>
>
> On Fri, 4 Oct 2002, Andrew Morton wrote:
> >
> > > Especially since I thought that O_DIRECT on the regular file (or block
> > > device) performed about as well as raw does anyway these days? Or is that
> > > just one of my LSD-induced flashbacks?
> > >
> >
> > Now we're not holding i_sem for O_DIRECT writes to blockdevs,
> > I don't think the raw driver offers any advantages at all. It's
> > a compatibility thing to save people from having to add "|O_DIRECT" to
> > their source and then typing `ln -s /dev/hda1 /dev/raw/raw0'.
>
> Maybe the raw driver could become a shell that just adds the O_DIRECT?
> Unless it can do something more, of course..
>
> Linus
>
>
Only issue would be the alignment restriction on blockdev versus raw device.
raw allows 512 byte alignment on userbuff, offset and length.
blockdevice might need 1024/2048 byte alignment.
If we get the alignment patch for DIO, this is not an issue.
(alignment patch - in case of defualt alignment problem,
go down to blkdev_hardsect_size() alignment.)
- Badari
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Oct 07 2002 - 22:00:48 EST