ide-scsi problems with multiple CD-Rs on same channel (2.2.14)

[Tkil]

>>>>> "Gadi" == Gadi Oxman <gadio en netvision.net.il> writes:

Gadi> I'd recommend changing choose_drive() in drivers/block/ide.c
Gadi> as follows -- replacing the "drive = hwgroup->drive;" line with
Gadi> "drive = hwgroup->drive->next;" line, and commenting out the
Gadi> part later which tests the best->nice1 flag (the second part
Gadi> of the patch can alternatively be done by using "echo nice1:0 >
Gadi> /proc/ide/hdx/settings for each IDE drive).

Gadi> This will select round-robin scheduling between the drives. This
Gadi> is usually not the best for throughput, but might be more
Gadi> appropriate when real-time constrains are involved.

Well, a few issues:

1. Switching to hwgroup->drive->next causes serious problems at boot.
   The four IDE devices are detected, but it stops at the "partition
   check" for hda.

   So I reverted this portion of the patch.  It might still be usable,
   but I suspect that we would need to fix the termination condition
   there too.

2. I'm not sure if this patch fixed it or not, but I can now start a
   burn onto hdc, and the second burn onto hdb, successfully.  Under
   the stock 2.2.14, I believe this caused problems, but I would have
   to run it again to make sure.

So, the short form is that I can get two burns (hdb and hdc) working
flawlessly.  Starting a third burn (so far just in -dummy mode) always
hoses the second burn (assuming hdb, hdc, hdd order).  I *can* read
from all three devices simultaneously, as my previous e-mail indicated
(with the exception of the "setup" period, which seems to block all
other traffic off that bus).  These drives do have 2MB buffers on
them, so we should be able to take advantage of any burst traffic we
can provide.

I will probably test 2.3.40 on this box pretty soon, and see if that
gives us any better luck.

Thanks again,
T.



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo en vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/