From: Stanislav Fomichev <sdf@fomichev.me>
To: Daniel Borkmann <daniel@iogearbox.net>
Cc: Stanislav Fomichev <sdf@google.com>,
netdev@vger.kernel.org, ast@kernel.org, davem@davemloft.net,
ecree@solarflare.com, quentin.monnet@netronome.com
Subject: Re: [PATCH bpf-next v2 4/6] selftests/bpf: skip verifier tests for unsupported map types
Date: Thu, 20 Dec 2018 14:51:46 -0800 [thread overview]
Message-ID: <20181220225146.GA2959@mini-arch.hsd1.ca.comcast.net> (raw)
In-Reply-To: <5f82a125-03ee-dddb-4a29-a29f08dcc394@iogearbox.net>
On 12/20, Daniel Borkmann wrote:
> On 12/20/2018 09:51 PM, Stanislav Fomichev wrote:
> > On Tue, Dec 18, 2018 at 4:02 PM Stanislav Fomichev <sdf@fomichev.me> wrote:
> >> On 12/19, Daniel Borkmann wrote:
> >>> On 12/17/2018 07:25 PM, Stanislav Fomichev wrote:
> >>>> Use recently introduced bpf_map_type_supported() to skip tests in the
> >>>> test_verifier if map creation (create_map) fails. It's handled
> >>>> explicitly for each fixup, i.e. if bpf_create_map returns negative fd,
> >>>> we probe the kernel for the appropriate map support and skip the
> >>>> test is map type is not supported.
> >>>>
> >>>> Signed-off-by: Stanislav Fomichev <sdf@google.com>
> >>>> ---
> >>>> tools/testing/selftests/bpf/test_verifier.c | 34 +++++++++++++++++++--
> >>>> 1 file changed, 31 insertions(+), 3 deletions(-)
> >>>>
> >>>> diff --git a/tools/testing/selftests/bpf/test_verifier.c b/tools/testing/selftests/bpf/test_verifier.c
> >>>> index 124d21306c27..d267f5248b5d 100644
> >>>> --- a/tools/testing/selftests/bpf/test_verifier.c
> >>>> +++ b/tools/testing/selftests/bpf/test_verifier.c
> >>>> @@ -14221,10 +14221,20 @@ static int create_cgroup_storage(bool percpu)
> >>>> return fd;
> >>>> }
> >>>>
> >>>> +static bool skip_unsupported_map(int ret, enum bpf_map_type map_type)
> >>>> +{
> >>>> + if (ret < 0 && !bpf_map_type_supported(map_type)) {
> >>>> + printf("SKIP (unsupported map type %d)\n", map_type);
> >>>> + skips++;
> >>>> + return true;
> >>>> + }
> >>>> + return false;
> >>>> +}
> >>>> +
> >>>> static char bpf_vlog[UINT_MAX >> 8];
> >>>>
> >>>> -static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> - struct bpf_insn *prog, int *map_fds)
> >>>> +static int do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> + struct bpf_insn *prog, int *map_fds)
> >>>> {
> >>>> int *fixup_map_hash_8b = test->fixup_map_hash_8b;
> >>>> int *fixup_map_hash_48b = test->fixup_map_hash_48b;
> >>>> @@ -14309,6 +14319,9 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>>
> >>>> if (*fixup_cgroup_storage) {
> >>>> map_fds[7] = create_cgroup_storage(false);
> >>>> + if (skip_unsupported_map(map_fds[7],
> >>>> + BPF_MAP_TYPE_CGROUP_STORAGE))
> >>>> + return -1;
> >>>> do {
> >>>> prog[*fixup_cgroup_storage].imm = map_fds[7];
> >>>> fixup_cgroup_storage++;
> >>>> @@ -14317,6 +14330,9 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>>
> >>>> if (*fixup_percpu_cgroup_storage) {
> >>>> map_fds[8] = create_cgroup_storage(true);
> >>>> + if (skip_unsupported_map(map_fds[8],
> >>>> + BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE))
> >>>> + return -1;
> >>>> do {
> >>>> prog[*fixup_percpu_cgroup_storage].imm = map_fds[8];
> >>>> fixup_percpu_cgroup_storage++;
> >>>> @@ -14325,6 +14341,8 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> if (*fixup_map_sockmap) {
> >>>> map_fds[9] = create_map(BPF_MAP_TYPE_SOCKMAP, sizeof(int),
> >>>> sizeof(int), 1);
> >>>> + if (skip_unsupported_map(map_fds[9], BPF_MAP_TYPE_SOCKMAP))
> >>>> + return -1;
> >>>> do {
> >>>> prog[*fixup_map_sockmap].imm = map_fds[9];
> >>>> fixup_map_sockmap++;
> >>>> @@ -14333,6 +14351,8 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> if (*fixup_map_sockhash) {
> >>>> map_fds[10] = create_map(BPF_MAP_TYPE_SOCKHASH, sizeof(int),
> >>>> sizeof(int), 1);
> >>>> + if (skip_unsupported_map(map_fds[10], BPF_MAP_TYPE_SOCKHASH))
> >>>> + return -1;
> >>>> do {
> >>>> prog[*fixup_map_sockhash].imm = map_fds[10];
> >>>> fixup_map_sockhash++;
> >>>> @@ -14341,6 +14361,8 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> if (*fixup_map_xskmap) {
> >>>> map_fds[11] = create_map(BPF_MAP_TYPE_XSKMAP, sizeof(int),
> >>>> sizeof(int), 1);
> >>>> + if (skip_unsupported_map(map_fds[11], BPF_MAP_TYPE_XSKMAP))
> >>>> + return -1;
> >>>> do {
> >>>> prog[*fixup_map_xskmap].imm = map_fds[11];
> >>>> fixup_map_xskmap++;
> >>>> @@ -14349,11 +14371,16 @@ static void do_test_fixup(struct bpf_test *test, enum bpf_map_type prog_type,
> >>>> if (*fixup_map_stacktrace) {
> >>>> map_fds[12] = create_map(BPF_MAP_TYPE_STACK_TRACE, sizeof(u32),
> >>>> sizeof(u64), 1);
> >>>> + if (skip_unsupported_map(map_fds[12],
> >>>> + BPF_MAP_TYPE_STACK_TRACE))
> >>>> + return -1;
> >>>
> >>> Nit: Could probably be slightly simplified by moving this into and by reworking
> >>> create_{map,cgroup_storage}() a bit.
> >> Yeah, I stated that option in the cover letter. I did that initially,
> >> but it required some additional argument (skip/supported) to the
> >> create_{map,cgroup_storage} and I scrapped this approach due to too
> >> much plumbing.
> >>
> >> But I think since we are not doing any parallel tests in the verifier,
> >> we can do something like the following patch below. WDYT?
> > Daniel, should this go as is or you'd like me to respin to the version
> > from my last reply (or something similar)?
>
> Your diff on top of the set looks good to me. My preference though is that
> we get both your work and Quentin's merged in the next cycle (given we're
> about to close bpf-next) and have plenty of time to consolidate the two and
> get it into good shape in order to then move the logic into libbpf as next
> step such that bpftool /and/ kselftests can make use of it.
Ack, I'll use that for a v3 submission when bpf-next opens again.
Maybe I can even drop my custom probes if Quentin's patch set is ready
by that point.
>
> Thanks a lot,
> Daniel
next prev parent reply other threads:[~2018-12-20 22:51 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-17 18:25 [PATCH bpf-next v2 0/6] skip verifier/map tests if kernel support is missing Stanislav Fomichev
2018-12-17 18:25 ` [PATCH bpf-next v2 1/6] selftests/bpf: add map/prog type probe helpers Stanislav Fomichev
2018-12-18 23:21 ` Daniel Borkmann
2018-12-18 23:35 ` Stanislav Fomichev
2018-12-17 18:25 ` [PATCH bpf-next v2 2/6] selftests/bpf: skip sockmap in test_maps if kernel doesn't have support Stanislav Fomichev
2018-12-17 18:25 ` [PATCH bpf-next v2 3/6] selftests/bpf: skip verifier tests for unsupported program types Stanislav Fomichev
2018-12-17 18:25 ` [PATCH bpf-next v2 4/6] selftests/bpf: skip verifier tests for unsupported map types Stanislav Fomichev
2018-12-18 23:25 ` Daniel Borkmann
2018-12-19 0:02 ` Stanislav Fomichev
2018-12-20 20:51 ` Stanislav Fomichev
2018-12-20 22:38 ` Daniel Borkmann
2018-12-20 22:51 ` Stanislav Fomichev [this message]
2018-12-17 18:25 ` [PATCH bpf-next v2 5/6] selftests/bpf: mark verifier test that uses bpf_trace_printk as BPF_PROG_TYPE_TRACEPOINT Stanislav Fomichev
2018-12-17 18:25 ` [PATCH bpf-next v2 6/6] bpf: BPF_PROG_TYPE_CGROUP_{SKB,SOCK,SOCK_ADDR} require cgroups enabled Stanislav Fomichev
2018-12-18 21:25 ` [PATCH bpf-next v2 0/6] skip verifier/map tests if kernel support is missing Alexei Starovoitov
2018-12-18 21:30 ` Stanislav Fomichev
2018-12-18 23:18 ` Daniel Borkmann
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181220225146.GA2959@mini-arch.hsd1.ca.comcast.net \
--to=sdf@fomichev.me \
--cc=ast@kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=ecree@solarflare.com \
--cc=netdev@vger.kernel.org \
--cc=quentin.monnet@netronome.com \
--cc=sdf@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).