- 08 Dec, 2016 37 commits
-
-
Eric Dumazet authored
Under UDP flood, many softirq producers try to add packets to UDP receive queue, and one user thread is burning one cpu trying to dequeue packets as fast as possible. Two parts of the per packet cost are : - copying payload from kernel space to user space, - freeing memory pieces associated with skb. If socket is under pressure, softirq handler(s) can try to pull in skb->head the payload of the packet if it fits. Meaning the softirq handler(s) can free/reuse the page fragment immediately, instead of letting udp_recvmsg() do this hundreds of usec later, possibly from another node. Additional gains : - We reduce skb->truesize and thus can store more packets per SO_RCVBUF - We avoid cache line misses at copyout() time and consume_skb() time, and avoid one put_page() with potential alien freeing on NUMA hosts. This comes at the cost of a copy, bounded to available tail room, which is usually small. (We might have to fix GRO_MAX_HEAD which looks bigger than necessary) This patch gave me about 5 % increase in throughput in my tests. skb_condense() helper could probably used in other contexts. Signed-off-by:
Eric Dumazet <edumazet@google.com> Cc: Paolo Abeni <pabeni@redhat.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Raghu Vatsavayi says: ==================== liquidio VF data path Following is V3 patch series that adds support for VF data path related features. It also has following changes related to previous comments: 1) Remove unnecessary "void *" casting. 2) Remove inline for functions and let gcc decide. Please apply patches in following order as some of them depend on earlier patches. ==================== Signed-off-by: -
Raghu Vatsavayi authored
Adds support for VF receive data control path. Signed-off-by:
Raghu Vatsavayi <raghu.vatsavayi@caviumnetworks.com> Signed-off-by:
Derek Chickles <derek.chickles@caviumnetworks.com> Signed-off-by:
Satanand Burla <satananda.burla@caviumnetworks.com> Signed-off-by:
Felix Manlunas <felix.manlunas@caviumnetworks.com> Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for freeing VF xmit buffers. Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for transmit functionality in VF. Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for VF scatter gather lists. Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for configuring mtu, multicast and mac address. Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for VF link status related changes. Signed-off-by:
-
Raghu Vatsavayi authored
Adds support for VF link initialization and offload features. Signed-off-by:
-
Eric Dumazet authored
RFS is not commonly used, so add a jump label to avoid some conditionals in fast path. Signed-off-by:
-
Zhang Shengju authored
Remove first zero for add, and use full attribute name in comments. Signed-off-by:
Zhang Shengju <zhangshengju@cmss.chinamobile.com> Signed-off-by:
-
David S. Miller authored
Niklas Cassel says: ==================== net: stmmac: make DMA programmable burst length more configurable Make DMA programmable burst length more configurable in the stmmac driver. This is done by adding support for independent pbl for tx/rx through DT. More fine grained tuning of pbl is possible thanks to a DT property saying that we should NOT multiply pbl values by x8/x4 in hardware. All new DT properties are optional, and created in a way that it will not affect any existing DT configurations. Changes since V1: Created cover-letter. Rebased patch set against next-20161205, since conflicting patches to stmmac_platform.c has been merged since V1. Changes since V2: Moved default value initialization of pbl to stmmac_platform.c and added a check for pbl != 0 in stmmac_main.c, to catch a possble pbl == 0 from pci glue. ==================== Signed-off-by: -
Niklas Cassel authored
The driver currently always sets the PBLx8/PBLx4 bit, which means that the pbl values configured via the pbl/txpbl/rxpbl DT properties are always multiplied by 8/4 in the hardware. In order to allow the DT to configure lower pbl values, while at the same time not changing behavior of any existing device trees using the pbl/txpbl/rxpbl settings, add a property to disable the multiplication of the pbl by 8/4 in the hardware. Suggested-by:
Rabin Vincent <rabinv@axis.com> Signed-off-by:
Niklas Cassel <niklas.cassel@axis.com> Acked-by:
Alexandre Torgue <alexandre.torgue@st.com> Signed-off-by:
-
Niklas Cassel authored
GMAC and newer supports independent programmable burst lengths for DMA tx/rx. Add new optional devicetree properties representing this. To be backwards compatible, snps,pbl will still be valid, but snps,txpbl/snps,rxpbl will override the value in snps,pbl if set. If the IP is synthesized to use the AXI interface, there is a register and a matching DT property inside the optional stmmac-axi-config DT node for controlling burst lengths, named snps,blen. However, using this register, it is not possible to control tx and rx independently. Also, this register is not available if the IP was synthesized with, e.g., the AHB interface. Signed-off-by:
