hi...
> One thing someone *did* mention though was that Linux(?) now has (or is
> working on?) low level functions to do this...I'm not sure what would be
> involved to use this functionatilty though...anyone in the Linux camp able
> to respond?
>
the latest raw i/o patches for 2.2 and 2.3 were released in august (by Stephen
C. Tweedie).. they are available at:
ftp://ftp.uk.linux.org/pub/linux/sct/fs/raw-io/
while these mods were designed w/Linus Torvalds and passed as OK by him, the
big question still remains as to whether this will ever make it into the main
kernel distribution...
and that probably is not going to happen unless/until there is a change of
heart in the kernel development leadership... a direct quote from Linus
Torvalds:
"I do not believe in raw IO - even for streaming audio it's just too common for
the data to have been available in the cache, and by using raw IO you (for
absolutely no good reason) just made the machine do more IO than it should
have.
There are very specific cases where the application knows that its dataset is
larger than physical memory, but those tend to be limited to quite large
problems. And they're getting larger. "
so, while the patches are there and "officially sanctioned" as it were, it
probably won't be a standard issue item and will remain in the form of patches
you have to fetch and apply yourself...
for the people who will benefit from raw io, however, this probably won't be a
huge issue as their installations will probably be highly tuned and tweaked
anyways..
in short: the support is there for raw io in the linux kernel, just not the
standard distribution. it will probably remain for as long as there is interest
(which seems to mostly come from database concerns)..
--
Aaron J. Seigo
