Re: [PATCH v3 bpf-next 3/3] selftests/bpf: Ensure cgroup/connect{4,6} programs can bind unpriv ICMP ping

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Sep 8, 2022 at 4:32 PM Martin KaFai Lau <martin.lau@xxxxxxxxx> wrote:
>
> On 9/8/22 4:16 PM, YiFei Zhu wrote:
> > +void test_connect_ping(void)
> > +{
> > +     struct connect_ping *skel;
> > +     int cgroup_fd;
> > +
> > +     if (!ASSERT_OK(unshare(CLONE_NEWNET | CLONE_NEWNS), "unshare"))
> > +             return;
> > +
> > +     /* overmount sysfs, and making original sysfs private so overmount
> > +      * does not propagate to other mntns.
> > +      */
> > +     if (!ASSERT_OK(mount("none", "/sys", NULL, MS_PRIVATE, NULL),
> > +                    "remount-private-sys"))
> > +             return;
> > +     if (!ASSERT_OK(mount("sysfs", "/sys", "sysfs", 0, NULL),
> > +                    "mount-sys"))
> > +             return;
> > +     if (!ASSERT_OK(mount("bpffs", "/sys/fs/bpf", "bpf", 0, NULL),
> > +                    "mount-bpf"))
> > +             goto clean_mount;
> > +
> > +     if (!ASSERT_OK(system("ip link set dev lo up"), "lo-up"))
> > +             goto clean_mount;
> > +     if (!ASSERT_OK(system("ip addr add 1.1.1.1 dev lo"), "lo-addr-v4"))
> > +             goto clean_mount;
> > +     if (!ASSERT_OK(system("ip -6 addr add 2001:db8::1 dev lo"), "lo-addr-v6"))
> > +             goto clean_mount;
> > +     if (write_sysctl("/proc/sys/net/ipv4/ping_group_range", "0 0"))
> > +             goto clean_mount;
> > +
> > +     cgroup_fd = test__join_cgroup("/connect_ping");
> > +     if (!ASSERT_GE(cgroup_fd, 0, "cg-create"))
> > +             goto clean_mount;
> > +
> > +     skel = connect_ping__open_and_load();
> > +     if (!ASSERT_OK_PTR(skel, "skel-load"))
> > +             goto close_cgroup;
> > +     skel->links.connect_v4_prog =
> > +             bpf_program__attach_cgroup(skel->progs.connect_v4_prog, cgroup_fd);
> > +     if (!ASSERT_OK_PTR(skel->links.connect_v4_prog, "cg-attach-v4"))
> > +             goto close_cgroup;
>
> connect_ping__destroy() is also needed in this error path.
>
> I had already mentioned that in v2.  I went back to v2 and noticed my
> editor some how merged my reply with the previous quoted line ">".  A
> similar thing happened in my recent replies also.  I hope it appears
> fine this time.

Oops, I missed that and thought that was part of a quote. Let me fix
that and send a v4.

> > +     skel->links.connect_v6_prog =
> > +             bpf_program__attach_cgroup(skel->progs.connect_v6_prog, cgroup_fd);
> > +     if (!ASSERT_OK_PTR(skel->links.connect_v6_prog, "cg-attach-v6"))
> > +             goto close_cgroup;
>
> Same here.



[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux