Re: [PATCH bpf-next v2 3/6] libbpf: Propagate EPERM to caller on program load

From: Alexei Starovoitov <alexei.starovoitov@gmail.com>
To: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Cc: "Toke Høiland-Jørgensen" <toke@redhat.com>,
	"Daniel Borkmann" <daniel@iogearbox.net>,
	"Alexei Starovoitov" <ast@kernel.org>,
	"Martin KaFai Lau" <kafai@fb.com>,
	"Song Liu" <songliubraving@fb.com>, "Yonghong Song" <yhs@fb.com>,
	"Jesper Dangaard Brouer" <brouer@redhat.com>,
	"David Miller" <davem@davemloft.net>,
	Networking <netdev@vger.kernel.org>, bpf <bpf@vger.kernel.org>
Subject: Re: [PATCH bpf-next v2 3/6] libbpf: Propagate EPERM to caller on program load
Date: Fri, 8 Nov 2019 15:17:58 -0800	[thread overview]
Message-ID: <20191108231757.7egzqebli6gcplfq@ast-mbp.dhcp.thefacebook.com> (raw)
In-Reply-To: <CAEf4BzZxcvhZG-FHF+0iqia72q3YA0dCgsgFchibiW7dkFQm2A@mail.gmail.com>

On Fri, Nov 08, 2019 at 02:50:43PM -0800, Andrii Nakryiko wrote:
> On Fri, Nov 8, 2019 at 1:33 PM Toke Høiland-Jørgensen <toke@redhat.com> wrote:
> >
> > From: Toke Høiland-Jørgensen <toke@redhat.com>
> >
> > When loading an eBPF program, libbpf overrides the return code for EPERM
> > errors instead of returning it to the caller. This makes it hard to figure
> > out what went wrong on load.
> >
> > In particular, EPERM is returned when the system rlimit is too low to lock
> > the memory required for the BPF program. Previously, this was somewhat
> > obscured because the rlimit error would be hit on map creation (which does
> > return it correctly). However, since maps can now be reused, object load
> > can proceed all the way to loading programs without hitting the error;
> > propagating it even in this case makes it possible for the caller to react
> > appropriately (and, e.g., attempt to raise the rlimit before retrying).
> >
> > Acked-by: Song Liu <songliubraving@fb.com>
> > Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
> > ---
> >  tools/lib/bpf/libbpf.c |    4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> > index cea61b2ec9d3..582c0fd16697 100644
> > --- a/tools/lib/bpf/libbpf.c
> > +++ b/tools/lib/bpf/libbpf.c
> > @@ -3721,7 +3721,7 @@ load_program(struct bpf_program *prog, struct bpf_insn *insns, int insns_cnt,
> >                 free(log_buf);
> >                 goto retry_load;
> >         }
> > -       ret = -LIBBPF_ERRNO__LOAD;
> > +       ret = (errno == EPERM) ? -errno : -LIBBPF_ERRNO__LOAD;

ouch. so libbpf was supressing all errnos for loading and that was a commit
from 2015. No wonder it's hard to debug. I grepped every where I could and it
doesn't look like anyone is using this code. There are other codes that can
come from sys_bpf(prog_load). Not sure why such decision was made back then. I
guess noone was really paying attention. I think we better propagate all codes.
I don't see why EPERM should be special.