- 05 Feb, 2015 40 commits
-
-
Markus Elfring authored
The kfree() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
Markus Elfring <elfring@users.sourceforge.net> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Markus Elfring authored
The kfree() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
-
Markus Elfring authored
The kfree() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
-
Markus Elfring authored
The release_firmware() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
-
Hariprasad Shenai authored
cxgb_busy_poll, corresponding to ndo_busy_poll, gets called by the socket waiting for data. With busy_poll enabled, improvement is seen in latency numbers as observed by collecting netperf TCP_RR numbers. Below are latency number, with and without busy-poll, in a switched environment for a particular msg size: netperf command: netperf -4 -H <ip> -l 30 -t TCP_RR -- -r1,1 Latency without busy-poll: ~16.25 us Latency with busy-poll : ~08.79 us Based on original work by Kumar Sanghvi <kumaras@chelsio.com> Signed-off-by:
Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by:
-
David S. Miller authored
Revert "bridge: Let bridge not age 'externally' learnt FDB entries, they are removed when 'external' entity notifies the aging" This reverts commit 9a05dde5. Requested by Scott Feldman. Signed-off-by:
-
Eric Dumazet authored
FQ has a fast path for skb attached to a socket, as it does not have to compute a flow hash. But for other packets, FQ being non stochastic means that hosts exposed to random Internet traffic can allocate million of flows structure (104 bytes each) pretty easily. Not only host can OOM, but lookup in RB trees can take too much cpu and memory resources. This patch adds a new attribute, orphan_mask, that is adding possibility of having a stochastic hash for orphaned skb. Its default value is 1024 slots, to mimic SFQ behavior. Note: This does not apply to locally generated TCP traffic, and no locally generated traffic will share a flow structure with another perfect or stochastic flow. This patch also handles the specific case of SYNACK messages: They are attached to the listener socket, and therefore all map to a single hash bucket. If listener have set SO_MAX_PACING_RATE, hoping to have new accepted socket inherit this rate, SYNACK might be paced and even dropped. This is very similar to an internal patch Google have used more than one year. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfsDavid S. Miller authored
More iov_iter work from Al Viro. Signed-off-by: -
Eric Dumazet authored
When we added pacing to TCP, we decided to let sch_fq take care of actual pacing. All TCP had to do was to compute sk->pacing_rate using simple formula: sk->pacing_rate = 2 * cwnd * mss / rtt It works well for senders (bulk flows), but not very well for receivers or even RPC : cwnd on the receiver can be less than 10, rtt can be around 100ms, so we can end up pacing ACK packets, slowing down the sender. Really, only the sender should pace, according to its own logic. Instead of adding a new bit in skb, or call yet another flow dissection, we tweak skb->truesize to a small value (2), and we instruct sch_fq to use new helper and not pace pure ack. Note this also helps TCP small queue, as ack packets present in qdisc/NIC do not prevent sending a data packet (RPC workload) This helps to reduce tx completion overhead, ack packets can use regular sock_wfree() instead of tcp_wfree() which is a bit more expensive. This has no impact in the case packets are sent to loopback interface, as we do not coalesce ack packets (were we would detect skb->truesize lie) In case netem (with a delay) is used, skb_orphan_partial() also sets skb->truesize to 1. This patch is a combination of two patches we used for about one year at Google. Signed-off-by:
-
David S. Miller authored
Herbert Xu says: ==================== rhashtable: Add iterators and use them The first patch fixes a potential crash with nft_hash destroying the table during a shrinking process. While the next patch adds rhashtable iterators to replace current manual walks used by netlink and netfilter. The final two patches make use of these iterators in netlink and netfilter. ==================== Signed-off-by: -
Herbert Xu authored
This patch gets rid of the manual rhashtable walk in nft_hash which touches rhashtable internals that should not be exposed. It does so by using the rhashtable iterator primitives. Note that I'm leaving nft_hash_destroy alone since it's only invoked on shutdown and it shouldn't be affected by changes to rhashtable internals (or at least not what I'm planning to change). Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
