On 5/5/20 10:21 PM, Andrii Nakryiko wrote:
On Sun, May 3, 2020 at 11:29 PM Yonghong Song <yhs@xxxxxx> wrote:
This patch added netlink and ipv6_route targets, using
the same seq_ops (except show() and minor changes for stop())
for /proc/net/{netlink,ipv6_route}.
The net namespace for these targets are the current net
namespace at file open stage, similar to
/proc/net/{netlink,ipv6_route} reference counting
the net namespace at seq_file open stage.
Since module is not supported for now, ipv6_route is
supported only if the IPV6 is built-in, i.e., not compiled
as a module. The restriction can be lifted once module
is properly supported for bpf_iter.
Signed-off-by: Yonghong Song <yhs@xxxxxx>
---
fs/proc/proc_net.c | 19 +++++++++
include/linux/proc_fs.h | 3 ++
net/ipv6/ip6_fib.c | 65 +++++++++++++++++++++++++++++-
net/ipv6/route.c | 27 +++++++++++++
net/netlink/af_netlink.c | 87 +++++++++++++++++++++++++++++++++++++++-
5 files changed, 197 insertions(+), 4 deletions(-)
[...]
int __init ip6_route_init(void)
{
int ret;
@@ -6455,6 +6474,14 @@ int __init ip6_route_init(void)
if (ret)
goto out_register_late_subsys;
+#if IS_BUILTIN(CONFIG_IPV6)
+#if defined(CONFIG_BPF_SYSCALL) && defined(CONFIG_PROC_FS)
+ ret = bpf_iter_register();
+ if (ret)
+ goto out_register_late_subsys;
Seems like bpf_iter infra is missing unregistering API.
ip6_route_init(), if fails, undoes all the registrations, so probably
should also unregister ipv6_route target as well?
Yes, it is. But not in this function. In this function,
bpf_iter_register() is the last one possibly causing error,
so there is no need to unregister here.
But there is another cleanup funciton called outside of this
function, I need to do proper unregister there.
Thanks for catching this.
+#endif
+#endif
+
for_each_possible_cpu(cpu) {
struct uncached_list *ul = per_cpu_ptr(&rt6_uncached_list, cpu);
[...]
+static void netlink_seq_stop(struct seq_file *seq, void *v)
+{
+ struct bpf_iter_meta meta;
+ struct bpf_prog *prog;
+
+ if (!v) {
+ meta.seq = seq;
+ prog = bpf_iter_get_info(&meta, true);
+ if (prog)
+ netlink_prog_seq_show(prog, &meta, v);
nit: netlink_prog_seq_show() can return failure (from BPF program),
but you are not returning it. Given seq_file's stop is not supposed to
fail, you can explicitly cast result to (void)? I think it's done in
Yes, we can do this. An explicit casting expressed the intention.
few other places in BPF code, when return result is explicitly
ignored.
+ }
+
+ netlink_native_seq_stop(seq, v);
+}
+#else
[...]
