| Message ID | 20180524222548.13217-2-fw@strlen.de |
|---|---|
| State | Accepted |
| Delegated to: | Pablo Neira |
| Headers | show |
| Series | netfilter: nf_tables: speed up commit phase | expand |
diff --git a/net/netfilter/nfnetlink.c b/net/netfilter/nfnetlink.c index 03ead8a9e90c..88c9e222b670 100644 --- a/net/netfilter/nfnetlink.c +++ b/net/netfilter/nfnetlink.c @@ -441,7 +441,14 @@ static void nfnetlink_rcv_batch(struct sk_buff *skb, struct nlmsghdr *nlh, kfree_skb(skb); goto replay; } else if (status == NFNL_BATCH_DONE) { - ss->commit(net, oskb); + err = ss->commit(net, oskb); + if (err == -EAGAIN) { + status |= NFNL_BATCH_REPLAY; + goto done; + } else if (err) { + ss->abort(net, oskb); + netlink_ack(oskb, nlmsg_hdr(oskb), err, NULL); + } } else { ss->abort(net, oskb); }
->commit() cannot fail at the moment. Followup-patch adds kmalloc calls in the commit phase, so we'll need to be able to handle errors. Make it so that -EGAIN causes a full replay, and make other errors cause the transaction to fail. Failing is ok from a consistency point of view as long as we perform all actions that could return an error before we increment the generation counter and the base seq. Signed-off-by: Florian Westphal <fw@strlen.de> --- net/netfilter/nfnetlink.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)