-
Herbert Xu authored
This patch gets rid of the manual rhashtable walk in netlink which touches rhashtable internals that should not be exposed. It does so by using the rhashtable iterator primitives. In fact the existing code was very buggy. Some sockets weren't shown at all while others were shown more than once. Signed-off-by:
-
Herbert Xu authored
Some existing rhashtable users get too intimate with it by walking the buckets directly. This prevents us from easily changing the internals of rhashtable. This patch adds the helpers rhashtable_walk_init/exit/start/next/stop which will replace these custom walkers. They are meant to be usable for both procfs seq_file walks as well as walking by a netlink dump. The iterator structure should fit inside a netlink dump cb structure, with at least one element to spare. Signed-off-by:
-
Herbert Xu authored
The current being_destroyed check in rhashtable_expand is not enough since if we start a shrinking process after freeing all elements in the table that's also going to crash. This patch adds a being_destroyed check to the deferred worker thread so that we bail out as soon as we take the lock. Signed-off-by:
-
Markus Elfring authored
The functions cpsw_ale_destroy() and of_dev_put() test whether their argument is NULL and then return immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
-
Markus Elfring authored
The of_dev_put() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
-
David S. Miller authored
Amir Vadai says: ==================== Mellanox drivers updates Feb-03-2015 This patchset introduces some small bug fixes and code cleanups in mlx4_core, mlx4_en and mlx5_core. I am sending it in parallel to the patchset sent by Or Gerlitz today [1] because this is the end of the time frame for 3.20. I also checked that there are no conflicts between those two patchsets (Or's patchset is focused on the bonding area while this on Mellanox drivers). The patchset was applied on top of commit 7d37d0c1 ('net: sctp: Deletion of an unnecessary check before the function call "kfree"') [1] - [PATCH 00/10] Add HA and LAG support to mlx4 RoCE and SRIOV services http://marc.info/?l=linux-netdev&m=142297582610254&w=2 ==================== Signed-off-by:
-
Ido Shamay authored
Notify users when TX vlan offload feature changed with ethtool. Relevant command - ethtool -K <eth> txvlan on/off. Signed-off-by:
Ido Shamay <idos@mellanox.com> Signed-off-by:
Amir Vadai <amirv@mellanox.com> Signed-off-by:
-
Ido Shamay authored
This patch improves memory utilization and therefore the packets rate for special MTU's. Instead of setting the frag_stride to the maximal hard coded frag_size, use the actual frag_size that is set according to the MTU, when setting the stride of the last frag. So, for example, for MTU 1600, where the frag_size of the 2nd frag is 86, the frag_size is set to 128 instead of 4096. See below: Before: frag:0 - size:1536 prefix:0 stride:1536 frag:1 - size:86 prefix:1536 stride:4096 frag 0 allocator: - size:32768 frags:21 frag 1 allocator: - size:32768 frags:8 After: frag:0 - size:1536 prefix:0 stride:1536 frag:1 - size:86 prefix:1536 stride:128 frag 0 allocator: - size:32768 frags:21 frag 1 allocator: - size:32768 frags:256 Signed-off-by:
-
Ido Shamay authored
After Initialization of page_alloc, print actual allocated page size and number of frags it contains. prints is done only when drv message level is set on the interface. Signed-off-by:
-
Or Gerlitz authored
Align the IDs in the code with the modinfo, lspci -n, etc tools outputs. Signed-off-by:
Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by:
-
Or Gerlitz authored
We do support cache line sizes of 32 and 64 bytes without activating the CQE stride feature. Fix a misleading print saying that these cache line sizes aren't supported. Signed-off-by:
-
Maor Gottlieb authored
Add Initialization to struct config_dev before filling and using it. Fix to warning: warning: config_dev.rx_checksum_val may be used uninitialized in this function Signed-off-by:
Maor Gottlieb <maorg@mellanox.com> Signed-off-by:
-
Maor Gottlieb authored
a) Previously, mlx4_mr_rereg_write filled the MPT's start and length with the old MPT's values. Fixing the initialization to take the new start and length. b) In addition access flags in mpt_status were initialized instead of status due to bad boolean operation. Fixing the operation. c) Initialization of pd_slave caused a protection error. Fix - removing this initialization. d) In resource_tracker.c: Fixing vf encoding to be one-based. Fixes: e630664c ('mlx4_core: Add helper functions to support MR re-registration') Signed-off-by:
-
David S. Miller authored
Or Gerlitz says: ==================== Add HA and LAG support to mlx4 RoCE and SRIOV services This series takes advanges of bonding mlx4 Ethernet devices to support a model of High-Availability and Link Aggregation for more environments. The mlx4 driver reacts on netdev events generated by bonding when slave state changes happen by programming a HW V2P (Virt-to-Phys) port table. Bonding was extended to expose these state changes through netdev events. When an mlx4 interface such as the mlx4 IB/RoCE driver is subject to this policy, QPs are created over virtual ports which are mapped to one of the two physical ports. When a failure happens, the re-programming of the V2P table allows traffic to keep flowing. The mlx4 Ethernet driver interfaces are not subject to this policy and act as usual. A 2nd use-case for this model would be to add HA and Link Aggregation support to single ported mlx4 Ethernet VFs. In this case, the PF Ethernet intrfaces are bonded, all the VFs see single port devices (which is supported already today), and VF QPs are subject to V2P. ==================== Signed-off-by: -
Moni Shoua authored
When the mlx4 IB (RoCE) device works in link aggregation mode, it exposes a single port to upper layers. Therefore, applications always set '1' in port_num attribute when modifying a QP or creating an address handle. To make sure that a node uses all available ports the mlx4 driver will override the port_num attribute with a round robin policy. Signed-off-by:
Moni Shoua <monis@mellanox.com> Signed-off-by:
-
Moni Shoua authored
In port aggregation mode flows for port #1 (the only port) should be mirrored on port #2. This is because packets can arrive from either physical ports. Signed-off-by:
-
Moni Shoua authored
Register the interface with the mlx4 core driver with port aggregation support and check for port aggregation mode when the 'add' function is called. In this mode, only one physical port is reported to the upper layer (RoCE/IB core stack and ULPs). Signed-off-by:
-
Moni Shoua authored
This function is implemented twice... get rid of one copy. Signed-off-by:
-
Moni Shoua authored
Capture NETDEV events generated by the bonding driver and based on that make decisions of how to configure port aggregation in the mlx4 core driver. This includes setting the V2P port table and re-creating the interested interfaces in bonded/non-bonded mode. Signed-off-by:
-
Moni Shoua authored
Supply interface functions to bond and unbond ports of a mlx4 internal interfaces. Example for such an interface is the one registered by the mlx4 IB driver under RoCE. There are 1. Functions to go in/out to/from bonded mode 2. Function to remap virtual ports to physical ports The bond_mutex prevents simultaneous access to data that keep status of the device in bonded mode. The upper mlx4 interface marks to the mlx4 core module that they want to be subject for such bonding by setting the MLX4_INTFF_BONDING flag. Interface which goes to/from bonded mode is re-created. The mlx4 Ethernet driver does not set this flag when registering the interface, the IB driver does. Signed-off-by:
-
Moni Shoua authored
Implement the hardware interface required for port aggregation. 1. Disable RX port check on receive - don't perform a validity check that matches to QP's port and the port where the packet is received. 2. Virtual to physical port remap - configure virtual to physical port mapping. Port remap capability for virtual functions. Signed-off-by:
-
Moni Shoua authored
Use notifier chain to dispatch an event upon a change in slave state. Event is dispatched with slave specific info. Signed-off-by:
-
Moni Shoua authored
Move slave state changes to a helper function, this is a pre-step for adding functionality of dispatching an event when this helper is called. This commit doesn't add new functionality. Signed-off-by:
-
Moni Shoua authored
Add event which provides an indication on a change in the state of a bonding slave. The event handler should cast the pointer to the appropriate type (struct netdev_bonding_info) in order to get the full info about the slave. Signed-off-by:
-
David S. Miller authored
Jon Maloy says: ==================== tipc: some small fixes During extensive testing and analysis of running dual links between nodes, we have encountered some issues that potentially may cause problems. We choose to fix those proactively in this series. ==================== Signed-off-by: -
Jon Paul Maloy authored
When a new link instance is created, it is trigged to start by sending it a TIPC_STARTING_EVT, whereafter a regular link reset is applied to it. The starting event is codewise treated as a timeout event, and prompts a link RESET message to be sent to the peer node, carrying a link session identifier. The later link_reset() call nudges this session identifier, whereafter all subsequent RESET messages will be sent out with the new identifier. The latter session number overrides the former, causing the peer to unconditionally accept it irrespective of its current working state. We don't think that this causes any problem, but it is not in accordance with the protocol spec, and may cause confusion when debugging TIPC sessions. To avoid this, we make the starting event distinct from the subsequent timeout events, by not allowing the former to send out any RESET message. This eliminates the described problem. Reviewed-by:
Erik Hugne <erik.hugne@ericsson.com> Reviewed-by:
Ying Xue <ying.xue@windriver.com> Signed-off-by:
Jon Maloy <jon.maloy@ericsson.com> Signed-off-by:
-
Jon Paul Maloy authored
Instances of struct node are created in the function tipc_disc_rcv() under the assumption that there is no race between received discovery messages arriving from the same node. This assumption is wrong. When we use more than one bearer, it is possible that discovery messages from the same node arrive at the same moment, resulting in creation of two instances of struct tipc_node. This may later cause confusion during link establishment, and may result in one of the links never becoming activated. We fix this by making lookup and potential creation of nodes atomic. Instead of first looking up the node, and in case of failure, create it, we now start with looking up the node inside node_link_create(), and return a reference to that one if found. Otherwise, we go ahead and create the node as we did before. Reviewed-by:
-
Jon Paul Maloy authored
During link failover it may happen that the remaining link goes down while it is still in the process of taking over traffic from a previously failed link. When this happens, we currently abort the failover procedure and reset the first failed link to non-failover mode, so that it will be ready to re-establish contact with its peer when it comes available. However, if the first link goes down because its bearer was manually disabled, it is not enough to reset it; it must also be deleted; which is supposed to happen when the failover procedure is finished. Otherwise it will remain a zombie link: attached to the owner node structure, in mode LINK_STOPPED, and permanently blocking any re- establishing of the link to the peer via the interface in question. We fix this by amending the failover abort procedure. Apart from resetting the link to non-failover state, we test if the link is also in LINK_STOPPED mode. If so, we delete it, using the conditional tipc_link_delete() function introduced in the previous commit. Reviewed-by:
-
Jon Paul Maloy authored
When a bearer is disabled, all pertaining links will be reset and deleted. However, if there is a second active link towards a killed link's destination, the delete has to be postponed until the failover is finished. During this interval, we currently put the link in zombie mode, i.e., we take it out of traffic, delete its timer, but leave it attached to the owner node structure until all missing packets have been received. When this is done, we detach the link from its node and delete it, assuming that the synchronous timer deletion that was initiated earlier in a different thread has finished. This is unsafe, as the failover may finish before del_timer_sync() has returned in the other thread. We fix this by adding an atomic reference counter of type kref in struct tipc_link. The counter keeps track of the references kept to the link by the owner node and the timer. We then do a conditional delete, based on the reference counter, both after the failover has been finished and when the timer expires, if applicable. Whoever comes last, will actually delete the link. This approach also implies that we can make the deletion of the timer asynchronous. Reviewed-by:
-
