| Message ID | cover.1716446893.git.tanggeliang@kylinos.cn |
|---|---|
| Headers | show |
| Series | fixes for test_sockmap | expand |
Geliang Tang wrote: > From: Geliang Tang <tanggeliang@kylinos.cn> > > The values of tx_prog_fd in run_options() should not be 0, so set it as -1 > in else branch, and test it using "if (tx_prog_fd > 0)" condition, not > "if (tx_prog_fd)" or "if (tx_prog_fd >= 0)". > > Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn> > --- Acked-by: John Fastabend <john.fastabend@gmail.com>
Geliang Tang wrote: > From: Geliang Tang <tanggeliang@kylinos.cn> > > The value of recv in msg_loop may be negative, like EWOULDBLOCK, so it's > necessary to check if it is positive before accumulating it to bytes_recvd. > > Fixes: 16962b2404ac ("bpf: sockmap, add selftests") > Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn> > --- > tools/testing/selftests/bpf/test_sockmap.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/tools/testing/selftests/bpf/test_sockmap.c b/tools/testing/selftests/bpf/test_sockmap.c > index 3654babfac59..4c7cb206b31d 100644 > --- a/tools/testing/selftests/bpf/test_sockmap.c > +++ b/tools/testing/selftests/bpf/test_sockmap.c > @@ -681,7 +681,8 @@ static int msg_loop(int fd, int iov_count, int iov_length, int cnt, > } > } > > - s->bytes_recvd += recv; > + if (recv > 0) > + s->bytes_recvd += recv; > > if (opt->check_recved_len && s->bytes_recvd > total_bytes) { > errno = EMSGSIZE; > -- > 2.43.0 > Acked-by: John Fastabend <john.fastabend@gmail.com>
Geliang Tang wrote: > From: Geliang Tang <tanggeliang@kylinos.cn> > > Switch attachments to bpf_link using bpf_program__attach_sockmap() instead > of bpf_prog_attach(). Sorry it took me a few days to get to this. Is there a reason to push this to links vs just leave it as is? I had a plan to port all the test_sockmap tests into prog_tests anyways. I'll try to push some initial patch next week. The one advantage of test_sockmap is we can have it run for longer runs by pushing different options through so might be worth keeping just for that. If you really want links here I'm OK with that I guess just asking. Thanks, John > > This patch adds a new array progs[] to replace prog_fd[] array, set in > populate_progs() for each program in bpf object. > > And another new array links[] to save the attached bpf_link. It is > initalized as NULL in populate_progs, set as the return valuses of > bpf_program__attach_sockmap(), and detached by bpf_link__detach(). > > Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn> > --- > tools/testing/selftests/bpf/test_sockmap.c | 59 ++++++++++++---------- > 1 file changed, 31 insertions(+), 28 deletions(-) > > diff --git a/tools/testing/selftests/bpf/test_sockmap.c b/tools/testing/selftests/bpf/test_sockmap.c > index e7dbf49a2ca6..d7581bbbc473 100644 > --- a/tools/testing/selftests/bpf/test_sockmap.c > +++ b/tools/testing/selftests/bpf/test_sockmap.c > @@ -64,6 +64,8 @@ int failed; > int map_fd[9]; > struct bpf_map *maps[9]; > int prog_fd[9]; > +struct bpf_program *progs[9]; > +struct bpf_link *links[9]; > > int txmsg_pass; > int txmsg_redir; > @@ -960,43 +962,39 @@ static int run_options(struct sockmap_options *options, int cg_fd, int test) > > /* Attach programs to sockmap */ > if (!txmsg_omit_skb_parser) { > - err = bpf_prog_attach(prog_fd[0], map_fd[0], > - BPF_SK_SKB_STREAM_PARSER, 0); > - if (err) { > + links[0] = bpf_program__attach_sockmap(progs[0], map_fd[0]); > + if (!links[0]) { > fprintf(stderr, > - "ERROR: bpf_prog_attach (sockmap %i->%i): %d (%s)\n", > - prog_fd[0], map_fd[0], err, strerror(errno)); > - return err; > + "ERROR: bpf_program__attach_sockmap (sockmap %i->%i): (%s)\n", > + bpf_program__fd(progs[0]), map_fd[0], strerror(errno)); > + return -1; > } > } > > - err = bpf_prog_attach(prog_fd[1], map_fd[0], > - BPF_SK_SKB_STREAM_VERDICT, 0); > - if (err) { > - fprintf(stderr, "ERROR: bpf_prog_attach (sockmap): %d (%s)\n", > - err, strerror(errno)); > - return err; > + links[1] = bpf_program__attach_sockmap(progs[1], map_fd[0]); > + if (!links[1]) { > + fprintf(stderr, "ERROR: bpf_program__attach_sockmap (sockmap): (%s)\n", > + strerror(errno)); > + return -1; > } > > /* Attach programs to TLS sockmap */ > if (txmsg_ktls_skb) { > if (!txmsg_omit_skb_parser) { > - err = bpf_prog_attach(prog_fd[0], map_fd[8], > - BPF_SK_SKB_STREAM_PARSER, 0); > - if (err) { > + links[2] = bpf_program__attach_sockmap(progs[0], map_fd[8]); > + if (!links[2]) { > fprintf(stderr, > - "ERROR: bpf_prog_attach (TLS sockmap %i->%i): %d (%s)\n", > - prog_fd[0], map_fd[8], err, strerror(errno)); > - return err; > + "ERROR: bpf_program__attach_sockmap (TLS sockmap %i->%i): (%s)\n", > + bpf_program__fd(progs[0]), map_fd[8], strerror(errno)); > + return -1; > } > } > > - err = bpf_prog_attach(prog_fd[2], map_fd[8], > - BPF_SK_SKB_STREAM_VERDICT, 0); > - if (err) { > - fprintf(stderr, "ERROR: bpf_prog_attach (TLS sockmap): %d (%s)\n", > - err, strerror(errno)); > - return err; > + links[3] = bpf_program__attach_sockmap(progs[2], map_fd[8]); > + if (!links[3]) { > + fprintf(stderr, "ERROR: bpf_program__attach_sockmap (TLS sockmap): (%s)\n", > + strerror(errno)); > + return -1; > } > } > > @@ -1281,10 +1279,11 @@ static int run_options(struct sockmap_options *options, int cg_fd, int test) > out: > /* Detatch and zero all the maps */ > bpf_prog_detach2(prog_fd[3], cg_fd, BPF_CGROUP_SOCK_OPS); > - bpf_prog_detach2(prog_fd[0], map_fd[0], BPF_SK_SKB_STREAM_PARSER); > - bpf_prog_detach2(prog_fd[1], map_fd[0], BPF_SK_SKB_STREAM_VERDICT); > - bpf_prog_detach2(prog_fd[0], map_fd[8], BPF_SK_SKB_STREAM_PARSER); > - bpf_prog_detach2(prog_fd[2], map_fd[8], BPF_SK_SKB_STREAM_VERDICT); > + > + for (i = 0; i < ARRAY_SIZE(links); i++) { > + if (links[i]) > + bpf_link__detach(links[i]); > + } > > if (tx_prog_fd > 0) > bpf_prog_detach2(tx_prog_fd, map_fd[1], BPF_SK_MSG_VERDICT); > @@ -1836,6 +1835,7 @@ static int populate_progs(char *bpf_file) > i = bpf_object__load(obj); > i = 0; > bpf_object__for_each_program(prog, obj) { > + progs[i] = prog; > prog_fd[i] = bpf_program__fd(prog); > i++; > } > @@ -1850,6 +1850,9 @@ static int populate_progs(char *bpf_file) > } > } > > + for (i = 0; i < ARRAY_SIZE(links); i++) > + links[i] = NULL; > + > return 0; > } > > -- > 2.43.0 >
On Mon, May 27, 2024 at 10:12 AM -07, John Fastabend wrote: > Geliang Tang wrote: >> From: Geliang Tang <tanggeliang@kylinos.cn> >> >> Switch attachments to bpf_link using bpf_program__attach_sockmap() instead >> of bpf_prog_attach(). > > Sorry it took me a few days to get to this. > > Is there a reason to push this to links vs just leave it as is? I had > a plan to port all the test_sockmap tests into prog_tests anyways. I'll > try to push some initial patch next week. > > The one advantage of test_sockmap is we can have it run for longer > runs by pushing different options through so might be worth keeping > just for that. > > If you really want links here I'm OK with that I guess just asking. It was me who suggested the switch to bpf_link in reaction to a series of cleanups to prog_type and prog_attach_type submitted by Geliang. Relevant threads: https://lore.kernel.org/bpf/9c10d9f974f07fcb354a43a8eca67acb2fafc587.1715926605.git.tanggeliang@kylinos.cn https://lore.kernel.org/bpf/20240522080936.2475833-1-jakub@cloudflare.com https://lore.kernel.org/bpf/e27d7d0c1e0e79b0acd22ac6ad5d8f9f00225303.1716372485.git.tanggeliang@kylinos.cn I thought bpf_links added more value than cleaning up "old style" attachments.