-
Niklas Cassel authored
DMA_BUS_MODE_RPBL_MASK is really 6 bits, just like DMA_BUS_MODE_PBL_MASK. Signed-off-by:
-
Niklas Cassel authored
commit 64c3b252 ("net: stmmac: fixed the pbl setting with DT") changed the parsing of the DT binding. Before 64c3b252, snps,fixed-burst and snps,mixed-burst were parsed regardless if the property snps,pbl existed or not. After the commit, fixed burst and mixed burst are only parsed if snps,pbl exists. Now when snps,aal has been added, it too is only parsed if snps,pbl exists. Since the DT binding does not specify that fixed burst, mixed burst or aal depend on snps,pbl being specified, undo changes introduced by 64c3b252. The issue commit 64c3b252 ("net: stmmac: fixed the pbl setting with DT") tries to address is solved in another way: The databook specifies that all values other than 1, 2, 4, 8, 16, or 32 results in undefined behavior, so snps,pbl = <0> is invalid. If pbl is 0 after parsing, set pbl to DEFAULT_DMA_PBL. This handles the case where the property is omitted, and also handles the case where the property is specified without any data. Signed-off-by:
-
Niklas Cassel authored
Use struct stmmac_dma_cfg *dma_cfg as an argument rather than using all the struct members as individual arguments. Signed-off-by:
-
Niklas Cassel authored
All drivers except pci glue layer calls stmmac_probe_config_dt. stmmac_probe_config_dt does a kzalloc dma_cfg. pci glue layer does kzalloc dma_cfg explicitly, so all current drivers does a kzalloc dma_cfg. Return an error if no DMA configuration is found, that way we can assume that the DMA configuration always exists. Signed-off-by:
-
Tobias Klauser authored
Use module_pci_driver() to get rid of some boilerplate code. Signed-off-by:
Tobias Klauser <tklauser@distanz.ch> Acked-by:
Lino Sanfilippo <LinoSanfilippo@gmx.de> Signed-off-by:
-
David S. Miller authored
Simon Horman says: ==================== net/sched: cls_flower: Support matching on ICMP this series adds support for matching on ICMP type and code to cls_flower. Changes v5->v6: * Restore missing signed-off-by Changes v4->v5: * Drop all helpers Changes v3->v4: * Do not add icmp to struct flow_keys, it is not needed * Do not test for ICMP protocols in packet in __skb_flow_dissect, this is also not needed Changes v2->v3: * Add FLOW_DISSECTOR_KEY_ICMP and use separate structure for ICMP Changes v1->v2: * Include all dissector helpers in first patch ==================== Signed-off-by: -
Simon Horman authored
Support matching on ICMP type and code. Example usage: tc qdisc add dev eth0 ingress tc filter add dev eth0 protocol ip parent ffff: flower \ indev eth0 ip_proto icmp type 8 code 0 action drop tc filter add dev eth0 protocol ipv6 parent ffff: flower \ indev eth0 ip_proto icmpv6 type 128 code 0 action drop Signed-off-by:
Simon Horman <simon.horman@netronome.com> Acked-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Simon Horman authored
Allow dissection of ICMP(V6) type and code. This should only occur if a packet is ICMP(V6) and the dissector has FLOW_DISSECTOR_KEY_ICMP set. There are currently no users of FLOW_DISSECTOR_KEY_ICMP. A follow-up patch will allow FLOW_DISSECTOR_KEY_ICMP to be used by the flower classifier. Signed-off-by:
-
Niklas Cassel authored
devicetree binding for stmmac states: - compatible: Should be "snps,dwmac-<ip_version>", "snps,dwmac" For backwards compatibility: "st,spear600-gmac" is also supported. Previously, when specifying "snps,dwmac-4.10a", "snps,dwmac" as your compatible string, plat_stmmacenet_data would have both has_gmac and has_gmac4 set. This would lead to stmmac_hw_init calling dwmac1000_setup rather than dwmac4_setup, resulting in a non-functional driver. This happened since the check for has_gmac is done before the check for has_gmac4. However, the order should not matter, so it does not make sense to have both set. If something is valid for both, you should do as the stmmac_interrupt does: if (priv->plat->has_gmac || priv->plat->has_gmac4) ... The places where it was obvious that the author actually meant if (has_gmac || has_gmac4) rather than if (has_gmac) has been updated. Signed-off-by:
-
Niklas Cassel authored
devicetree binding for stmmac states: - compatible: Should be "snps,dwmac-<ip_version>", "snps,dwmac" For backwards compatibility: "st,spear600-gmac" is also supported. Since dwmac-generic.c calls stmmac_probe_config_dt explicitly, another alternative would have been to remove all compatible strings other than "snps,dwmac" and "st,spear600-gmac" from dwmac-generic.c. However, that would probably do more good than harm, since when trying to figure out what hardware a certain driver supports, you usually look at the compatible strings in the struct of_device_id, and not in some function defined in a completely different file. No functional change intended. Signed-off-by:
