Messages in this thread |  | | Date | Tue, 1 Nov 2016 22:59:05 +0000 | From | Wei Yang <> | Subject | Re: [PATCH] kbuild: add %/built-in.o target in top Makefile |
| |
On Tue, Nov 01, 2016 at 05:27:34PM +0100, Michal Marek wrote: >On 2016-10-24 14:45, Wei Yang wrote: > >> diff --git a/Makefile b/Makefile >> index addb235..546998e 100644 >> --- a/Makefile >> +++ b/Makefile >> @@ -1627,6 +1627,8 @@ endif >> $(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@) >> %.symtypes: %.c prepare scripts FORCE >> $(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@) >> +%/built-in.o: prepare scripts >> + $(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@) > >This rule gets in the way of > > 975 # The actual objects are generated when descending, > 976 # make sure no implicit rule kicks in > 977 $(sort $(vmlinux-deps)): $(vmlinux-dirs) ; >
Yes, second level built-in.o is built because of this rule, while for deeper level built-in.o we don't have a rule now. This patch is trying to add a rule for those deeper level built-in.o.
>Quickly looking at the Makefile, it may be fine, but it needs to be >confirmed. This also explains why the current behavior with */built-in.o >is inconsistent. >
Agree with you.
One of my concern is whether it is necessary to add "FORCE" to the prerequest. I have tried on my machine, it works fine without "FORCE", while maybe it won't work fine with some version of "make".
This is what I want to confirm. Any other potential issue, just let me know.
Thanks for your comment :-)
>Michal
-- Wei Yang Help you, Help me
|  |