henry@utzoo.UUCP (Henry Spencer) (10/18/84)
Well, the RM driver code is yielding up some of its secrets. There is
an explicit bypass in the code for write-lock errors; they are assumed
incurable and so the driver doesn't retry them. An explicit error message
should probably be produced, though, since we don't normally run any of
our drives write-protected.
There is a curious indication, however, that the driver's error-handling
behavior is not all it should be. There are signs that it is only doing
four retries, instead of the statutory 28, on some classes of errors.
There is code in the driver that explicitly attempts a Recalibrate
operation after -- you guessed it -- four retries, and we suspect that
the Recalibrate is being botched somehow. This may be a subtle effect
of differences between the RP06 that the driver was written for, and
the RM03 emulator that it is actually running here. Stay tuned.
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,linus,decvax}!utzoo!henry