Giuseppe Cavallaro <peppe.cavallaro@st.com> Signed-off-by:
-
Niklas Cassel authored
snps,tso was previously placed under AXI BUS Mode parameters, suggesting that the property should be in the stmmac-axi-config node. TSO (TCP Segmentation Offloading) has nothing to do with AXI BUS Mode parameters, and the parser actually expects it to be in the root node, not in the stmmac-axi-config. Also added a note about snps,tso only being available on GMAC4 and newer. Signed-off-by:
Rob Herring <robh@kernel.org> Acked-by:
-
Tobias Klauser authored
Do not open code getting the MAC address exclusively from the "local-mac-address" property, but instead use of_get_mac_address() which looks up the MAC address using the 3 typical property names. Also avoid casting away the const qualifier of the return value by making temac_init_mac_address() take a const void* address. Follows commit b34296a9 ("net: ethoc: Utilize of_get_mac_address()"). Cc: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Tobias Klauser authored
Do not open code getting the MAC address exclusively from the "local-mac-address" property, but instead use of_get_mac_address() which looks up the MAC address using the 3 typical property names. Also avoid casting away the const qualifier of the return value by making axienet_set_mac_address() take a const void* address. Follows commit b34296a9 ("net: ethoc: Utilize of_get_mac_address()"). Cc: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
David S. Miller authored
Or Gerlitz says: ==================== net/sched: cls_flower: Add support for matching on dissection flags This series add the UAPI to provide set of flags for matching, where the flags provided from user-space are mapped to flow-dissector flags. The 1st flag allows to match on whether the packet is an IP fragment and corresponds to the FLOW_DIS_IS_FRAGMENT flag. v2->v3: - replace BIT() with << (kbuild test robot) v1->v2: - dropped the flow dissector patch (#1) as no changes are needed there (Jiri) - applied code review comments from Jiri to the flower patch ==================== Signed-off-by: -
Or Gerlitz authored
Enable offloading of matching on packets being fragments. Signed-off-by:
Or Gerlitz <ogerlitz@mellanox.com> Reviewed-by:
Paul Blakey <paulb@mellanox.com> Reviewed-by:
-
Or Gerlitz authored
Add UAPI to provide set of flags for matching, where the flags provided from user-space are mapped to flow-dissector flags. The 1st flag allows to match on whether the packet is an IP fragment and corresponds to the FLOW_DIS_IS_FRAGMENT flag. Signed-off-by:
-
Dan Carpenter authored
These two statements were not indented correctly so it's sort of confusing. Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Dan Carpenter authored
We never set "slots" in this function. Fixes: a9380b0f ("drivers: net: xgene: Add support for Jumbo frame") Signed-off-by:
Iyappan Subramanian <isubramanian@apm.com> Signed-off-by:
-
Peng Tao authored
test_and_set_bit() already implies a memory barrier. Signed-off-by:
Peng Tao <bergwolf@gmail.com> Reviewed-by:
Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by:
-
Peng Tao authored
Signed-off-by:
-
Zhang Shengju authored
Currently, icmp_rcv() always return zero on a packet delivery upcall. To make its behavior more compliant with the way this API should be used, this patch changes this to let it return NET_RX_SUCCESS when the packet is proper handled, and NET_RX_DROP otherwise. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-nextDavid S. Miller authored
Pablo Neira Ayuso says: ==================== Netfilter/IPVS updates for net-next The following patchset contains a large Netfilter update for net-next, to summarise: 1) Add support for stateful objects. This series provides a nf_tables native alternative to the extended accounting infrastructure for nf_tables. Two initial stateful objects are supported: counters and quotas. Objects are identified by a user-defined name, you can fetch and reset them anytime. You can also use a maps to allow fast lookups using any arbitrary key combination. More info at: http://marc.info/?l=netfilter-devel&m=148029128323837&w=2 2) On-demand registration of nf_conntrack and defrag hooks per netns. Register nf_conntrack hooks if we have a stateful ruleset, ie. state-based filtering or NAT. The new nf_conntrack_default_on sysctl enables this from newly created netnamespaces. Default behaviour is not modified. Patches from Florian Westphal. 3) Allocate 4k chunks and then use these for x_tables counter allocation requests, this improves ruleset load time and also datapath ruleset evaluation, patches from Florian Westphal. 4) Add support for ebpf to the existing x_tables bpf extension. From Willem de Bruijn. 