Am Samstag, den 07.03.2015, 10:57 +0000 schrieb Mark Brown: > On Fri, Mar 06, 2015 at 03:16:55PM +0100, Philipp Zabel wrote: > > > #include <linux/sort.h> > > -#include <trace/events/regmap.h> > > > > #include "internal.h" > > > > +#include <trace/events/regmap.h> > > + > > The change is basically OK but the above is a bit too funky for my > taste - it feels like it's asking for annoying compile breakage if > anyone changes anything. I'd be more inclined to move the header > somewhere where the events stuff can see it than do this.
Good point. As I see it I have three possiblities now: a) Just #include "../../../drivers/base/regmap/internal.h" in include/trace/events/regmap.h b) Move drivers/base/regmap/internal.h to include/linux/regmap-internal.h and #include <linux/regmap-internal.h> in include/trace/events/regmap.h c) Move include/trace/events/regmap.h to drivers/base/regmap/trace.h and just #include "internal.h" in drivers/base/regmap/trace.h Which, if any, of these options would you prefer? Here is an unscientific survey: a) $ git grep "#include \"\.\." include/ include/linux/libfdt.h:#include "../../scripts/dtc/libfdt/fdt.h" include/linux/libfdt.h:#include "../../scripts/dtc/libfdt/libfdt.h" include/xen/interface/io/vscsiif.h:#include "../grant_table.h" b) $ find include/ -name "*internal.h" include/linux/timekeeper_internal.h c) $ find drivers/ -name trace.h drivers/net/wireless/ath/ath10k/trace.h drivers/net/wireless/ath/trace.h drivers/net/wireless/ath/wil6210/trace.h drivers/net/wireless/ath/ath5k/trace.h drivers/net/wireless/ath/ath6kl/trace.h drivers/usb/dwc3/trace.h drivers/misc/cxl/trace.h a) and b) aren't very common, and c) is mostly used by driver-level, not subsystem-level trace events. regards Philipp -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
