diff options
author | Jon Maxwell <jmaxwell37@gmail.com> | 2019-04-23 09:42:13 +1000 |
---|---|---|
committer | Jiri Pirko <jiri@mellanox.com> | 2019-05-02 08:42:05 +0200 |
commit | 21bdbb95a70e182dc7d3d8cafc04c52090c34ce3 (patch) | |
tree | 77c0acb9f8c06919a287347b80b7481c0bda1572 | |
parent | 3adfafe8cec5254c7be9e08b83a4db140855ccb5 (diff) | |
download | libteam-21bdbb95a70e182dc7d3d8cafc04c52090c34ce3.tar.gz |
libteam: double NETLINK_RCVBUF to fix -ENOMEM error
v1: Change 96k to 192k in the comments
We are seeing the following errors on some systems configured for LACP:
eth0: Failed to set "priority".
Loop callback failed with: Cannot allocate memory
Failed loop callback: libteam_events, 0x
The slave is then rejected and does not become part of the team. We debugged
this down to -ENOMEM netlink error getting returned by nl_recvmsgs()called
by send_and_recv(). Doubling the buffer size fixed the problem.
Signed-off-by: Jon Maxwell <jmaxwell37@gmail.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
-rw-r--r-- | libteam/libteam.c | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/libteam/libteam.c b/libteam/libteam.c index ce0467e..0065a7f 100644 --- a/libteam/libteam.c +++ b/libteam/libteam.c @@ -551,9 +551,9 @@ int team_destroy(struct team_handle *th) /* \endcond */ /* libnl uses default 32k socket receive buffer size, - * whicn can get too small. Use 96k for all sockets. + * which can get too small. Use 192k for all sockets. */ -#define NETLINK_RCVBUF 98304 +#define NETLINK_RCVBUF 196608 /** * @param th libteam library context |