January 21, 2026 at 20:55, "Jiayuan Chen" <[email protected] mailto:[email protected]?to=%22Jiayuan%20Chen%22%20%3Cjiayuan.chen%40linux.dev%3E > wrote:
> > January 21, 2026 at 17:36, "Jakub Sitnicki" <[email protected] > mailto:[email protected]?to=%22Jakub%20Sitnicki%22%20%3Cjakub%40cloudflare.com%3E > > wrote: > > > > > On Tue, Jan 20, 2026 at 04:00 PM +01, Jakub Sitnicki wrote: > > > > > > On Tue, Jan 13, 2026 at 10:50 AM +08, Jiayuan Chen wrote: > > > > [...] > > > > > > > > > > diff --git a/net/ipv4/udp_bpf.c b/net/ipv4/udp_bpf.c > > > index 0735d820e413..91233e37cd97 100644 > > > --- a/net/ipv4/udp_bpf.c > > > +++ b/net/ipv4/udp_bpf.c > > > @@ -5,6 +5,7 @@ > > > #include <net/sock.h> > > > #include <net/udp.h> > > > #include <net/inet_common.h> > > > +#include <asm/ioctls.h> > > > > > > #include "udp_impl.h" > > > > > > @@ -111,12 +112,26 @@ enum { > > > static DEFINE_SPINLOCK(udpv6_prot_lock); > > > static struct proto udp_bpf_prots[UDP_BPF_NUM_PROTS]; > > > > > > +static int udp_bpf_ioctl(struct sock *sk, int cmd, int *karg) > > > +{ > > > + if (cmd != SIOCINQ) > > > + return udp_ioctl(sk, cmd, karg); > > > + > > > + /* Since we don't hold a lock, sk_receive_queue may contain data. > > > + * BPF might only be processing this data at the moment. We only > > > + * care about the data in the ingress_msg here. > > > + */ > > > > > I think we should strive for a design where FIONREAD does not go down > > after you add your socket to sockmap, if there was no recvmsg call in > > between. To show what I mean, I added this test that's currently failing > > for udp: > > > > ---8<--- > > diff --git a/tools/testing/selftests/bpf/prog_tests/sockmap_basic.c > > b/tools/testing/selftests/bpf/prog_tests/sockmap_basic.c > > index 1f1289f5a8c2..123c96fcaef0 100644 > > --- a/tools/testing/selftests/bpf/prog_tests/sockmap_basic.c > > +++ b/tools/testing/selftests/bpf/prog_tests/sockmap_basic.c > > @@ -1229,6 +1229,66 @@ static void test_sockmap_copied_seq(bool strp) > > test_sockmap_pass_prog__destroy(skel); > > } > > > > +/* Test FIONREAD when data exists in sk_receive_queue before sockmap > > insertion */ > > +static void test_sockmap_fionread_pre_insert(int sotype) > > +{ > > + int map, err, sent, recvd, zero = 0, avail = 0; > > + struct test_sockmap_pass_prog *skel = NULL; > > + int c = -1, p = -1; > > + char buf[10] = "0123456789", rcv[11]; > > + struct bpf_program *prog; > > + > > + skel = test_sockmap_pass_prog__open_and_load(); > > + if (!ASSERT_OK_PTR(skel, "open_and_load")) > > + return; > > + > > + prog = skel->progs.prog_skb_verdict; > > + map = bpf_map__fd(skel->maps.sock_map_rx); > > + > > + err = bpf_prog_attach(bpf_program__fd(prog), map, > > BPF_SK_SKB_STREAM_VERDICT, 0); > > + if (!ASSERT_OK(err, "bpf_prog_attach verdict")) > > + goto end; > > + > > + err = create_pair(AF_INET, sotype, &c, &p); > > + if (!ASSERT_OK(err, "create_pair")) > > + goto end; > > + > > + /* Step 1: Send data BEFORE sockmap insertion - lands in > > sk_receive_queue */ > > + sent = xsend(p, buf, sizeof(buf), 0); > > + if (!ASSERT_EQ(sent, sizeof(buf), "xsend pre-insert")) > > + goto end; > > + > > + /* Step 2: Verify FIONREAD reports data in sk_receive_queue */ > > + err = poll_read(c, IO_TIMEOUT_SEC); > > + if (!ASSERT_OK(err, "poll_read pre-insert")) > > + goto end; > > + err = ioctl(c, FIONREAD, &avail); > > + ASSERT_OK(err, "ioctl(FIONREAD) pre-insert error"); > > + ASSERT_EQ(avail, sizeof(buf), "ioctl(FIONREAD) pre-insert"); > > + > > + /* Step 3: Insert socket into sockmap */ > > + err = bpf_map_update_elem(map, &zero, &c, BPF_ANY); > > + if (!ASSERT_OK(err, "bpf_map_update_elem(c)")) > > + goto end; > > + > > + /* Step 4: FIONREAD should still report the data in sk_receive_queue */ > > + err = ioctl(c, FIONREAD, &avail); > > + ASSERT_OK(err, "ioctl(FIONREAD) post-insert error"); > > + ASSERT_EQ(avail, sizeof(buf), "ioctl(FIONREAD) post-insert"); > > + > > + /* Verify