> Odd, both my hard disks are UDMA. You're saying unmaskirq has
> no effect on UDMA systems?
absolutely. the whole point of unmaskirq is to drop the mask
during a PIO rep insw/outsw.
> What about CDROMs? My guess is that they still use PIO and
no, all the ones I see on the market now are DMA and even UDMA.
though I think that the actual ATAPI packet is transfered with PIO
(but it's trivial in size/time).
> thus unmaskirq will help there (your HDD can receive
> interrupts while your CDROM is seeking).
afaik, unmaskirq refers to global interrupt masking that takes
place only during PIO transfers. note that this masking does not
happen during a seek, only a transfer.
there's another issue, regarding concurrency of master and slave
on a single channel. normally, the controller can only talk to
one drive at a time. the new disconnect stuff (ata5 "overlapped
feature set") permits the controller to select the other drive
if one drive chooses to disconnect. the issue here is really
arbitration among master and slave for access to the channel's
service interrupt.
> Well, from the source code it seems the unmaskirq is always used
> except you're saying that with modern systems the period that
> interrupts are disabled is so short to not matter?
no, modern systems use DMA/UDMA, for which unmaskirq is irrelevant.
regards, mark hahn.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 23 2000 - 21:00:19 EST