- 06 Mar, 2015 32 commits
-
-
Fan Du authored
As per RFC4821 7.3. Selecting Probe Size, a probe timer should be armed once probing has converged. Once this timer expired, probing again to take advantage of any path PMTU change. The recommended probing interval is 10 minutes per RFC1981. Probing interval could be sysctled by sysctl_tcp_probe_interval. Eric Dumazet suggested to implement pseudo timer based on 32bits jiffies tcp_time_stamp instead of using classic timer for such rare event. Signed-off-by:
Fan Du <fan.du@intel.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Fan Du authored
Current probe_size is chosen by doubling mss_cache, the probing process will end shortly with a sub-optimal mss size, and the link mtu will not be taken full advantage of, in return, this will make user to tweak tcp_base_mss with care. Use binary search to choose probe_size in a fine granularity manner, an optimal mss will be found to boost performance as its maxmium. In addition, introduce a sysctl_tcp_probe_threshold to control when probing will stop in respect to the width of search range. Test env: Docker instance with vxlan encapuslation(82599EB) iperf -c 10.0.0.24 -t 60 before this patch: 1.26 Gbits/sec After this patch: increase 26% 1.59 Gbits/sec Signed-off-by:
John Heffner <johnwheffner@gmail.com> Signed-off-by:
-
Fan Du authored
Quotes from RFC4821 7.2. Selecting Initial Values It is RECOMMENDED that search_low be initially set to an MTU size that is likely to work over a very wide range of environments. Given today's technologies, a value of 1024 bytes is probably safe enough. The initial value for search_low SHOULD be configurable. Moreover, set a small value will introduce extra time for the search to converge. So set the initial probe base mss size to 1024 Bytes. Signed-off-by:
-
Eric W. Biederman authored
Other users users of the neighbour table use neigh->output as the method to decided when and which link-layer header to place on a packet. DECnet has been using neigh->output to decide which DECnet headers to place on a packet depending which neighbour the packet is destined for. The DECnet usage isn't totally wrong but it can run into problems if the neighbour output function is run for a second time as the teql driver and the bridge netfilter code can do. Therefore to avoid pathologic problems later down the line and make the neighbour code easier to understand by refactoring the decnet output code to only use a neighbour method to add a link layer header to a packet. This is done by moving the neigbhour operations lookup from dn_to_neigh_output to dn_neigh_output_packet. Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Mahesh Bandewar authored
This patches implements the poll_controller support for all bonding driver. If the slaves have poll_controller net_op defined, this implementation calls them. This is mode agnostic implementation and iterates through all slaves (based on mode) and calls respective handler. Signed-off-by:
Mahesh Bandewar <maheshb@google.com> Signed-off-by:
-
Scott Feldman authored
Signed-off-by:
Scott Feldman <sfeldma@gmail.com> Acked-by:
Jiri Pirko <jiri@resnulli.us> Signed-off-by:
-
Scott Feldman authored
Signed-off-by:
-
David S. Miller authored
net/ipv4/fib_trie.c: In function ‘fib_table_flush_external’: net/ipv4/fib_trie.c:1572:6: warning: unused variable ‘found’ [-Wunused-variable] int found = 0; ^ net/ipv4/fib_trie.c:1571:16: warning: unused variable ‘slen’ [-Wunused-variable] unsigned char slen; ^ Signed-off-by: -
David S. Miller authored
Scott Feldman says: ==================== switchdev: add IPv4 routing offload v4: - Add NETIF_F_NETNS_LOCAL to rocker port feature list to keep rocker ports in the default netns. Rocker hardware can't be partitioned to support multiple namespaces, currently. It would be interesting to add netns support to rocker device by basically adding another match field to each table to match on some unique netns ID, with a port knowing it's netns ID. Future work TDB. - Up-level the RTNH_F_EXTERNAL marking of routes installed to offload device from driver to switchdev common code. Now driver can't skip routes. Either it can install the route or it cannot. Yes or No. If no on any route, all offloading is aborted by removing routes from offload device and setting ipv4.fib_offload_disabled so no more routes can be offloaded. This is harsh, but it's our starting point. We can refine the policies in follow-up work. - Add new net.ipv4.fib_offload_disabled bool that is set if anything goes wrong with route offloading. We can refine this later to make the setting per-device or per-device-port-netdev, but let's start here simple and refine in follow-up work. - Rebase against Alex's latest FIB changes. I think I did