5) Update layer 4 checksum if any of the pseudoheader fields is updated. This provides a limited form of 1:1 stateless NAT that make sense in specific scenario, eg. load balancing. 6) Add support to flush sets in nf_tables. This series comes with a new set->ops->deactivate_one() indirection given that we have to walk over the list of set elements, then deactivate them one by one. The existing set->ops->deactivate() performs an element lookup that we don't need. 7) Two patches to avoid cloning packets, thus speed up packet forwarding via nft_fwd from ingress. From Florian Westphal. 8) Two IPVS patches via Simon Horman: Decrement ttl in all modes to prevent infinite loops, patch from Dwip Banerjee. And one minor refactoring from Gao feng. 9) Revisit recent log support for nf_tables netdev families: One patch to ensure that we correctly handle non-ethernet packets. Another patch to add missing logger definition for netdev. Patches from Liping Zhang. 10) Three patches for nft_fib, one to address insufficient register initialization and another to solve incorrect (although harmless) byteswap operation. Moreover update xt_rpfilter and nft_fib to match lbcast packets with zeronet as source, eg. DHCP Discover packets (0.0.0.0 -> 255.255.255.255). Also from Liping Zhang. 11) Built-in DCCP, SCTP and UDPlite conntrack and NAT support, from Davide Caratti. While DCCP is rather hopeless lately, and UDPlite has been broken in many-cast mode for some little time, let's give them a chance by placing them at the same level as other existing protocols. Thus, users don't explicitly have to modprobe support for this and NAT rules work for them. Some people point to the lack of support in SOHO Linux-based routers that make deployment of new protocols harder. I guess other middleboxes outthere on the Internet are also to blame. Anyway, let's see if this has any impact in the midrun. 12) Skip software SCTP software checksum calculation if the NIC comes with SCTP checksum offload support. From Davide Caratti. 13) Initial core factoring to prepare conversion to hook array. Three patches from Aaron Conole. 14) Gao Feng made a wrong conversion to switch in the xt_multiport extension in a patch coming in the previous batch. Fix it in this batch. 15) Get vmalloc call in sync with kmalloc flags to avoid a warning and likely OOM killer intervention from x_tables. From Marcelo Ricardo Leitner. 16) Update Arturo Borrero's email address in all source code headers. ==================== Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 40GbE Intel Wired LAN Driver Updates 2016-12-07 This series contains updates to i40e and i40evf only. Filip modifies the i40e to log link speed change and when the link is brought up and down. Mitch replaces i40e_txd_use_count() with a new function which is slightly faster and better documented so the dim witted can better follow the code. Fixes the locking of the service task so that it is actually done in the service task and not in the scheduling function which calls the service task. Jacob, being the busy little beaver he is, provides most of the changes starting restores a workaround that is still needed in some configurations, specifically the Ethernet Controller XL710 for 40GbE QSFP+. Removes duplicate code and simplifies the i40e_vsi_add_vlan() and i40e_vsi_kill_vlan() functions. Removes detection of PTP frames over L4 (UDP) on the XL710 MAC, since there was a product decision to defeature it. Fixed a previous refactor of active filters which caused issues in the accounting of active_filters. Remaining work was done in the VLAN filters to improve readability and simplify code as much as possible to reduce inconsistencies. Alex fixes foul budget accounting in core code by returning actual work done, capped to budget-1. Henry fixes the "ethtool -p" function for 1G BaseT PHYs. Carolyn adds support for 25G devices for i40e and i40evf. Michal adds functions to apply the correct access method for external PHYs which could use Clause22 or Clause45 depending on the PHY. v2: dropped last patch from previous series, since changes are needed based on feedback from Sergei Shtylyov ==================== Signed-off-by:
-
- 07 Dec, 2016 3 commits
-
-
Zhang Shengju authored
After commit 61e84623 ("net: centralize net_device min/max MTU checking"), the mtu range for dummy device becomes [68, 1500]. This patch extends it to [0, 65535]. Signed-off-by:
-
Zhang Shengju authored
Since commit 61e84623 ("net: centralize net_device min/max MTU checking"), mtu range is checked at dev_set_mtu(). This patch adds min_mtu for nlmon device and remove unnecessary ndo_change_mtu() function. Signed-off-by:
-
Gao Feng authored
When free macvlan_port in macvlan_port_destroy, it is safe to free directly because netdev_rx_handler_unregister could enforce one grace period. So it is unnecessary to use kfree_rcu for macvlan_port. Signed-off-by:
Gao Feng <fgao@ikuai8.com> Signed-off-by:
-
