On Thu, Jul 24, 2025 at 12:15:11PM +0100, Lorenzo Stoakes wrote: > On Thu, Jul 24, 2025 at 01:12:49PM +0200, Andy Shevchenko wrote: > > On Thu, Jul 24, 2025 at 1:10 PM Lorenzo Stoakes > > <lorenzo.stoa...@oracle.com> wrote: > > > On Thu, Jul 24, 2025 at 11:24:29AM +0200, Bartosz Golaszewski wrote: > > > > +extern const void *kmemdup_const(const void *src, size_t len, gfp_t > > > > gfp); > > > > > > Please drop extern, it's unnecessary. > > > > It's all over the header. This should be done as a precursor patch and > > I know that usually people push back on doing that. I gave up on this. > > Kernel is going to rot sooner or later... :-( > > In mm we just update as we go, this is probably the best approach to avoid > unnecessary churn.
I agree on the idea of eliminating it, but also I agree on the consistency over redundancy. That's why I prefer to see this done at once for all (in the same header) than doing one-by-one. And this approach got a lot of pushes back, while the former even more pushed back on the (in)consistency matters. -- With Best Regards, Andy Shevchenko
