Perhaps EDOOFUS

Noticed same issue in D26959, but commenting on already committed code so it could be fixed if my observation is correct.

Using sizeof() this way here looks bogus, shouldn't it be either sizeof(array)/sizeof(array[0]) or simply nitems(array)?

Same sizeof() issue.

Just pass errors that are outside the range AS-IS.

This function has no consumers at this point, so I'll just remove it - it's kind of pointless to guess what the semantics should be. The table is necessary for fuse(4), but it includes linux_errno.inc directly, to avoid dependency on linux KLD.

D'oh! Thanks for reporting it.

Should this code bet written like this?

if (error >= 0 && error <= ELAST)
  return (linux_errtbl[error]);
else
  return (XXX);

Not really; this is for errors generated by the kernel, so we trust the error is correct.