The rule is that you don't get to call scheduler-related functions (e.g.
wakeup()) above the clock interrupt. Going to statclock unnecessarily
hoses e.g. serial interrupts on the SPARC.
Vrkiu_cngetc() had returned WSCONS_EVENT_ALL_KEYS_UP if there are no event to
report. Modification keys like shift-key had been canceled by that event.
be issued/completed in order; that is, provide a barrier for I/O queues.
- Change the buffer driver queue links to a TAILQ, rather than using
a home-grown equivalent. Provide BUFQ_*() macros to manipulate buffer
queues; these deal with the barrier provided by B_ORDERED.
- Update disksort() accordingly, and provide 3 versions:
- disksort_cylinder(): historical disksort(), which keys on
b_cylinder (and b_blkno for the case when b_cylinder matches).
- disksort_blkno(): sorts only on b_blkno. Essentially the
same as disksort_cylinder(), but with fewer comparisons.
- disksort_tail(): requests are simply inserted into the queue
at the tail. This is provided as an option so that drivers
can simply have a pointer to the appropriate sort function.
Note that disksort() now pays attention to B_ORDERED.
* Improve communication between the `hard' and `soft' interrupt handlers
to better distinguish various interrupt sources.
* Eliminate several race conditions where we would set a time out handler
only after starting the command on the hardware.
* Handle most timeouts by resetting the controller; there isn't much chance
of recovery in any other way. Currently, the exception is a timeout
on I/O, in which case we first try to pulse the controller's TC line
in order to abort the pseudo-dma sequence. Apparently, "normal"
conditions can induce such a timeout when there's no disk in the drive.
* Reduce the formatting gap parameter to 0x54.
* On the obio bus, interpret the `status' attribute.
* Minimize console diagnostic output if the errors we get appear to be
caused by the absence of a disk.
