Allain Legacy [Tue, 28 Mar 2017 11:54:08 +0000 (07:54 -0400)]
net/avp: handle interrupt migration
This commit introduces changes required to support VM live-migration. This
is done by registering and responding to interrupts coming from the host to
signal that the memory is about to be made invalid and replaced with a new
memory zone on the destination compute node.
Enabling and disabling of the interrupts are maintained outside of the
start/stop functions because they must be enabled for the lifetime of the
device. This is so that host interrupts are serviced and acked even in
cases where the app may have stopped the device.
Signed-off-by: Allain Legacy <allain.legacy@windriver.com> Signed-off-by: Matt Peters <matt.peters@windriver.com> Acked-by: Vincent Jardin <vincent.jardin@6wind.com>
Allain Legacy [Tue, 28 Mar 2017 11:54:07 +0000 (07:54 -0400)]
net/avp: add device start and stop operations
Adds support for device start and stop functions. This allows an
application to control the administrative state of an AVP device. Stopping
the device will notify the host application to stop sending packets on that
device's receive queues.
Signed-off-by: Allain Legacy <allain.legacy@windriver.com> Signed-off-by: Matt Peters <matt.peters@windriver.com> Acked-by: Vincent Jardin <vincent.jardin@6wind.com>
Allain Legacy [Tue, 28 Mar 2017 11:54:06 +0000 (07:54 -0400)]
net/avp: add device promiscuous functions
Adds support for setting and clearing promiscuous mode on an AVP device.
When enabled the _mac_filter function will allow packets destined to any
MAC address to be processed by the receive functions.
Signed-off-by: Allain Legacy <allain.legacy@windriver.com> Signed-off-by: Matt Peters <matt.peters@windriver.com> Acked-by: Vincent Jardin <vincent.jardin@6wind.com>
Allain Legacy [Tue, 28 Mar 2017 11:54:04 +0000 (07:54 -0400)]
net/avp: add packet transmit functions
Adds support for packet transmit functions so that an application can send
packets to the host application via an AVP device queue. Both the simple
and scattered functions are supported.
Signed-off-by: Allain Legacy <allain.legacy@windriver.com> Signed-off-by: Matt Peters <matt.peters@windriver.com> Acked-by: Vincent Jardin <vincent.jardin@6wind.com>
Allain Legacy [Tue, 28 Mar 2017 11:53:57 +0000 (07:53 -0400)]
net/avp: add public header files
Adds public/exported header files for the AVP PMD. The AVP device is a
shared memory based device. The structures and constants that define the
method of operation of the device must be visible by both the PMD and the
host DPDK application. They must not change without proper version
controls and updates to both the hypervisor DPDK application and the PMD.
The hypervisor DPDK application is a Wind River Systems proprietary
virtual switch.
Signed-off-by: Allain Legacy <allain.legacy@windriver.com> Signed-off-by: Matt Peters <matt.peters@windriver.com> Acked-by: Vincent Jardin <vincent.jardin@6wind.com>
Vasily Philipov [Tue, 28 Mar 2017 10:18:49 +0000 (13:18 +0300)]
net/mlx: enhance Rx scatter mode detection
Toggle Rx scatter mode based on the scatter_enable flag and the maximum
packet size only instead of deriving this information from the jumbo_frame
setting and the MTU configuration.
Wenzhuo Lu [Fri, 24 Feb 2017 03:24:29 +0000 (11:24 +0800)]
net/i40e: allocate VF TC bandwidth from PF
Allocate all TCs' relative bandwidth (percentage) on
a specific VF.
It can be taken as relative min bandwidth setting.
User can call the API to set VF TC's min bandwidth
from PF.
Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com> Acked-by: Jingjing Wu <jingjing.wu@intel.com>
Shijith Thotton [Sat, 25 Mar 2017 06:24:28 +0000 (11:54 +0530)]
net/liquidio: add APIs for response list
Add APIs to setup and process response list. Response list holds soft
commands waiting for response from device. Entries of this list are
processed to check for command response or timeout.
Shijith Thotton [Sat, 25 Mar 2017 06:24:24 +0000 (11:54 +0530)]
net/liquidio: add APIs to allocate and free IQ
Instruction queue (IQ) is used to send control and data packets to
device from host. IQ 0 is used to send device configuration commands
during initialization and later re-allocated as per application
requirement.
Wenzhuo Lu [Fri, 24 Mar 2017 02:51:03 +0000 (10:51 +0800)]
net/ixgbe: fix TC bandwidth setting
4 and 8 TCs are supported on ixgbe. By default there're
8 TCs. So when initializing the device, the bandwidth for
8 TCs is set.
When changing the TC number, it's only considered setting
the bandwidth for 4 TCs. If the user change the number
from 4 to 8, the TCs' bandwidth is not right.
Pascal Mazon [Wed, 22 Mar 2017 08:40:01 +0000 (09:40 +0100)]
net/tap: add link status notification
As tap is a virtual device, there's no physical way a link can be cut.
However, it has an associated kernel netdevice and possibly a remote
netdevice too. These netdevices link status may change outside of the
DPDK scope, through an external command such as:
ip link set dev tapX down
This commit implements link status notification through netlink.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Pascal Mazon [Wed, 22 Mar 2017 08:40:00 +0000 (09:40 +0100)]
net/tap: improve link update
Reflect device link status according to the state of the tap netdevice
and the remote netdevice (if any). If both are UP and RUNNING, then the
device link status is set to ETH_LINK_UP, otherwise ETH_LINK_DOWN.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Pascal Mazon [Thu, 23 Mar 2017 08:42:11 +0000 (09:42 +0100)]
net/tap: add remote netdevice traffic capture
By default, a tap netdevice is of no use when not fed by a separate
process. The ability to automatically feed it from another netdevice
allows applications to capture any kind of traffic normally destined to
the kernel stack.
This patch implements this ability through a new optional "remote"
parameter.
Packets matching filtering rules created with the flow API are matched
on the remote device and redirected to the tap PMD, where the relevant
action will be performed.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Olga Shern <olgas@mellanox.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Pascal Mazon [Thu, 16 Mar 2017 08:59:21 +0000 (09:59 +0100)]
net/tap: support segmented mbufs
Support for segmented packets (scatter/gather) is mandatory for most
purposes, regardless of the MTU size. Tx packets are often the result of
mbuf concatenation, and an mbuf is not necessarily large enough for Rx
packets to fit in a single one.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Jerin Jacob [Mon, 20 Mar 2017 14:10:40 +0000 (19:40 +0530)]
net/thunderx: sync mailbox definitions with Linux PF driver
- bgx_link_status mbox definition was changed in Linux
commit 1cc702591bae ("net: thunderx: Add ethtool support")
- NIC_MBOX_MSG_RES_BIT related changes were never part of Linux PF driver
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Pascal Mazon [Thu, 23 Mar 2017 08:33:57 +0000 (09:33 +0100)]
net/tap: add basic flow API patterns and actions
Supported flow rules are now mapped to TC rules on the tap netdevice.
The netlink message used for creating the TC rule is stored in struct
rte_flow. That way, by simply changing a metadata in it, we can require
for the rule deletion without further parsing.
Supported items:
- eth: src and dst (with variable masks), and eth_type (0xffff mask).
- vlan: vid, pcp, tpid, but not eid.
- ipv4/6: src and dst (with variable masks), and ip_proto (0xffff mask).
- udp/tcp: src and dst port (0xffff) mask.
Supported actions:
- DROP
- QUEUE
- PASSTHRU
It is generally not possible to provide a "last" item. However, if the
"last" item, once masked, is identical to the masked spec, then it is
supported.
Only IPv4/6 and MAC addresses can use a variable mask. All other
items need a full mask (exact match).
Support for VLAN requires kernel headers >= 4.9, checked using
auto-config.sh.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Olga Shern <olgas@mellanox.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Pascal Mazon [Thu, 23 Mar 2017 08:33:56 +0000 (09:33 +0100)]
net/tap: add netlink back-end for flow API
Each kernel netdevice may have queueing disciplines set for it, which
determine how to handle the packet (mostly on egress). That's part of
the TC (Traffic Control) mechanism.
