On Wed, May 04, 2022 at 10:19:20AM -0700, "T.J. Mercier" <tjmercier@xxxxxxxxxx> wrote: > Should I export these now for this series? Hehe, _I_ don't know. Depends on the likelihood this lands in and is built upon. > No, except maybe the gpucg_bucket name which I can add an accessor > function for. Won't this mean depending on LTO for potential inlining > of the functions currently implemented in the header? Yes. Also depends how much inlining here would be performance relevant. I suggested this with an OS vendor hat on, i.e. the less such ABI, the simpler. > I'm happy to make this change, but I wonder why some parts of the > kernel take this approach and others do not. I think there is no convention (see also Documentation/process/stable-api-nonsense.rst ;-)). Regards, Michal