On Thu, May 23, 2024 at 02:49 PM +08, Geliang Tang wrote: > From: Geliang Tang <tanggeliang@kylinos.cn> > > This patchset contains some fixes and improvements for test_sockmap. > > 3-5: switching attachments to bpf_link as Jakub suggested in [1]. > 1-2, 6-8: Small fixes. > > [1] > https://lore.kernel.org/bpf/87zfsiw3a3.fsf@cloudflare.com/ > > Geliang Tang (8): > selftests/bpf: Fix tx_prog_fd values in test_sockmap > selftests/bpf: Drop duplicate definition of i in test_sockmap > selftests/bpf: Use bpf_link attachments in test_sockmap > selftests/bpf: Replace tx_prog_fd with tx_prog in test_sockmap > selftests/bpf: Drop prog_fd array in test_sockmap > selftests/bpf: Fix size of map_fd in test_sockmap > selftests/bpf: Check length of recv in test_sockmap > selftests/bpf: Drop duplicate bpf_map_lookup_elem in test_sockmap > > .../selftests/bpf/progs/test_sockmap_kern.h | 3 - > tools/testing/selftests/bpf/test_sockmap.c | 101 +++++++++--------- > 2 files changed, 51 insertions(+), 53 deletions(-) Tested-by: Jakub Sitnicki <jakub@cloudflare.com>
On Thu, May 30, 2024 at 04:45 PM -07, John Fastabend wrote: > Geliang Tang wrote: >> On Mon, 2024-05-27 at 21:36 +0200, Jakub Sitnicki wrote: >> > On Mon, May 27, 2024 at 10:12 AM -07, John Fastabend wrote: >> > > Geliang Tang wrote: [...] >> > > The one advantage of test_sockmap is we can have it run for longer >> > > runs by pushing different options through so might be worth keeping >> > > just for that. >> > > >> > > If you really want links here I'm OK with that I guess just asking. >> > >> > It was me who suggested the switch to bpf_link in reaction to a >> > series >> > of cleanups to prog_type and prog_attach_type submitted by Geliang. >> >> Yes, patches 3-5 address Jakub's suggestion: switching attachments to >> bpf_link. > > OK. Lets just take them the series lgtm. Jakub any other comments? Gave it a run - all looks well. Thanks for the patches. Geliang, is there some MPTCP+sockmap use-case you're working towards?
