You've already forked torvalds-GuitarPedal
mirror of
https://github.com/torvalds/GuitarPedal.git
synced 2026-08-17 13:09:42 +00:00
5c1d15e687
test-bench, test-analog and measure-load all took whatever enumerated first, and test-audio took its card from one enumeration and its MIDI port from another - so with two boards attached it could have measured the audio of one and talked to the other. test-bench and test-analog at least refused to guess and skipped; the other two did not. All four go through pedal.discover() now and take --target (-t for measure-load), which is passed to pedal.find() - already written to refuse ambiguity rather than answer with the first hit. Refusing is still the default: no target and more than one board is a skip, not a guess. The Makefile passes TARGET= through, the same way check-boot already did. This is not hypothetical tidying. Two boards of one revision differ only in their serial, and the run that matters is the one where a suspected board is compared against a known-good one - which is exactly when picking the wrong one is both easy and invisible. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>