we can still read the data */ > > + recvd = recv_timeout(c, rcv, sizeof(rcv), MSG_DONTWAIT, IO_TIMEOUT_SEC); > > + ASSERT_EQ(recvd, sizeof(buf), "recv post-insert"); > > + ASSERT_OK(memcmp(buf, rcv, recvd), "data mismatch"); > > + > > +end: > > + if (c >= 0) > > + close(c); > > + if (p >= 0) > > + close(p); > > + test_sockmap_pass_prog__destroy(skel); > > +} > > + > > /* it is used to send data to via native stack and BPF redirecting */ > > static void test_sockmap_multi_channels(int sotype) > > { > > @@ -1373,4 +1433,8 @@ void test_sockmap_basic(void) > > test_sockmap_multi_channels(SOCK_STREAM); > > if (test__start_subtest("sockmap udp multi channels")) > > test_sockmap_multi_channels(SOCK_DGRAM); > > + if (test__start_subtest("sockmap tcp fionread pre-insert")) > > + test_sockmap_fionread_pre_insert(SOCK_STREAM); > > + if (test__start_subtest("sockmap udp fionread pre-insert")) > > + test_sockmap_fionread_pre_insert(SOCK_DGRAM); > > } > > --->8--- > > > > > > > > + *karg = sk_msg_first_len(sk); > > > + return 0; > > > +} > > > + > > > > > > > I've been thinking about this some more and came to the conclusion that > > this udp_bpf_ioctl implementation is actually what we want, while > > tcp_bpf_ioctl *should not* be checking if the sk_receive_queue is > > non-empty. > > > > Why? Because the verdict prog might redirect or drop the skbs from > > sk_receive_queue once it actually runs. The messages might never appear > > on the msg_ingress queue. > > > > What I think we should be doing, in the end, is kicking the > > sk_receive_queue processing on bpf_map_update_elem, if there's data > > ready. > > > > The API semantics I'm proposing is: > > > > 1. ioctl(FIONREAD) -> reports N bytes > > 2. bpf_map_update_elem(sk) -> socket inserted into sockmap > > 3. poll() for POLLIN -> wait for socket to be ready to read > > 5. ioctl(FIONREAD) -> report N bytes if verdict prog didn't > > redirect or drop it > > > > We don't have to add the the queue kick on map update in this series. > > > > If you decide to leave it for later, can I ask that you open an issue at > > our GH project [1]? > > > > I don't want it to fall through the cracks. And I sometimes have people > > asking what they could help with in sockmap. > > > > Thanks, > > -jkbs > > > > [1] https://github.com/sockmap-project/sockmap-project/issues > > > Hi Jakub, > > Thanks for taking the time to think through this carefully. I agree with your > analysis - reporting sk_receive_queue length is misleading since the verdict > prog might redirect or drop those skbs. > > There's no rush to merge this patch. > > Since the kick queue on bpf_map_update_elem addresses a closely related issue, > I think it makes sense to include it in this patchset for easier tracking > rather > than splitting it out. > > I'll spend more time looking into this and come back with an updated version. > > Thanks, > Jiayuan > Hi Jakub, I've been thinking about this more, and I realize the problem is not as simple as it seems. Regarding kicking the sk_receive_queue on bpf_map_update_elem: the BPF program may not be fully initialized at that point. For example, with a redirect program, the destination fd might not yet be inserted into the map. If we kick the data through the BPF program immediately, the redirect lookup would fail, leading to unexpected behavior (data being dropped or passed to the wrong socket). I also considered triggering the kick in poll/select via sk_msg_is_readable(). However, this approach doesn't work for TCP because tcp_poll() -> tcp_stream_is_readable() -> tcp_epollin_ready() will return early when sk_receive_queue has data, before ever calling sk_is_readable(). In the next version, I'll address your other nits and remove the sk_receive_queue check from tcp_bpf_ioctl. I'll also open an issue on the GH project to track this problem so we can continue exploring better solutions. Thanks, Jiayuan
