Conor Dooley <conor@kernel.org> says:

From: Conor Dooley <conor.dooley@microchip.com>

I noticed ~today~ while looking at the isa manual that I had not
accounted for another couple of edge cases with my regex. As before, I
think attempting to validate the canonical order for multiletter stuff
makes no sense - but we should totally try to avoid false-positives for
combinations that are known to be valid.

* b4-shazam-merge:
  dt-bindings: riscv: fix single letter canonical order
  dt-bindings: riscv: fix underscore requirement for multi-letter extensions

Link: https://lore.kernel.org/r/20221205174459.60195-1-conor@kernel.orgSigned-off-by: Palmer Dabbelt <palmer@rivosinc.com>