everything correctly, and didn't run into any issues with testing, but I'd like Alex to look over the changes and maybe follow-up with any cleanups. v3: Changes based on v2 review comments: - Move check for custom rules up earlier in patch set, to keep git bisect safe. - Simplify the route add/modify failure handling to simple try until failure, and then on failure, undo everything. The switchdev driver will return err when route can normally be installed to device, but the install fails for one reason or another (no space left on device, etc). If a failure happens, uninstall all routes from the device, punting forwarding for all routes back to the kernel. - Scan route's full nexthop list, ensuring all nexthop devs belong to the same switchdev device, otherwise don't try to install route to device. v2: Changes based on v1 review comments and discussions at netconf: - Allow route modification, but use same ndo op used for adding route. Driver/device is expected to modify route in-place, if it can, to avoid interruption of service. - Add new RTNH_F_EXTERNAL flag to mark FIB entries offloaded externally. - Don't offload routes if using custom IP rules. If routes are already offloaded, and custom IP rules are turned on, flush routes from offload device. (Offloaded routes are marked with RTNH_F_EXTERNAL). - Use kernel's neigh resolution code to resolve route's nexthops' neigh MAC addrs. (Thanks davem, works great!). - Use fib->fib_priority in rocker driver to give priorities to routes in OF-DPA unicast route table. v1: This patch set adds L3 routing offload support for IPv4 routes. The idea is to mirror routes installed in the kernel's FIB down to a hardware switch device to offload the data forwarding path for L3. Only the data forwarding path is intercepted. Control and management of the kernel's FIB remains with the kernel. ==================== Signed-off-by: -
Scott Feldman authored
The driver implements ndo_switch_fib_ipv4_add/del ops to add/del/mod IPv4 routes to/from switchdev device. Once a route is added to the device, and the route's nexthops are resolved to neighbor MAC address, the device will forward matching pkts rather than the kernel. This offloads the L3 forwarding path from the kernel to the device. Note that control and management planes are still mananged by Linux; only the data plane is offloaded. Standard routing control protocols such as OSPF and BGP run on Linux and manage the kernel's FIB via standard rtm netlink msgs...nothing changes here. A new hash table is added to rocker to track neighbors. The driver listens for neighbor updates events using netevent notifier NETEVENT_NEIGH_UPDATE. Any ARP table updates for ports on this device are recorded in this table. Routes installed to the device with nexthops that reference neighbors in this table are "qualified". In the case of a route with nexthops not resolved in the table, the kernel is asked to resolve the nexthop. The driver uses fib_info->fib_priority for the priority field in rocker's unicast routing table. The device can only forward to pkts matching route dst to resolved nexthops. Currently, the device only supports single-path routes (i.e. routes with one nexthop). Equal Cost Multipath (ECMP) route support will be added in followup patches. This patch is driver support for unicast IPv4 routing only. Followup patches will add driver and infrastructure for IPv6 routing and multicast routing. Signed-off-by:
-
Scott Feldman authored
Call into the switchdev driver any time an IPv4 fib entry is added/modified/deleted from the kernel's FIB. The switchdev driver may or may not install the route to the offload device. In the case where the driver tries to install the route and something goes wrong (device's routing table is full, etc), then all of the offloaded routes will be flushed from the device, route forwarding falls back to the kernel, and no more routes are offloading. We can refine this logic later. For now, use the simplist model of offloading routes up to the point of failure, and then on failure, undo everything and mark IPv4 offloading disabled. Signed-off-by:
-
Scott Feldman authored
If something goes wrong with IPv4 FIB offload, mark entire net offload disabled. This is brute force policy to basically shut down IPv4 FIB offload permanently if there is a problem offloading any route to an external device. We can refine the policy in the future, to handle failures on a per-device or per-route basis, but for now, this policy is per-net. What we're trying to avoid is an inconsistent split between the kernel's FIB and the offload device's FIB. We don't want the device to fwd a pkt inconsitent with what the kernel would do. An example of a split is if device has 10.0.0.0/16 and kernel has 10.0.0.0/16 and 10.0.0.0/24, the device wouldn't see the longest prefix 10.0.0.0/24 and potentially forward pkts incorrectly. Limited capacity or limited capability are two ways a route may fail to install to the offload device. We'll not differentiate between failures at this time, and treat any failure as fatal and mark the net as fib_offload_disabled. Signed-off-by:
-
Scott Feldman authored
Flesh out ndo wrappers to call into device driver. To call into device driver, the wrapper must interate over route's nexthops to ensure all nexthop devs belong to the same switch device. Currently, there is no support for route's nexthops spanning offloaded and non-offloaded devices, or spanning ports of multiple offload devices. Since switch device ports may be stacked under virtual interfaces (bonds and/or bridges), and the route's nexthop may be on the virtual interface, the wrapper will traverse the nexthop dev down to the base dev. It's the base dev that's passed to the switchdev driver's ndo ops. Signed-off-by:
-
Scott Feldman authored
Keep switchdev FIB offload model simple for now and don't allow custom ip rules. Signed-off-by:
-
Scott Feldman authored
Add IPv4 fib ndo wrapper funcs and stub them out for now. Signed-off-by:
-
Scott Feldman authored
Add two new ndo ops for IPv4 fib offload support, add and del. Add uses modifiy semantics if fib entry already offloaded. Drivers implementing the new ndo ops will return err<0 if programming device fails, for example if device's tables are full. Signed-off-by:
-
Scott Feldman authored
Add new RTNH_F_EXTERNAL flag to mark fib entries offloaded externally, for example to a switchdev switch device. Signed-off-by:
-
Eric Dumazet authored
Using napi_complete_done() instead of napi_complete() allows us to use /sys/class/net/ethX/gro_flush_timeout GRO layer can aggregate more packets if the flush is delayed a bit, without having to set too big coalescing parameters that impact latencies. Tested: lpx:~# echo 0 >/sys/class/net/eth1/gro_flush_timeout lpx:~# sar -n DEV 1 10 | grep eth1 10:36:25 AM eth1 81290.00 40617.00 120479.67 2777.01 0.00 0.00 0.00 10:36:26 AM eth1 81283.00 40608.00 120481.81 2778.13 0.00 0.00 1.00 10:36:27 AM eth1 81304.00 40639.00 120518.42 2778.28 0.00 0.00 0.00 10:36:28 AM eth1 81255.00 40605.00 120437.34 2775.95 0.00 0.00 1.00 10:36:29 AM eth1 81306.00 40630.00 120521.44 2777.70 0.00 0.00 0.00 10:36:30 AM eth1 81286.00 40564.00 120480.20 2773.31 0.00 0.00 0.00 10:36:31 AM eth1 81256.00 40599.00 120438.81 2776.27 0.00 0.00 0.00 10:36:32 AM eth1 81287.00 40594.00 120480.69 2776.69 0.00 0.00 0.00 10:36:33 AM eth1 81279.00 40601.00 120478.53 2775.84 0.00 0.00 0.00 10:36:34 AM eth1 81277.00 40610.00 120476.94 2776.25 0.00 0.00 0.00 Average: eth1 81282.30 40606.70 120479.39 2776.54 0.00 0.00 0.20 lpx:~# echo 13000 >/sys/class/net/eth1/gro_flush_timeout lpx:~# sar -n DEV 1 10 | grep eth1 10:36:43 AM eth1 81257.00 7747.00 120437.44 530.00 0.00 0.00 0.00 10:36:44 AM eth1 81278.00 7748.00 120480.00 529.85 0.00 0.00 0.00 10:36:45 AM eth1 81282.00 7752.00 120479.09 531.09 0.00 0.00 0.00 10:36:46 AM eth1 81282.00 7751.00 120478.80 530.90 0.00 0.00 0.00 10:36:47 AM eth1 81276.00 7745.00 120478.31 529.64 0.00 0.00 0.00 10:36:48 AM eth1 81278.00 7747.00 120478.50 529.81 0.00 0.00 0.00 10:36:49 AM eth1 81282.00 7749.00 120478.88 530.01 0.00 0.00 0.00 10:36:50 AM eth1 81284.00 7751.00 120481.52 530.20 0.00 0.00 0.00 10:36:51 AM eth1 81299.00 7769.00 120481.74 533.81 0.00 0.00 0.00 10:36:52 AM eth1 81281.00 7748.00 120478.62 529.96 0.00 0.00 0.00 Average: eth1 81279.90 7750.70 120475.29 530.53 0.00 0.00 0.00 Signed-off-by:
Eric Dumazet <edumazet@google.com> Acked-by:
Michael Chan <mchan@broadcom.com> Signed-off-by:
-
David S. Miller authored
Florian Fainelli says: ==================== net: dsa: code re-organization This pull request contains the first part of the patches required to implement the grand plan detailed here: http://www.spinics.net/lists/netdev/msg295942.html These are mostly code re-organization and function bodies re-arrangement to allow different callers of lower-level initialization functions for 'struct dsa_switch' and 'struct dsa_switch_tree' to be later introduced. There is no functional code change at this point. ==================== Signed-off-by:
-
Florian Fainelli authored
Extract the core logic that setups a 'struct dsa_switch_tree' and removes it, update dsa_probe() and dsa_remove() to use the two helper functions. This will be useful to allow for other callers to setup this structure differently. Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Florian Fainelli authored
