On Mon, 10 Jun 2019, Lucas De Marchi <lucas.de.marchi@xxxxxxxxx> wrote: > On Thu, Jun 06, 2019 at 11:42:46AM +0300, Jani Nikula wrote: >>On Tue, 04 Jun 2019, Imre Deak <imre.deak@xxxxxxxxx> wrote: >>From the Nitpicks Department, please add a blank line here. >> >>> +#include "intel_display.h" >>> +#include "i915_drv.h" >>> +#include "intel_tc.h" >> >>And sort the includes. > > humn... In some other projects I use the convention of always having the > correspondent header to be included first and the rest sorted. Here it > would be like: > > #include "intel_tc.h" > > #include "intel_display.h" > #include "intel_drv.h" > > This way we guarantee that intel_tc.h header is self-contained: anything > it needs, it includes or declares. If we adopted such convention I don't > see why we would still need Makefile.header-test for example. I think the long standing convention is to have <asm>, <linux>, and <drm> headers included before "local" headers. Eventually the version of header tests that I've submitted to kbuild upstream will obsolete the Makefile.header-test files, and I expect us to have simply header-test-y := $(notdir $(wildcard $(src)/*.h)) in there, testing everything. BR, Jani. -- Jani Nikula, Intel Open Source Graphics Center _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx