On Mon, Feb 19, 2024 at 05:58:41PM +0106, John Ogness wrote: > On 2024-02-19, Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote: > >>> The headers in this file is a mess. But here you can at least keep the > >>> piece ordered, can you? > >> > >> Just to clarify, you would like to see this ordering and inclusion? > > > > Roughly, yes. Ideally it is quite likely that kernel.h is being used as > > a 'proxy' header. Nowadays, it's rare the code needs kernel.h. > > So I took the time to painfully discover every header that is required > for nbcon.c without any proxy usage. It came down to this: > > #include <linux/atomic.h> > #include <linux/bug.h> > #include <linux/compiler.h> This is guaranteed to be included by types.h, can be dropped. > #include <linux/console.h> > #include <linux/delay.h> > #include <linux/errno.h> > #include <linux/export.h> > #include <linux/init.h> > #include <linux/irqflags.h> > #include <linux/minmax.h> > #include <linux/percpu-defs.h> This... > #include <linux/preempt.h> > #include <linux/serial_core.h> > #include <linux/slab.h> > #include <linux/smp.h> ...and this I believe can be represented by percpu.h as most likely that is the "main" library you are using. > #include <linux/stddef.h> > #include <linux/string.h> > #include <linux/types.h> > #include "internal.h" > > For the next version of this series I will only add the includes you > suggested, but will follow-up with a patch that fixes all proxy headers > for nbcon.c. As a separate patch it will help with bisecting in case the > ordering causes an explosion on some config/architecture. Sure, thanks! -- With Best Regards, Andy Shevchenko