In order to support the new DSA device driver model, a dsa_switch should be able to advertise the type of tagging protocol supported by the underlying switch device. This also removes constraints on how tagging can be stacked to each other. Signed-off-by:
-
Florian Fainelli authored
Split the part of dsa_switch_setup() which is responsible for allocating and initializing a 'struct dsa_switch' and the part which is doing a given switch device setup and slave network device creation. This is a preliminary change to allow a separate caller of dsa_switch_setup_one() which may have externally initialized the dsa_switch structure, outside of dsa_switch_setup(). Signed-off-by:
-
Florian Fainelli authored
In preparation for allowing a different model to register DSA switches, update dsa_of_probe() and dsa_probe() to return -EPROBE_DEFER where appropriate. Failure to find a phandle or Device Tree property is still fatal, but looking up the internal device structure associated with a Device Tree node is something that might need to be delayed based on driver probe ordering. Signed-off-by:
-
Florian Fainelli authored
In preparation for allowing a different mechanism to register DSA switch devices and driver, update dsa_of_probe and dsa_of_remove to take a struct device pointer since neither of these two functions uses the struct platform_device pointer. Signed-off-by:
-
Eric Dumazet authored
timewait sockets now share a common base with established sockets. inet_twsk_diag_dump() can use inet_diag_bc_sk() instead of duplicating code, granted that inet_diag_bc_sk() does proper userlocks initialization. twsk_build_assert() will catch any future changes that could break the assumptions. Signed-off-by:
-
Jeff Kirsher authored
As pointed out by Ben Hutchings, ioremap uses unsigned long as its parameter type, so we should be using that instead of u32 or int. Reported-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Erik Hugne authored
The ip/udp bearer can be configured in a point-to-point mode by specifying both local and remote ip/hostname, or it can be enabled in multicast mode, where links are established to all tipc nodes that have joined the same multicast group. The multicast IP address is generated based on the TIPC network ID, but can be overridden by using another multicast address as remote ip. Signed-off-by:
Erik Hugne <erik.hugne@ericsson.com> Reviewed-by:
Jon Maloy <jon.maloy@ericsson.com> Reviewed-by:
Ying Xue <ying.xue@windriver.com> Signed-off-by:
-
Erik Hugne authored
The payload area following the TIPC discovery message header is an opaque area defined by the media. INT_H_SIZE was enough for Ethernet/IB/IPv4 but needs to be expanded to carry IPv6 addressing information. Signed-off-by:
-
WANG Cong authored
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by:
Cong Wang <xiyou.wangcong@gmail.com> Acked-by:
Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by:
-
WANG Cong authored
Cc: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-nextDavid S. Miller authored
Jeff Kirsher says: ==================== This series contains updates to i40e only. Greg provides fixes for the NPAR transmit scheduler where the driver initialization caused the BW configurations to not take effect, so use a BW configuration read and write back to "kick" the transmit scheduler into action. Fixes the ethtool offline test, where we were not actually taking the device offline before doing the testing. Matt modifies the get and set LED functions so they ignore activity LEDs since we are required to blink the link LEDs only. Neerav provides a workaround for whenever a DCBX configuration is changed, where the firmware doe not set the operational status bit of the application TLV status as returned from the "Get CEE DCBX Oper Cfg" admin queue command. So remove the check for the operational and sync bits of the application TLV status until a firmware fix is provided. Shannon changes the driver to grab the NVM devstarter version and not the image version, since it is the more useful version and is what should be displayed. Moves the IRQ tracking setup and tear down into the same routines that do the IRQ setup and tear down. This keeps like activities together and allows us to track exactly the number of vectors reserved from the OS, which may be fewer than are available from the hardware. Jesse provides a fix to use a more portable sign extension by replacing 0xffff.... with ~(u64)0 or ~(u32)0. Also fixes XPS mask when resetting, where the driver would accidentally clear the XPS mask for all queues back to 0. This caused higher CPU utilization and had some other performance impacts for transmit tests. Cleans up some whitespace formatting. Catherine provides a fix where some firmware versions are incorrectly reporting a breakout cable as PHY type 0x3 when it should be 0x16 (I40E_PHY_TYPE_10GBASE_SFPP_CU). Adds the 10G and 40G AOC PHY types to the case statement in get_media_type and ethtool get_settings so that the correct information gets reported back to the user. Anjali provides IOREMAP changes for future device support, where we do not want to map the whole CSR space since some of it is mapped by other drivers with different mapping methods. Mitch changes the i40e driver to not "spam" the system log with messages about VF VSI when VFs are created and when they are reset to reduce user annoyance. ==================== Signed-off-by: -
