attaching to running processes, it completely breaks normal debugging.
A better fix is in the works, but cannot be properly tested until
the problem with gdb hanging the system in -current is solved.
WWNs correctly (Again!) - this time for the case that we're not going
to fully init the adapter if isp_init is called (with ISP_CFG_NOINIT
set in options). The pupose for this is to bring the adapter up to
almost ready to go, get info out of NVRAM, but to not start it up- leaving
it until later to actually start things up if wanted (and possibly with
different roles selected).
process. This fixes a problem when attaching to a process in gdb
and the process staying in the STOP'd state after quiting gdb.
This whole process seems a bit suspect, but this seems to work.
Reviewed by: peter
with the driver locking up under load.
- Restructure so that we use a static pool of commands/FIBs, rather than
allocating them in clusters. The cluster allocation just made things
more complicated, and allowed us to waste more memory in peak load
situations.
- Make queueing macros more like my other drivers. This adds queue stats
for free. Add some debugging to take advantage of this.
- Reimplement the periodic timeout scan. Kick the interrupt handler
and the start routine every scan as well, just to be safe. Track busy
commands properly.
- Bring resource cleanup into line with resource allocation. We should
now clean up correctly after a failed probe/unload/etc.
- Try to start new commands when old ones are completed. We weren't doing
this before, which could lead to deadlock when the controller was full.
- Don't try to build a new command if we have found a deferred command.
This could cause us to lose the deferred command.
- Use diskerr() to report I/O errors.
- Don't bail if the AdapterInfo structure is the wrong size. Some variation
seems to be normal. We need to improve our handing of 2.x firmware sets.
- Improve some comments in an attempt to try to make things clearer.
- Restructure to avoid some warnings.
in 4.2-REL which I ripped out in -stable and -current when implementing the
low-memory handling solution. However, maxlaunder turns out to be the saving
grace in certain very heavily loaded systems (e.g. newsreader box). The new
algorithm limits the number of pages laundered in the first pageout daemon
pass. If that is not sufficient then suceessive will be run without any
limit.
Write I/O is now pipelined using two sysctls, vfs.lorunningspace and
vfs.hirunningspace. This prevents excessive buffered writes in the
disk queues which cause long (multi-second) delays for reads. It leads
to more stable (less jerky) and generally faster I/O streaming to disk
by allowing required read ops (e.g. for indirect blocks and such) to occur
without interrupting the write stream, amoung other things.
NOTE: eventually, filesystem write I/O pipelining needs to be done on a
per-device basis. At the moment it is globalized.