On Fri, May 31, 2024 at 01:13:39PM +0200, Jakub Sitnicki wrote: > On Thu, May 30, 2024 at 04:45 PM -07, John Fastabend wrote: > > Geliang Tang wrote: > >> On Mon, 2024-05-27 at 21:36 +0200, Jakub Sitnicki wrote: > >> > On Mon, May 27, 2024 at 10:12 AM -07, John Fastabend wrote: > >> > > Geliang Tang wrote: > > [...] > > >> > > The one advantage of test_sockmap is we can have it run for longer > >> > > runs by pushing different options through so might be worth keeping > >> > > just for that. > >> > > > >> > > If you really want links here I'm OK with that I guess just asking. > >> > > >> > It was me who suggested the switch to bpf_link in reaction to a > >> > series > >> > of cleanups to prog_type and prog_attach_type submitted by Geliang. > >> > >> Yes, patches 3-5 address Jakub's suggestion: switching attachments to > >> bpf_link. > > > > OK. Lets just take them the series lgtm. Jakub any other comments? > > Gave it a run - all looks well. Thanks for the patches. > > Geliang, is there some MPTCP+sockmap use-case you're working towards? Yes, indeed. I have been working on a task related to MPTCP+sockmap recently, at least related to this test_sockmap.c selftest. We recently received an issue with MPTCP [1], that is TLS cannot be set on MPTCP sockets. The reason is that both MPTCP and TLS are implemented on TCP ULP. And each socket only supports one type of TCP ULP. I simply modified this test_sockmap.c selftest to support MPTCP, so that it can be used as the first version of test for MPTCP+TLS. So I spent some time reading and debugging this test. The development of MPTCP+TLS is still ongoing, and currently only setsockopt part has been successfully supported. The idea is simple, use an array of tcp_ulp_ops in a socket, instead of a single one: struct inet_connection_sock { ... ... const struct tcp_ulp_ops *icsk_ulp_ops[ULP_INDEX_MAX]; void __rcu *icsk_ulp_data[ULP_INDEX_MAX]; } The entire patch is in my commit "mptcp: tls support" [2]. It's not finish yet, but I really want to hear your opinions, especially John's. [1] https://github.com/multipath-tcp/mptcp_net-next/issues/480 [2] https://github.com/geliangtang/mptcp_net-next/commit/bba00a6cde75bab5a2c1c196d49812b4ed6addb0 Thanks, -Geliang
Hello: This series was applied to bpf/bpf-next.git (master) by Daniel Borkmann <daniel@iogearbox.net>: On Thu, 23 May 2024 14:49:56 +0800 you wrote: > From: Geliang Tang <tanggeliang@kylinos.cn> > > This patchset contains some fixes and improvements for test_sockmap. > > 3-5: switching attachments to bpf_link as Jakub suggested in [1]. > 1-2, 6-8: Small fixes. > > [...] Here is the summary with links: - [bpf-next,1/8] selftests/bpf: Fix tx_prog_fd values in test_sockmap https://git.kernel.org/bpf/bpf-next/c/d95ba15b9784 - [bpf-next,2/8] selftests/bpf: Drop duplicate definition of i in test_sockmap https://git.kernel.org/bpf/bpf-next/c/a9f0ea175948 - [bpf-next,3/8] selftests/bpf: Use bpf_link attachments in test_sockmap https://git.kernel.org/bpf/bpf-next/c/3f32a115f61d - [bpf-next,4/8] selftests/bpf: Replace tx_prog_fd with tx_prog in test_sockmap https://git.kernel.org/bpf/bpf-next/c/24bb90a42633 - [bpf-next,5/8] selftests/bpf: Drop prog_fd array in test_sockmap https://git.kernel.org/bpf/bpf-next/c/467a0c79b551 - [bpf-next,6/8] selftests/bpf: Fix size of map_fd in test_sockmap https://git.kernel.org/bpf/bpf-next/c/dcb681b659f2 - [bpf-next,7/8] selftests/bpf: Check length of recv in test_sockmap https://git.kernel.org/bpf/bpf-next/c/de1b5ea789dc - [bpf-next,8/8] selftests/bpf: Drop duplicate bpf_map_lookup_elem in test_sockmap https://git.kernel.org/bpf/bpf-next/c/49784c797932 You are awesome, thank you!
From: Geliang Tang <tanggeliang@kylinos.cn> This patchset contains some fixes and improvements for test_sockmap. 3-5: switching attachments to bpf_link as Jakub suggested in [1]. 1-2, 6-8: Small fixes. [1] https://lore.kernel.org/bpf/87zfsiw3a3.fsf@cloudflare.com/ Geliang Tang (8): selftests/bpf: Fix tx_prog_fd values in test_sockmap selftests/bpf: Drop duplicate definition of i in test_sockmap selftests/bpf: Use bpf_link attachments in test_sockmap selftests/bpf: Replace tx_prog_fd with tx_prog in test_sockmap selftests/bpf: Drop prog_fd array in test_sockmap selftests/bpf: Fix size of map_fd in test_sockmap selftests/bpf: Check length of recv in test_sockmap selftests/bpf: Drop duplicate bpf_map_lookup_elem in test_sockmap .../selftests/bpf/progs/test_sockmap_kern.h | 3 - tools/testing/selftests/bpf/test_sockmap.c | 101 +++++++++--------- 2 files changed, 51 insertions(+), 53 deletions(-)