2021-03-17 10:54:10

by Kumar Kartikeya Dwivedi

Subject: [PATCH] libbpf: use SOCK_CLOEXEC when opening the netlink socket

Otherwise, there exists a small window between the opening and closing
of the socket fd where it may leak into processes launched by some other
thread.

Signed-off-by: Kumar Kartikeya Dwivedi <[email protected]>
---
tools/lib/bpf/netlink.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/lib/bpf/netlink.c b/tools/lib/bpf/netlink.c
index 4dd73de00..d2cb28e9e 100644
--- a/tools/lib/bpf/netlink.c
+++ b/tools/lib/bpf/netlink.c
@@ -40,7 +40,7 @@ static int libbpf_netlink_open(__u32 *nl_pid)
memset(&sa, 0, sizeof(sa));
sa.nl_family = AF_NETLINK;

- sock = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+ sock = socket(AF_NETLINK, SOCK_RAW | SOCK_CLOEXEC, NETLINK_ROUTE);
if (sock < 0)
return -errno;

--
2.30.2

2021-03-17 11:09:45

by Toke Høiland-Jørgensen

Subject: Re: [PATCH] libbpf: use SOCK_CLOEXEC when opening the netlink socket

Kumar Kartikeya Dwivedi <[email protected]> writes:

> Otherwise, there exists a small window between the opening and closing
> of the socket fd where it may leak into processes launched by some other
> thread.
>
> Signed-off-by: Kumar Kartikeya Dwivedi <[email protected]>

FYI, you should be tagging patches with the tree they are targeting. In
this case probably the 'bpf' tree (so [PATCH bpf] in the subject).

Also, a Fixes: tag would be nice here?

-Toke