On Tue, Mar 15, 2022 at 4:10 PM Daniel Xu <dxu@xxxxxxxxx> wrote: > > Hi Alexei, > > On Tue, Mar 15, 2022, at 2:38 PM, Alexei Starovoitov wrote: > > On Sun, Mar 13, 2022 at 4:01 PM Daniel Xu <dxu@xxxxxxxxx> wrote: > >> > >> A concern about potential GPL violations came up at the new $DAYJOB when > >> I tried to vendor the vmlinux.h output. The central point was that the > >> generated vmlinux.h does not embed a license string -- making the > >> licensing of the file non-obvious. > >> > >> This commit adds a LGPL-2.1 OR BSD-2-Clause SPDX license identifier to > >> the generated vmlinux.h output. This is line with what bpftool generates > >> in object file skeletons. > >> > >> Signed-off-by: Daniel Xu <dxu@xxxxxxxxx> > >> --- > >> tools/bpf/bpftool/btf.c | 1 + > >> 1 file changed, 1 insertion(+) > >> > >> diff --git a/tools/bpf/bpftool/btf.c b/tools/bpf/bpftool/btf.c > >> index a2c665beda87..fca810a27768 100644 > >> --- a/tools/bpf/bpftool/btf.c > >> +++ b/tools/bpf/bpftool/btf.c > >> @@ -425,6 +425,7 @@ static int dump_btf_c(const struct btf *btf, > >> if (err) > >> return err; > >> > >> + printf("/* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */\n\n"); > > > > I don't think we can add any kind of license identifier > > to the auto generated output. > > vmlinux.h is a pretty printed dwarfdump. > > Just so I understand better, when you say "I don't think we can", > do you mean: > > 1) There may be legal issues w/ adding the license identifier > 2) It doesn't make sense to add the license header > 3) Something else? 2