Stephen Rothwell authored
Fixes this build error: net/mpls/af_mpls.c: In function 'resize_platform_label_table': net/mpls/af_mpls.c:767:4: error: implicit declaration of function 'vzalloc' [-Werror=implicit-function-declaration] labels = vzalloc(size); ^ Fixes: 7720c01f ("mpls: Add a sysctl to control the size of the mpls label table") Signed-off-by:Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
-
- 05 Mar, 2015 8 commits
-
-
David S. Miller authored
Hariprasad Shenai says: ==================== cxgb4: RX Queue related cleanup and fixes This patch series adds a common function to allocate RX queues and queue allocation changes to RDMA CIQ The patches series is created against 'net-next' tree. And includes patches on cxgb4 driver. We have included all the maintainers of respective drivers. Kindly review the change and let us know in case of any review comments. ==================== Signed-off-by: -
Hariprasad Shenai authored
To allow for better scalability on systems with large core counts, we will try and allocate enough RDMA Concentrator IQs and MSI/X vectors as we have cores. If we cannot get enough MSI/X vectors, fall back to the minimum required: 1 per adapter rx channel. Also clean up cxgb_enable_msix() to make it readable and correct a bug where the vectors are not correctly assigned if the driver doesn't get the full amount requested. Signed-off-by:
Steve Wise <swise@opengridcomputing.com> Signed-off-by:
Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by:
-
Hariprasad Shenai authored
Adds a common function for all Rx queue allocation. Signed-off-by:
-
Jouni Malinen authored
This extends the design in commit 95850116 ("bridge: Add support for IEEE 802.11 Proxy ARP") with optional set of rules that are needed to meet the IEEE 802.11 and Hotspot 2.0 requirements for ProxyARP. The previously added BR_PROXYARP behavior is left as-is and a new BR_PROXYARP_WIFI alternative is added so that this behavior can be configured from user space when required. In addition, this enables proxyarp functionality for unicast ARP requests for both BR_PROXYARP and BR_PROXYARP_WIFI since it is possible to use unicast as well as broadcast for these frames. The key differences in functionality: BR_PROXYARP: - uses the flag on the bridge port on which the request frame was received to determine whether to reply - block bridge port flooding completely on ports that enable proxy ARP BR_PROXYARP_WIFI: - uses the flag on the bridge port to which the target device of the request belongs - block bridge port flooding selectively based on whether the proxyarp functionality replied Signed-off-by:
Jouni Malinen <jouni@codeaurora.org> Signed-off-by:
-
kbuild test robot authored
> > >> net/ax25/ax25_ip.c:225:26: error: unknown type name 'sturct' > netdev_tx_t ax25_ip_xmit(sturct sk_buff *skb) > ^ > > vim +/sturct +225 net/ax25/ax25_ip.c > > 219 unsigned short type, const void *daddr, > 220 const void *saddr, unsigned int len) > 221 { > 222 return -AX25_HEADER_LEN; > 223 } > 224 > > 225 netdev_tx_t ax25_ip_xmit(sturct sk_buff *skb) > 226 { > 227 kfree_skb(skb); > 228 return NETDEV_TX_OK; Ooops I misspelled struct... Signed-off-by: -
Sravanthi Tangeda authored
Bump i40e to 1.2.11 and i40evf to 1.2.5 Change-ID: Ie13375941606b0a027e5b5dbc235f5f5f03b75c8 Signed-off-by:
Sravanthi Tangeda <sravanthi.tangeda@intel.com> Tested-by:
Jim Young <james.m.young@intel.com> Signed-off-by:
-
Mitch A Williams authored
The PF driver spams the system log with messages about VF VSI when VFs are created, as well as each time they are reset. This is annoying, and the information isn't even useful most of the time. Remove this message to reduce user annoyance. Change-ID: I8de90d05380f54b038c9c8c3265150be87c9242c Signed-off-by:
Mitch Williams <mitch.a.williams@intel.com> Tested-by:
-
Shannon Nelson authored
Move the IRQ tracking setup and teardown into the same routines that do the IRQ setup and teardown. This keeps like activities together and allows us to track exactly the number of vectors reserved from the OS, which may be fewer than are available from the HW. Change-ID: I6b2b1a955c5f0ac6b94c3084304ed0b2ea6777cf Signed-off-by:
Shannon Nelson <shannon.nelson@intel.com> Signed-off-by:
-
