- 13 Oct, 2017 23 commits
-
-
Jon Maloy authored
We introduce an end-to-end flow control mechanism for group broadcast messages. This ensures that no messages are ever lost because of destination receive buffer overflow, with minimal impact on performance. For now, the algorithm is based on the assumption that there is only one active transmitter at any moment in time. Signed-off-by:
Jon Maloy <jon.maloy@ericsson.com> Acked-by:
Ying Xue <ying.xue@windriver.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jon Maloy authored
Like with any other service, group members' availability can be subscribed for by connecting to be topology server. However, because the events arrive via a different socket than the member socket, there is a real risk that membership events my arrive out of synch with the actual JOIN/LEAVE action. I.e., it is possible to receive the first messages from a new member before the corresponding JOIN event arrives, just as it is possible to receive the last messages from a leaving member after the LEAVE event has already been received. Since each member socket is internally also subscribing for membership events, we now fix this problem by passing those events on to the user via the member socket. We leverage the already present member synch- ronization protocol to guarantee correct message/event order. An event is delivered to the user as an empty message where the two source addresses identify the new/lost member. Furthermore, we set the MSG_OOB bit in the message flags to mark it as an event. If the event is an indication about a member loss we also set the MSG_EOR bit, so it can be distinguished from a member addition event. Signed-off-by:
-
Jon Maloy authored
With group communication, it becomes important for a message receiver to identify not only from which socket (identfied by a node:port tuple) the message was sent, but also the logical identity (type:instance) of the sending member. We fix this by adding a second instance of struct sockaddr_tipc to the source address area when a message is read. The extra address struct is filled in with data found in the received message header (type,) and in the local member representation struct (instance.) Signed-off-by:
-
Jon Maloy authored
As a preparation for introducing flow control for multicast and datagram messaging we need a more strictly defined framework than we have now. A socket must be able keep track of exactly how many and which other sockets it is allowed to communicate with at any moment, and keep the necessary state for those. We therefore introduce a new concept we have named Communication Group. Sockets can join a group via a new setsockopt() call TIPC_GROUP_JOIN. The call takes four parameters: 'type' serves as group identifier, 'instance' serves as an logical member identifier, and 'scope' indicates the visibility of the group (node/cluster/zone). Finally, 'flags' makes it possible to set certain properties for the member. For now, there is only one flag, indicating if the creator of the socket wants to receive a copy of broadcast or multicast messages it is sending via the socket, and if wants to be eligible as destination for its own anycasts. A group is closed, i.e., sockets which have not joined a group will not be able to send messages to or receive messages from members of the group, and vice versa. Any member of a group can send multicast ('group broadcast') messages to all group members, optionally including itself, using the primitive send(). The messages are received via the recvmsg() primitive. A socket can only be member of one group at a time. Signed-off-by: -
Jon Maloy authored
We often see a need for a linked list of destination identities, sometimes containing a port number, sometimes a node identity, and sometimes both. The currently defined struct u32_list is not generic enough to cover all cases, so we extend it to contain two u32 integers and rename it to struct tipc_dest_list. Signed-off-by:
-
Jon Maloy authored
We see an increasing need to send multiple single-buffer messages of TIPC_SYSTEM_IMPORTANCE to different individual destination nodes. Instead of looping over the send queue and sending each buffer individually, as we do now, we add a new help function tipc_node_distr_xmit() to do this. Signed-off-by:
-
Jon Maloy authored
In the following commits we will need to handle multiple incoming and rejected/returned buffers in the function socket.c::filter_rcv(). As a preparation for this, we generalize the function by handling buffer queues instead of individual buffers. We also introduce a help function tipc_skb_reject(), and rename filter_rcv() to tipc_sk_filter_rcv() in line with other functions in socket.c. Signed-off-by:
-
Jon Maloy authored
In the coming commits, functions at the socket level will need the ability to read the availability status of a given node. We therefore introduce a new function for this purpose, while renaming the existing static function currently having the wanted name. Signed-off-by:
-
Jon Maloy authored
The address given to tipc_connect() is not completely sanity checked, under the assumption that this will be done later in the function __tipc_sendmsg() when the address is used there. However, the latter functon will in the next commits serve as caller to several other send functions, so we want to move the corresponding sanity check there to the beginning of that function, before we possibly need to grab the address stored by tipc_connect(). We must therefore be able to trust that this address already has been thoroughly checked. We do this in this commit. Signed-off-by:
-
Jon Maloy authored
As preparation for introducing communication groups, we add the ability to issue topology subscriptions and receive topology events from kernel space. This will make it possible for group member sockets to keep track of other group members. Signed-off-by:
-
Florian Westphal authored
Signed-off-by:
Florian Westphal <fw@strlen.de> Reviewed-by:
Sabrina Dubroca <sd@queasysnail.net> Signed-off-by:
-
Geert Uytterhoeven authored
The module clock is used for two purposes: - Wake-on-LAN (WoL), which is optional, - gPTP Timer Increment (GTI) configuration, which is mandatory. As the clock is needed for GTI configuration anyway, WoL is always available. Hence remove duplication and repeated obtaining of the clock by making GTI use the stored clock for WoL use. Signed-off-by:
Geert Uytterhoeven <geert+renesas@glider.be> Reviewed-by:
Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Reviewed-by:
Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by:
-
David S. Miller authored
Rafał Miłecki says: ==================== net: support bgmac with B50212E B1 PHY I got a report that a board with BCM47189 SoC and B50212E B1 PHY doesn't work well some devices as there is massive ping loss. After analyzing PHY state it has appeared that is runs in slave mode and doesn't auto switch to master properly when needed. This patchset fixes this by: 1) Adding new flag support to the PHY driver for setting master mode 2) Modifying bgmac to request master mode for reported hardware ==================== Signed-off-by: -
Rafał Miłecki authored
There are 4 very similar PHYs: 0x600d84a1: BCM54210E (rev B0) 0x600d84a2: BCM54210E (rev B1) 0x600d84a5: B50212E (rev B0) 0x600d84a6: B50212E (rev B1) that need setting master mode manually. It's because they run in slave mode by default with Automatic Slave/Master configuration disabled which can lead to unreliable connection with massive ping loss. So far it was reported for a board with BCM47189 SoC and B50212E B1 PHY connected to the bgmac supported ethernet device. Telling PHY driver to setup PHY properly solves this issue. Signed-off-by:
Rafał Miłecki <rafal@milecki.pl> Signed-off-by:
-
Rafał Miłecki authored
Some of Broadcom's PHYs run by default in slave mode with Automatic Slave/Master configuration disabled. It stops them from working properly with some devices. So far it has been verified for BCM54210E and BCM50212E which don't work well with Intel's I217-LM and I218-LM: http://ark.intel.com/products/60019/Intel-Ethernet-Connection-I217-LM http://ark.intel.com/products/71307/Intel-Ethernet-Connection-I218-LM I was told there is massive ping loss. This commit adds support for a new flag which can be set by an ethernet driver to fixup PHY setup. Signed-off-by:
-
Mahesh Bandewar authored
If the underlying master ever changes its L2 (e.g. bonding device), then make sure that the IPvlan slaves always emit packets with the current L2 of the master instead of the stale mac addr which was copied during the device creation. The problem can be seen with following script - #!/bin/bash # Create a vEth pair ip link add dev veth0 type veth peer name veth1 ip link set veth0 up ip link set veth1 up ip link show veth0 ip link show veth1 # Create an IPvlan device on one end of this vEth pair. ip link add link veth0 dev ipvl0 type ipvlan mode l2 ip link show ipvl0 # Change the mac-address of the vEth master. ip link set veth0 address 02:11:22:33:44:55 Fixes: 2ad7bf36 ("ipvlan: Initial check-in of the IPVLAN driver.") Signed-off-by:
Mahesh Bandewar <maheshb@google.com> Signed-off-by:
-
David S. Miller authored
Alexander Aring says: ==================== sched: act: ife: UAPI checks and performance tweaks this patch series contains at first a patch which adds a check for IFE_ENCODE and IFE_DECODE when a ife act gets created or updated and adding handling of these cases only inside the act callback only. The second patch use per-cpu counters and move the spinlock around so that the spinlock is less being held in act callback. The last patch use rcu for update parameters and also move the spinlock for the same purpose as in patch 2. Notes: - There is still a spinlock around for protecting the metalist and a rw-lock for another list. Should be migrated to a rcu list, ife possible. - I use still dereference in dump callback, so I think what I didn't got was what happened when rcu_assign_pointer will do when rcu read lock is held. I suppose the pointer will be updated, then we don't have any issue here. ==================== Signed-off-by: -
