On Mon, Oct 11, 2021 at 11:20:57AM +0100, Jonathan Cameron wrote: > Mark Brown <broonie@xxxxxxxxxx> wrote: > > As for SVE we will track a per task SME vector length for tasks. Convert > > the existing storage for the vector length into an array and update > > fpsimd_flush_task() to initialise this in a function. > I'm clearly having a trivial comment day. Given reduction in indenting > it would be nice perhaps to reformat comments to take that into account. Again I'll have done this to make it clearer that things are just being moved about - as a reviewer I do find things like that very helpful. > I'm also unconvinced the trivial wrappers are worthwhile. (maybe you drop > those later?) They do hide what we're doing from the rest of the series which makes the whole thing a lot easier to work with, especially if we change what data structures we use or there's some debate as to the names of the constants. A bunch of them directly map onto existing trivial wrappers too, there's been some stylistic preference for that. If people want to remove the wrappers I'd propose leaving them for now then adding a patch afterwards which removes them, or at least waiting until we've got very firm agreement from Catalin and Will on the data structures and constants.
Attachment:
signature.asc
Description: PGP signature
