On Sun, May 19, 2024 at 11:02 AM Kent Overstreet <kent.overstreet@xxxxxxxxx> wrote: > > On Sun, May 19, 2024 at 09:48:49AM -0700, Linus Torvalds wrote: > > On Sun, 19 May 2024 at 08:32, Linus Torvalds > > <torvalds@xxxxxxxxxxxxxxxxxxx> wrote: > > > > > > I'm going to take this pull and fix up the cases I find, but I'm not > > > happy with this kind of trivial C preprocessor misuse. > > > > I did some other maco handling cleanup too and tried to regularize > > some of this all, and it seems to work for me. But somebody should > > double-check, and it's possible these patterns should all be > > regularized further with a few helper macros for the whole "add > > __GFP_ZERO to argument list" or similar. > > I just double checked slab.h, gfp.h and percpu.h, and scanned through > the diff vs. 6.9 for include/linux/ - looks like you got everything. Sorry about that. Yeah, I could not find any other place that was not fixed. Thanks for noticing and fixing them! > > I think we can slim down the API surface of slab.h some more too, we're > now exposing three different ways of saying "trace/track this allocation > here": _trace, _track_caller and _noprof vs. normal; I think after a > cycle we can see if the old variants are still needed or can be > consolidated somehow. >
