Neil,
I think Linus knows this, and from what I've seen, it may break a ton of
shit. I think folks need to let Linus know these issues. NT and
NetWare both support variable length and offset sectors read/write
requests. This fixed block size stuff is tough for non-Linux file
systems like NTFS and NWFS.
Jeff
Neil Brown wrote:
>
> On Saturday September 2, jmerkey@timpanogas.com wrote:
> >
> > Linus,
> >
> > The attached patch is submitted to enable variable sector size block
> > chaining via ll_rw_block() in the I/O subsystem layer.
> >
> > Jeff904a905,907
> > > /****
> > > // This code is being commented out to allow support for variable chained
> > > // block I/O requests. Jeff V. Merkey
> > 915a919
> > > *****/
>
> While I think this patch is a good idea, I feel I should let you know
> that using the functionality on a Software-RAID-5 device will cause
> terrible performance.
> Whenever raid5.c notices the block size has changed, it flushes it's
> stripe cache, and rebuilds it with the new size. This effectively
> serialises requests around any size change.
>
> Fixing this in on my todo list, but I haven't come up with a really
> neat model yet.
>
> NeilBrown
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> Please read the FAQ at http://www.tux.org/lkml/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Sep 07 2000 - 21:00:22 EST