summaryrefslogtreecommitdiff
path: root/include/net/gro_cells.h
AgeCommit message (Collapse)AuthorFilesLines
2017-02-08gro_cells: move to net/core/gro_cells.cEric Dumazet1-82/+4
We have many gro cells users, so lets move the code to avoid duplication. This creates a CONFIG_GRO_CELLS option. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-16gro_cells: mark napi struct as not busy poll candidatesEric Dumazet1-0/+3
Rolf Neugebauer reported very long delays at netns dismantle. Eric W. Biederman was kind enough to look at this problem and noticed synchronize_net() occurring from netif_napi_del() that was added in linux-4.5 Busy polling makes no sense for tunnels NAPI. If busy poll is used for sessions over tunnels, the poller will need to poll the physical device queue anyway. netif_tx_napi_add() could be used here, but function name is misleading, and renaming it is not stable material, so set NAPI_STATE_NO_BUSY_POLL bit directly. This will avoid inserting gro_cells napi structures in napi_hash[] and avoid the problematic synchronize_net() (per possible cpu) that Rolf reported. Fixes: 93d05d4a320c ("net: provide generic busy polling to all NAPI drivers") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Rolf Neugebauer <rolf.neugebauer@docker.com> Reported-by: Eric W. Biederman <ebiederm@xmission.com> Acked-by: Cong Wang <xiyou.wangcong@gmail.com> Tested-by: Rolf Neugebauer <rolf.neugebauer@docker.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-07-22gro_cells: gro_cells_receive now return error codePaolo Abeni1-6/+5
so that the caller can update stats accordingly, if needed Signed-off-by: Paolo Abeni <pabeni@redhat.com> Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-09-01gro_cells: remove spinlock protecting receive queuesEric Dumazet1-13/+5
As David pointed out, spinlock are no longer needed to protect the per cpu queues used in gro cells infrastructure. Also use new napi_complete_done() API so that gro_flush_timeout tweaks have an effect. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-01-18ip_tunnel: Create percpu gro_cellMartin KaFai Lau1-17/+12
In the ipip tunnel, the skb->queue_mapping is lost in ipip_rcv(). All skb will be queued to the same cell->napi_skbs. The gro_cell_poll is pinned to one core under load. In production traffic, we also see severe rx_dropped in the tunl iface and it is probably due to this limit: skb_queue_len(&cell->napi_skbs) > netdev_max_backlog. This patch is trying to alloc_percpu(struct gro_cell) and schedule gro_cell_poll to process the skb in the same core. Signed-off-by: Martin KaFai Lau <kafai@fb.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2013-01-28gro: Fix kcalloc argument orderJoe Perches1-2/+2
First number, then size. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-12-11net: fix a race in gro_cell_poll()Eric Dumazet1-5/+9
Dmitry Kravkov reported packet drops for GRE packets since GRO support was added. There is a race in gro_cell_poll() because we call napi_complete() without any synchronization with a concurrent gro_cells_receive() Once bug was triggered, we queued packets but did not schedule NAPI poll. We can fix this issue using the spinlock protected the napi_skbs queue, as we have to hold it to perform skb dequeue anyway. As we open-code skb_dequeue(), we no longer need to mask IRQS, as both producer and consumer run under BH context. Bug added in commit c9e6bc644e (net: add gro_cells infrastructure) Reported-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Tested-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2012-10-02net: add gro_cells infrastructureEric Dumazet1-0/+103
This adds a new include file (include/net/gro_cells.h), to bring GRO (Generic Receive Offload) capability to tunnels, in a modular way. Because tunnels receive path is lockless, and GRO adds a serialization using a napi_struct, I chose to add an array of up to DEFAULT_MAX_NUM_RSS_QUEUES cells, so that multi queue devices wont be slowed down because of GRO layer. skb_get_rx_queue() is used as selector. In the future, we might add optional fanout capabilities, using rxhash for example. With help from Ben Hutchings who reminded me netif_get_num_default_rss_queues() function. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>