|
I have updated my patches to
work against 2.3.22, and these should be up on http://www.andante.org/scsi_queue.html.
I have also finished dealing with the single_lun case, which was the only
outstanding outage that I am aware of.
FWIW, I have tested it more
extensively on my test system, and it seemed to stand up, at least until one of
the disks that I was using for testing purposes died. The previous
problems that I was having with the aic7xxx driver went away when I switched to
a newer Adaptec card - for some reason the interrupt line was getting stuck on
with the old card, and I need to stay focussed on getting these patches
shepherded into the distribution kernel.
The testing that I have done
with the live system was on 2.3.17 - I am getting bitten by an inability to
assemble bootsect.s with 2.3.22 (.org directive moves IP backwards), and I need
to browse linux-kernel to figure out what got broken and how to fix
it.
My thinking is that I would like
to submit this into the main distribution kernel within the next two weeks
or so, barring any problems or strong objections. Given that it is
possible for people to select the old queueing code at kernel config time, the
main objective is to make it easier for people to test, and to reduce the number
of stupid merge conflicts that I have to deal with.
-Eric
|
