On Tue, 2009-01-27 at 12:15 +0100, Florian Sesser wrote: > Hi! > > Thanks for your comment. My intention was having a logical seperation to > enable better discussion of my patch, which I think is pretty much of a > monster and not yet ready for inclusion? > > What do you think about my selection of mesh_path_sel_ops and the > storing of mesh PP/PM/CC id? > > If you deem it OK I would be happy to reformat the patch (into three or > so seperatly compiling patches) and include KConfigs. I also see a lot of EXPORT_SYMBOLs that should not be necessary, but I need to look through it in more detail. I think your layering may be broken in some parts. Why, for example, do you need to look up path selection algorithms by name? I also think you should wrap things like osdata->u.mesh.mesh_pp->ops->nexthop_lookup_fn into small static inlines that can go away if only one algorithm is built in, for example. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
