Re: [PATCH v1] tools/resolve_btfids: Install subcmd headers

From: Jiri Olsa <olsajiri@gmail.com>
To: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <olsajiri@gmail.com>,
	Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andrii@kernel.org>,
	Martin KaFai Lau <martin.lau@linux.dev>,
	Song Liu <song@kernel.org>, Yonghong Song <yhs@fb.com>,
	John Fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>,
	Stanislav Fomichev <sdf@google.com>, Hao Luo <haoluo@google.com>,
	Connor OBrien <connoro@google.com>,
	Nathan Chancellor <nathan@kernel.org>,
	Kumar Kartikeya Dwivedi <memxor@gmail.com>,
	bpf@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v1] tools/resolve_btfids: Install subcmd headers
Date: Mon, 16 Jan 2023 19:20:04 +0100	[thread overview]
Message-ID: <Y8WVVLEIkXgYvePd@krava> (raw)
In-Reply-To: <CAP-5=fXjjnivhuVSRj7yjaea+6E-YfVKP+Tz2BsTNSqUBDvm4g@mail.gmail.com>

On Mon, Jan 16, 2023 at 09:20:39AM -0800, Ian Rogers wrote:
> On Mon, Jan 16, 2023 at 3:22 AM Jiri Olsa <olsajiri@gmail.com> wrote:
> >
> > On Wed, Jan 11, 2023 at 04:40:24PM -0800, Ian Rogers wrote:
> > > Previously tools/lib/subcmd was added to the include path, switch to
> > > installing the headers and then including from that directory. This
> > > avoids dependencies on headers internal to tools/lib/subcmd. Add the
> > > missing subcmd directory to the affected #include.
> > >
> > > Signed-off-by: Ian Rogers <irogers@google.com>
> > > ---
> > >  tools/bpf/resolve_btfids/Makefile | 19 ++++++++++++++-----
> > >  tools/bpf/resolve_btfids/main.c   |  2 +-
> > >  2 files changed, 15 insertions(+), 6 deletions(-)
> >
> > this depends on changes that went to Arnaldo's tree right?
> > I can't apply this on bpf-next/master
> 
> Hmm.. sorry for that. I did the work on the master branch of
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> 
> > >
> > > diff --git a/tools/bpf/resolve_btfids/Makefile b/tools/bpf/resolve_btfids/Makefile
> > > index 19a3112e271a..de7d29cf43d6 100644
> > > --- a/tools/bpf/resolve_btfids/Makefile
> > > +++ b/tools/bpf/resolve_btfids/Makefile
> > > @@ -35,21 +35,29 @@ SUBCMD_SRC := $(srctree)/tools/lib/subcmd/
> > >  BPFOBJ     := $(OUTPUT)/libbpf/libbpf.a
> > >  LIBBPF_OUT := $(abspath $(dir $(BPFOBJ)))/
> > >  SUBCMDOBJ  := $(OUTPUT)/libsubcmd/libsubcmd.a
> > > +SUBCMD_OUT := $(abspath $(dir $(SUBCMDOBJ)))/
> > >
> > >  LIBBPF_DESTDIR := $(LIBBPF_OUT)
> > >  LIBBPF_INCLUDE := $(LIBBPF_DESTDIR)include
> > >
> > > +SUBCMD_DESTDIR := $(SUBCMD_OUT)
> > > +SUBCMD_INCLUDE := $(SUBCMD_DESTDIR)include
> > > +
> > >  BINARY     := $(OUTPUT)/resolve_btfids
> > >  BINARY_IN  := $(BINARY)-in.o
> > >
> > >  all: $(BINARY)
> > >
> > > +prepare: $(SUBCMDOBJ)
> >
> > do we need special target for that? we already have BPFOBJ dependency
> > placed in the BINARY_IN as prereq
> 
> BPFOBJ is $(OUTPUT)/libbpf/libbpf.a which is needed at link time. The
> prepare step is one we have elsewhere and it creates things like the
> header files necessary to compile the C code, so we need it earlier is
> the answer.
> 
> > why not place both as BINARY_IN prereq, or is there some other reason
> > for new 'prepare' target?
> 
> I was trying to follow the convention elsewhere in the tree of having
> a prepare target that does things like get the necessary header files
> ready. Having prepare is useful because if an additional dependency is
> added, then it just needs to be added to prepare. It could be tedious
> to list all the dependencies for every C file, although Makefile.build

ok, could we maybe add the BPFOBJ in prepare target as well?

> handles most of that. It isn't clear to me why $(BPFOBJ) is a
> dependency of $(BINARY_IN) as it is already a dependency of $(BINARY).

I think that if you specify OUTPUT then we need the libbpf headers
to be created before we go to compile resolve_btfids objects

thanks,
jirka