[kvnloo]

Summary

Clarify the difference between:

./rpiboot -d ./mass-storage-gadget64

and running rpiboot with no payload directory.

The explicit -d form uses files from the current checkout, while the no-argument form uses the installed payload under:

INSTALL_PREFIX/share/rpiboot/mass-storage-gadget64

Also document using rpiboot -v to verify the paths actually loaded and reinstalling with make install when testing local changes.

This should make stale installed-payload mismatches much easier to identify.

Refs #391

[kvnloo]

Integrated current master 0fd2663f869b7f325f4481049c60d6394855260e without rewriting the original commit. Tested/pushed head 745fb73; only Readme.md differs from current master. The doc now preserves the installed → CWD → executable-adjacent fallback order and the explicit-checkout/verbose-path guidance; rebuilding alone does not refresh installed payloads.

Native ARM64 software checks: isolated-prefix make, make install, existing test/validate-hsm-wrapper.sh signing comparisons, installed-payload hash comparison, and real rpiboot -V all passed. Explicit checkout and installed -d … -v invocations print/open their expected bootfiles.bin, then exit255 at libusb_init in this environment. No-directory selection fails before fallback selection, so that fallback is source-checked only; no physical USB boot or payload-transfer claim.

New-head CI39102394743 is action_required, zero jobs executed, not a test failure or green CI. Marking the documentation change ready for review; historical approval was on the old head, not new-head acceptance. Original authorship is preserved; current integration/verification was AI-assisted.

[pelwell]

Apart from the spurious ./s the change looks okay, which is in contrast to the unreadable slop that is the most recent comment.

[pelwell]

The leading ./ on the directory name is completely unnecessary - it does not belong in official documentation.

[pelwell]

Ditto here.

[timg236]

I'm seeing some merge conflicts here