's entry is stale -- self-healing check
* against the cell's discriminator bit and identity, same pattern
* resolve_resident_cell() uses in stadium_words.c): attempts starter-grant
* admission -- pulls STADIUM_BLOCK_HEAT_QUANTUM (same floor), builds an
* unpinned MIGRATE candidate (identity = lbn, mass = 1, payload unused --
* the block's real content is never copied here), calls stadium_admit(). On
* refusal, pushes the pulled quantum back (rollback). On success, records
* the mapping. If the hash table itself is full and has no slot for this
* (vm_id, lbn) pair, this touch is silently skipped -- Stadium's own
* capacity already bounds real residency, so a table miss under load is
* graceful degradation, not an error.
*
* No-op if vm_id holds no Stadium quota, or stadium_blocks_init() has not
* run.
*
* @param vm_id Owning VM -- vm->stadium_vm_id at every call site,
* same quota-isolation reasoning stadium_word_
* dispatch() already documents (LBN numbering is
* global, not per-VM, but quota scoping still keeps
* two VMs' admissions from evicting each other).
* @param lbn Logical block number being touched.
* @param heartbeat_ticks Current vm->heartbeat.tick_count.
*/
void stadium_block_dispatch(VMUuid vm_id, uint32_t lbn, uint64_t heartbeat_ticks);
/*
* stadium_blocks_print_boot_diagnostics - Console output mirroring
* stadium_words_print_boot_diagnostics(): promotions/evictions for vm_id's
* own block-touch table, plus a conservation check
* (Σ(resident block heat) + reservoir against Q48_ONE is NOT a standalone
* invariant here -- word-execution heat and any other resident shares the
* same reservoir, so this prints the block-only resident sum as a
* diagnostic term, not a claim that it alone should equal Q48_ONE).
*
* @param vm_id The VM whose block-touch table/reservoir to read.
*/
void stadium_blocks_print_boot_diagnostics(VMUuid vm_id);
#endif /* __STARKERNEL__ */
#endif /* STARKERNEL_VM_STADIUM_BLOCKS_H */