Re: Status of the IO scheduler fixes for 2.4

[Chris Mason <mason@suse.com>]

On Thu, 2003-07-03 at 08:31, Marc-Christian Petersen wrote:
> On Thursday 03 July 2003 04:02, Chris Mason wrote:
> 
> Hi Chris,
> 
> > So, the patch attached includes the q->full code but has it off by
> > default.  I've got code locally for an elvtune interface that can toggle
> > q->full check on a per device basis, as well as tune the max io per
> > queue.  I've got two choices on how to submit it, I can either add a new
> > ioctl or abuse the max_bomb_segments field in the existing ioctl.
> > If we can agree on the userland tuning side, I can have some kind of
> > elvtune patch tomorrow.
> what about /proc ?

Always an option.  If elvtune didn't exist at all I'd say proc was a
better choice.  But I do want to be able to tune things on a per device
basis, which probably means a new directory tree somewhere in proc.  Our
chances are only 50/50 of getting that patch in without a long thread
about the one true way to access kernel tunables through an fs 
interface visible to userland ;-)

For the most part I'm only visiting drivers/block/*.c right now, so I'll
code whatever interface the long term maintainers hate the least.

-chris