On Mon, Jul 01, 2024 at 05:50:35PM -0700, Andrii Nakryiko wrote: > On Mon, Jul 1, 2024 at 5:32 PM Brian Norris <briannorris@xxxxxxxxxxxx> wrote: > > --- a/tools/lib/bpf/Makefile > > +++ b/tools/lib/bpf/Makefile > > @@ -153,7 +153,11 @@ $(BPF_IN_SHARED): force $(BPF_GENERATED) > > echo "Warning: Kernel ABI header at 'tools/include/uapi/linux/if_xdp.h' differs from latest version at 'include/uapi/linux/if_xdp.h'" >&2 )) || true > > $(Q)$(MAKE) $(build)=libbpf OUTPUT=$(SHARED_OBJDIR) CFLAGS="$(CFLAGS) $(SHLIB_FLAGS)" > > > > -$(BPF_IN_STATIC): force $(BPF_GENERATED) > > +$(STATIC_OBJDIR): > > + $(Q)mkdir -p $@ > > + > > +$(BPF_IN_STATIC): force $(BPF_GENERATED) | $(STATIC_OBJDIR) > > wouldn't $(BPF_IN_SHARED) target need a similar treatment? Hmm, probably. I'll admit, I only debugged errors that show up in the top-level kernel build. The only tools/bpf stuff tied to the top-level build is resolve_btfids, which uses the static library, not the shared. And now that I poke around at some other build targets, this highlights that my patch introduced some problems with the independent libbpf build. Particularly, $(OUTPUT) is a relative path in some cases, and that relative path gets interpreted differently in recursive make (when we change cwd). So please don't accept this patch as-is. Brian > > + $(SILENT_MAKE) -C $(srctree)/tools/build CFLAGS= LDFLAGS= OUTPUT=$(STATIC_OBJDIR) $(STATIC_OBJDIR)fixdep > > $(Q)$(MAKE) $(build)=libbpf OUTPUT=$(STATIC_OBJDIR) > > > > $(BPF_HELPER_DEFS): $(srctree)/tools/include/uapi/linux/bpf.h
