| Message ID | 20200123165934.9584-3-lmb@cloudflare.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [bpf,1/4] selftests: bpf: use a temporary file in test_sockmap | expand |
On Thu, Jan 23, 2020 at 04:59:31PM +0000, Lorenz Bauer wrote: > The reuseport tests currently suffer from a race condition: RST > packets count towards DROP_ERR_SKB_DATA, since they don't contain > a valid struct cmd. Tests will spuriously fail depending on whether > check_results is called before or after the RST is processed. > > Exit the BPF program early if FIN is set. Make sense. Is it a RST or FIN? The earlier commit message said RST.
On Thu, 23 Jan 2020 at 17:16, Martin Lau <kafai@fb.com> wrote: > > On Thu, Jan 23, 2020 at 04:59:31PM +0000, Lorenz Bauer wrote: > > The reuseport tests currently suffer from a race condition: RST > > packets count towards DROP_ERR_SKB_DATA, since they don't contain > > a valid struct cmd. Tests will spuriously fail depending on whether > > check_results is called before or after the RST is processed. > > > > Exit the BPF program early if FIN is set. > Make sense. > Is it a RST or FIN? The earlier commit message said RST. FIN, sorry. I'll update in a follow up.
On Thu, Jan 23, 2020 at 04:59:31PM +0000, Lorenz Bauer wrote: > The reuseport tests currently suffer from a race condition: RST > packets count towards DROP_ERR_SKB_DATA, since they don't contain > a valid struct cmd. Tests will spuriously fail depending on whether > check_results is called before or after the RST is processed. > > Exit the BPF program early if FIN is set. btw, it needs a Fixes tag. Patch 4 and Patch 1 also need a Fixes tag.
On Thu, 23 Jan 2020 at 21:54, Martin Lau <kafai@fb.com> wrote: > > btw, it needs a Fixes tag. > > Patch 4 and Patch 1 also need a Fixes tag. This makes me wonder, should these go via bpf or bpf-next? Do I have to split the series then?
On 1/24/20 10:00 AM, Lorenz Bauer wrote: > On Thu, 23 Jan 2020 at 21:54, Martin Lau <kafai@fb.com> wrote: >> >> btw, it needs a Fixes tag. >> >> Patch 4 and Patch 1 also need a Fixes tag. > > This makes me wonder, should these go via bpf or bpf-next? Do I have > to split the series then? Lets do all of these for bpf-next since timing is very close before v5.5 release. If needed, we can later have them picked up for 5.5 stable. Thanks, Daniel
Lorenz Bauer wrote: > The reuseport tests currently suffer from a race condition: RST > packets count towards DROP_ERR_SKB_DATA, since they don't contain > a valid struct cmd. Tests will spuriously fail depending on whether > check_results is called before or after the RST is processed. > > Exit the BPF program early if FIN is set. > > Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> > --- > .../selftests/bpf/progs/test_select_reuseport_kern.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c > index d69a1f2bbbfd..26e77dcc7e91 100644 > --- a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c > +++ b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c > @@ -113,6 +113,12 @@ int _select_by_skb_data(struct sk_reuseport_md *reuse_md) > data_check.skb_ports[0] = th->source; > data_check.skb_ports[1] = th->dest; > > + if (th->fin) > + /* The connection is being torn down at the end of a > + * test. It can't contain a cmd, so return early. > + */ > + return SK_PASS; > + > if ((th->doff << 2) + sizeof(*cmd) > data_check.len) > GOTO_DONE(DROP_ERR_SKB_DATA); > if (bpf_skb_load_bytes(reuse_md, th->doff << 2, &cmd_copy, > -- > 2.20.1 > Acked-by: John Fastabend <john.fastabend@gmail.com>
diff --git a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c index d69a1f2bbbfd..26e77dcc7e91 100644 --- a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c +++ b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c @@ -113,6 +113,12 @@ int _select_by_skb_data(struct sk_reuseport_md *reuse_md) data_check.skb_ports[0] = th->source; data_check.skb_ports[1] = th->dest; + if (th->fin) + /* The connection is being torn down at the end of a + * test. It can't contain a cmd, so return early. + */ + return SK_PASS; + if ((th->doff << 2) + sizeof(*cmd) > data_check.len) GOTO_DONE(DROP_ERR_SKB_DATA); if (bpf_skb_load_bytes(reuse_md, th->doff << 2, &cmd_copy,
The reuseport tests currently suffer from a race condition: RST packets count towards DROP_ERR_SKB_DATA, since they don't contain a valid struct cmd. Tests will spuriously fail depending on whether check_results is called before or after the RST is processed. Exit the BPF program early if FIN is set. Signed-off-by: Lorenz Bauer <lmb@cloudflare.com> --- .../selftests/bpf/progs/test_select_reuseport_kern.c | 6 ++++++ 1 file changed, 6 insertions(+)