On Thu, Mar 23, 2017 at 03:23:28PM +0100, Daniel Vetter wrote: > On Wed, Mar 22, 2017 at 04:22:38PM +0200, Joonas Lahtinen wrote: > > + Daniel for the rsvd2 > > I'd go with a shadow struct definition which matches the uapi one exactly, > except for having a proper name and type for this ... Or we do the > anynomous union thing again. Definitely not an anon union since this is not part of the ABI, just taking advantage of the overallocation in the kernel copy of the user structs. I've plonked it behind a macro so that we have only one point of failure, though realistically expanding the struct or using a lut[] will take more massaging than just renaming the field. It's not a big job so I don't feel like the code is hamstringing our future selves. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
