Files

BlackMark 8f6319c068 pureboot 7: the same features in fewer words on every chip

Four cuts, none touching what the loader can do. The entry stub stops
re-doing the reset logic's own SP write where the datasheet guarantees
RAMEND (stack::hardware — the classic megas keep theirs). The autobaud
unit moves into GPIOR2:GPIOR1 wherever the chip has the pair: one-word
accesses, no RAM object, and the host's measured-clock peek follows it
by version and geometry. 'J' rides the unified decode, carrying a
selector it ignores so its address is the same two reads as every other
command — the tool sends the bare form to older residents. run_app stops
insisting on a body of its own. The fleet lands at 358–410 B stock and
438–474 B autobaud; the tightest image in the space — the 1284s'
autobaud on a USART's own pins with the OSCCAL trim — drops from 510 to
484 of its 512. Every chip's suite is green on the wire that changed,
and the README's table is machine-checked against the built images.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-07-28 17:02:38 +02:00

3.8 KiB

Raw Permalink Blame History

Atmel Studio

master carries bootloader.atsln, so this branch does too: ide/bootloader.atsln builds the loaders from the same sources Ninja does, to a byte-identical .text — 390 B for the 328P pureboot loader, 510 B for the tsb_asm tier in its 512-byte section. CMake remains the build system; the solution is here so the port opens in Studio as its predecessor did.

The two projects, and why two

pureboot is a chip × backend × clock × baud matrix — pureboot_add_loader() resolves a deployment into compile definitions — and a .cppproj is one binary at one set of flags, so a project can only ever be one point of it. pureboot is that point: the stock 328P deployment, USART0 at 115200 on a 16 MHz crystal, an 8-second activation window. tsb_asm is the TinySafeBoot tier that occupies the same 512-byte section master's tsb project targeted.

The other three tsb tiers (tsb_pure, tsb_tricks, tsb_policy) are not here. They differ from tsb_asm in their source file, their section size, and — for tsb_policy — two loop flags; nothing about that is a Studio concern, and what they exist to demonstrate is a size gradient only the CMake size tests measure. Adding one is a copy of tsb_asm/tsb_asm.cppproj in its own directory, with its name, its GUID, its source path and its --section-start changed (0x7c00 for the 1 KiB tiers), plus four lines in the solution.

avrdevice is a project property, so each project gets its own directory: Studio builds into <project dir>/<Configuration> whatever OutputDirectory says, and two projects sharing a directory would share one object file.

Debug keeps `-Os`

Both configurations compile at -Os; Debug adds only -gdwarf-4. The .text is therefore identical in both, which is the point — a loader's section is a correctness bound and not a budget. -Og builds this same source to 590 B, and linking it at --section-start=.text=0x7e00 on a 32 KiB part puts 78 bytes past flash end without a diagnostic: rcall/rjmp targets there wrap modulo flash size, so the image dies right after activation. A debug configuration that silently produces that is worse than none, and DWARF costs no flash, so the optimisation level stays where correctness needs it.

What Studio needs from the machine

libavr from the submodule, found at $(MSBuildProjectDirectory)\..\..\libavr\include — correct by construction, and anchored to the project because a plain relative path resolves against the generated makefile's directory (the configuration's output directory), not the project's. There is no LIBAVR_ROOT escape hatch: a variable exported in a shell is invisible to Studio launched from the Start menu, and the failure reads as a missing libavr/libavr.hpp — which is what the submodule answers.

A GCC 16.1 toolchain registered as flavour avr-g++-16.1.0, nothing older reaching -std=c++26.

Generating and gating

One generated file is required before a project will load at all, and one command per project checks the flags have not drifted (both from libavr's tools/atmelstudio/):

for name in pureboot tsb_asm; do
    python libavr/tools/atmelstudio/componentinfo.py \
        "ide/$name/$name.componentinfo.xml" --device ATmega328P
    python libavr/tools/atmelstudio/check-flags.py \
        --solution ide/bootloader.atsln --project "$name" --target "$name" \
        --compile-commands build/atmega328p-generated/compile_commands.json \
        --log "build/as-$name.log"
done

--project because one reference describes one binary; --target because pureboot.cpp is compiled by every point of the size matrix and the flags differ per point, so the basename alone does not name a reference. Release is what the gate compares — the presets define no debug build, and Debug differs from Release only in -gdwarf-4.

Legacy (the yazoalfa-era submodules) stays on master.

3.8 KiB Raw Permalink Blame History Unescape Escape