On Tue, Jun 20, 2017 at 6:38 PM, Jason Gunthorpe <jgunthorpe@xxxxxxxxxxxxxxxxxxxx> wrote: > On Tue, Jun 20, 2017 at 03:03:36PM +0300, Matan Barak wrote: > >> What about the specification deceleration? >> Since this patch-set builds everything statically, we need somehow to >> hash the entities correctly >> according to their higher ids. > > Don't try and build everything statically. > > Do what I suggested in Santa Fe - have a .rodata description that is > easy for the programmer to build and understand, 'compile' that into a > fast-access version (eg a hash table, etc) that is suitable for runtime > use. > I started coding this. Actually, method becomes just some info + an array of attributes at some arbitrary order. Every attribute carries its own id. At driver initialization, we build the compact hash. >> When a developer wants to add additional attributes he can't just >> add them wherever he wants. He needs to understand the >> implications. Attributes which are added as common attribute ids are > > We can look at this stuff based on the attribute ID value and where in > the code it lives, not based on the type of a .rodata structure. > I guess the static->dynamic mechanism already solves that :) > Jason Matan -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