Alexander Aring authored
This patch changes the parameter updating via RCU and not protected by a spinlock anymore. This reduce the time that the spinlock is being held. Signed-off-by:
Alexander Aring <aring@mojatatu.com> Acked-by:
Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by:
-
Alexander Aring authored
This patch migrates the current counter handling which is protected by a spinlock to a per-cpu counter handling. This reduce the time where the spinlock is being held. Signed-off-by:
-
Alexander Aring authored
This patch adds the check of the two possible ife handlings encode and decode to the init callback. The decode value is for usability aspect and used in userspace code only. The current code offers encode else decode only. This patch avoids any other option than this. Signed-off-by:
-
David S. Miller authored
Roman Mashak says: ==================== net: sched: Fix IFE meta modules loading Adjust module alias names of IFE meta modules and fix the bug that prevented auto-loading IFE modules in run-time. ==================== Signed-off-by: -
Roman Mashak authored
Macro __stringify_1() can stringify a macro argument, however IFE_META_* are enums, so they never expand, however request_module expects an integer in IFE module name, so as a result it always fails to auto-load. Fixes: ef6980b6 ("introduce IFE action") Signed-off-by:
Roman Mashak <mrv@mojatatu.com> Acked-by:
Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by:
-
Roman Mashak authored
Make style of module alias name consistent with other subsystems in kernel, for example net devices. Fixes: 084e2f65 ("Support to encoding decoding skb mark on IFE action") Fixes: 200e10f4 ("Support to encoding decoding skb prio on IFE action") Fixes: 408fbc22 ("net sched ife action: Introduce skb tcindex metadata encap decap") Signed-off-by:
-
- 12 Oct, 2017 17 commits
-
-
Christos Gkekas authored
Delete unused channel variables in vxge-traffic. Signed-off-by:
Christos Gkekas <chris.gekas@gmail.com> Signed-off-by:
-
Florian Fainelli authored
This file contains unnecessary whitespaces as newlines, remove them, found by looking at what struct tc_mirred looks like. Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Arnd Bergmann authored
When af_mpls is built-in but the tunnel support is a module, we get a link failure: net/mpls/af_mpls.o: In function `mpls_init': af_mpls.c:(.init.text+0xdc): undefined reference to `ip_tunnel_encap_add_ops' This adds a Kconfig statement to prevent the broken configuration and force mpls to be a module as well in this case. Fixes: bdc47641 ("ip_tunnel: add mpls over gre support") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Acked-by:
Amine Kherbouche <amine.kherbouche@6wind.com> Signed-off-by:
-
David S. Miller authored
Ursula Braun says: ==================== net/smc: ib_query_gid() patches triggered by Parav Pandit here are 2 cleanup patches for usage of ib_query_gid() in the smc-code. Thanks, Ursula v2 changes advised by Parav Pandit: extra check is_vlan_dev() in patch 2/2 "RoCE" spelling added "Reported-by" added "Reviewed-by" added "Fixes" ==================== Signed-off-by: -
Ursula Braun authored
For RoCEs ib_query_gid() takes a reference count on the net_device. This reference count must be decreased by the caller. Signed-off-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Reported-by:
Parav Pandit <parav@mellanox.com> Reviewed-by:
-
Ursula Braun authored
SMC should not open code the function pointer get_netdev of the IB device. Replacing ib_query_gid(..., NULL) with ib_query_gid(..., gid_attr) allows access to the netdev. Signed-off-by:
-
David S. Miller authored
Florian Fainelli says: ==================== Enable ACB for bcm_sf2 and bcmsysport This patch series enables Broadcom's Advanced Congestion Buffering mechanism which requires cooperation between the CPU/Management Ethernet MAC controller and the switch. I took the notifier approach because ultimately the information we need to carry to the master network device is DSA specific and I saw little room for generalizing beyond what DSA requires. Chances are that this is highly specific to the Broadcom HW as I don't know of any HW out there that supports something nearly similar for similar or identical needs. ==================== Signed-off-by: -
Florian Fainelli authored
Now that we have established the queue mapping between the switch port egress queues and the SYSTEMPORT egress queues, we can turn on Advanced Congestion Buffering (ACB) at the SYSTEMPORT level. This enables the Ethernet MAC controller to get out of band flow control information directly from the switch port and queue that it monitors such that its internal TDMA can be appropriately backpressured. Signed-off-by:
-
Florian Fainelli authored
Turn on the out of band Advanced Congestion Buffering (ACB) mechanism at the switch level now that we have properly established the queue mapping between the switch egress queues and the SYSTEMPORT egress queues. This allows the switch to correctly backpressure the host system when one of its queue drops below the configured thresholds. This is also helping achieve so called "lossless" behavior by adapting the TX interrupt pacing to the actual speed and capacity of the switch port. Signed-off-by:
-
Florian Fainelli authored
Establish a queue mapping between the DSA slave network device queues created that correspond to switch port queues, and the transmit queue that SYSTEMPORT manages. We need to configure the SYSTEMPORT transmit queue with the switch port number and switch port queue number in order for the switch and SYSTEMPORT hardware to utilize the out of band congestion notification. This hardware mechanism works by looking at the switch port egress queue and determines whether there is enough buffers for this queue, with that class of service for a successful transmission and if not, backpressures the SYSTEMPORT queue that is being used. For this to work, we implement a notifier which looks at the DSA_PORT_REGISTER event. When DSA network devices are registered, the framework calls the DSA notifiers when that happens, extracts the number of queues for these devices and their associated port number, remembers that in the driver private structure and linearly maps those queues to TX rings/queues that we manage. This scheme works because DSA slave network deviecs always transmit through SYSTEMPORT so when DSA slave network devices are destroyed/brought down, the corresponding SYSTEMPORT queues are no longer used. Also, by design of the DSA framework, the master network device (SYSTEMPORT) is registered first. For faster lookups we use an array of up to DSA_MAX_PORTS * number of queues per port, and then map pointers to bcm_sysport_tx_ring such that our ndo_select_queue() implementation can just index into that array to locate the corresponding ring index. Signed-off-by:
-
Florian Fainelli authored
We need to tell the DSA master network device doing the actual transmission what the desired switch port and queue number is for it to resolve that to the internal transmit queue it is mapped to. Signed-off-by:
-
Florian Fainelli authored
In preparation for communicating a given DSA network device's port number and switch index, create a specialized DSA notifier and two events: DSA_PORT_REGISTER and DSA_PORT_UNREGISTER that communicate: the slave network device (slave_dev), port number and switch number in the tree. This will be later used for network device drivers like bcmsysport which needs to cooperate with its DSA network devices to set-up queue mapping and scheduling. Signed-off-by:
-
Timur Tabi authored
This reverts commit df1ec1b9. It turns out that memory allocated via dma_alloc_coherent is always aligned to the size of the buffer, so there's no way the RRD and RFD can ever be in separate 32-bit regions. Signed-off-by:
Timur Tabi <timur@codeaurora.org> Signed-off-by:
-
Eric Dumazet authored
Remove three inline helpers that are no longer needed. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Colin Ian King authored
Variable old_flags is being assigned but is never read; it is redundant and can be removed. Cleans up clang warning: Value stored to 'old_flags' is never read Signed-off-by:
Colin Ian King <colin.king@canonical.com> Acked-by:
Alexei Starovoitov <ast@kernel.org> Acked-by:
Daniel Borkmann <daniel@iogearbox.net> Signed-off-by:
-
David S. Miller authored
Tariq Toukan says: ==================== mlx4_en XDP TX improvements This patchset contains performance improvements to the XDP_TX use case in the mlx4 Eth driver. Patch 1 is a simple change in a function parameter type. Patch 2 replaces a call to a generic function with the relevant parts inlined. Patch 3 moves the write of descriptors' constant values from data path to control path. Series generated against net-next commit: 833e0e2f net: dst: move cpu inside ifdef to avoid compilation warning ==================== Signed-off-by:
-
Tariq Toukan authored
In XDP_TX, some fields in tx_info and tx_desc are constants across all entries of the different XDP_TX rings. Assign values to these fields on ring creation time, rather than in data-path. Patchset performance tests: Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz Single queue no-RSS optimization ON. XDP_TX packet rate: ------------------------------ Before | After | Gain | 13.7 Mpps | 14.0 Mpps | %2.2 | ------------------------------ Signed-off-by:
Tariq Toukan <tariqt@mellanox.com> Signed-off-by:
-
