> On Tue, Nov 02, 2004 at 03:51:03PM +0100, Stefan Bader wrote: > > Personally I would rather see path information moving into the path > > selector > > We need some concrete examples of what information is missing and > how other sorts of path selectors would benefit from it. What doesn't work: a selector cannot parse its own arguments. Parsing is done from mpath. And the information about paths is actually not required above selector level. I think it is only needed because path group layer and selector layer are mixed. I don't would it help to provide an example patch? > > There are even some fields and functions in dm-mpath.c > > (current_count) that are rather specific to a selector. > > That's a generic optimisation - you normally don't want to be > recalculating which path to use for every single I/O - choose > a path, then send the next X I/O down it before reassessing > which to use. But you would want to let the selector decide that and not the generic code. There might be selectors that want to send IOs to different paths based on other requirements. Ok, you can just set count to 0 then. /Stefan Bader
