Artem Kachitchkine wrote:
>
>> Yes, I intend to try it out.
>
> Great.
>
>> I'm not specifically targetting ZFS, but I can't see why it wouldn't
>> work, other than the device id concern raised below.
>
> Another vector to keep an eye on, wrt ZFS, is write cache (6424510 usb
> ignores DKIOCFLUSHWRITECACHE).
So does SDcard, but that is only because it has no cache. I added the
B2S_CMD_SYNC command in my original proposal for flushing out cache, but
I'm starting to wonder if that is too simplistic...
SCSI can enable/disable various cache tuning parameters. Maybe I need
to support that. Possibly later, if some device that has write cache
ever shows up. As I said, right now I don't have it, so its a non-issue.
-- Garrett