Through TC, it is possible to set filter rules that match specific
packets, and act according to what is in the rule. This is a perfect
candidate to implement the flow API for the tap PMD, as it has an
associated kernel netdevice automatically.
Each flow API rule will be translated into its TC counterpart.
To leverage TC, it is necessary to communicate with the kernel using
netlink. This patch introduces a library to help that communication.
Inside netlink.c, functions are generic for any netlink messaging.
Inside tcmsgs.c, functions are specific to deal with TC rules.
Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com> Acked-by: Olga Shern <olgas@mellanox.com> Acked-by: Keith Wiles <keith.wiles@intel.com>
Yongseok Koh [Tue, 21 Mar 2017 17:50:51 +0000 (10:50 -0700)]
net/mlx5: fix reusing Rx/Tx queues
When configuring Rx/Tx queue, if queue already exists, it is reused. But if
the queue size is changed, it must be resized to not access/overwrite
invalid memory.
Fixes: 2e22920b85d9 ("mlx5: support non-scattered Tx and Rx") Cc: stable@dpdk.org Signed-off-by: Yongseok Koh <yskoh@mellanox.com> Acked-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Jingjing Wu [Wed, 22 Mar 2017 09:24:59 +0000 (17:24 +0800)]
net/i40e/base: new AQ commands for cloud filter
Add new admin queue function and extended fields for cloud filter:
- Add admin queue function for Replace filter command (Opcode: 0x025F)
- Define big buffer for extended general fields in Add/Remove
Cloud filters command
Signed-off-by: Laura Stroe <laura.stroe@intel.com> Signed-off-by: Bernard Iremonger <bernard.iremonger@intel.com> Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
Jingjing Wu [Wed, 22 Mar 2017 09:24:58 +0000 (17:24 +0800)]
net/i40e/base: add VF offload flags
This patch adds:
- ENCAP offload negotiation flag. Use the existing ENCAP_CSUM offload
flag to negotiate GSO_UDP_TUNNEL_CSUM capability and create new ENCAP
flag for negotiating offloads for encapsulated packets
- RX_ENCAP_CSUM offload negotiation flag for VF to negotiate RX
checksum capability for tunnelled packet types.
Jingjing Wu [Wed, 22 Mar 2017 09:24:57 +0000 (17:24 +0800)]
net/i40e/base: reduce wait time for adminq command
When sending an adminq command, we wait for the command to complete in
a loop. This loop waits for an entire millisecond, when in practice the
adminq command is processed often much faster.
Change the loop to use i40e_usec_delay instead, and wait for 50 usecs
each time instead. This appears to be about the minimum time required,
based on some manual observation and testing.
The primary benefit of this change is reducing latency of various
operations in the PF driver, especially when related to having a large
number of VFs enabled.
Jingjing Wu [Wed, 22 Mar 2017 09:24:55 +0000 (17:24 +0800)]
net/i40e/base: fix potential out of bound array access
This is fix for klocwork issue where dcbcfg->numapps could
be greater than size of array (i.e dcbcfg->app[I40E_DCBX_MAX_APPS]).
The fix makes sure the array is not accessed past size of array
(i.e. I40E_DCBX_MAX_APPS).
Keith Wiles [Tue, 21 Mar 2017 15:12:12 +0000 (10:12 -0500)]
net/bonding: reduce slave starvation on Rx poll
When polling the bonded ports for RX packets the old driver would
always start with the first slave in the list. If the requested
number of packets is filled on the first port in a two port config
then the second port could be starved or have larger number of
missed packet errors.
The code attempts to start with a different slave each time RX poll
is done to help eliminate starvation of slave ports. The effect of
the previous code was much lower performance for two slaves in the
bond then just the one slave.
The performance drop was detected when the application can not poll
the rings of Rx packets fast enough and the packets per second for
two or more ports was at the threshold throughput of the application.
At this threshold the slaves would see very little or no drops in
the case of one slave. Then enable the second slave you would see
a large drop rate on the two slave bond and reduction in throughput.
Signed-off-by: Keith Wiles <keith.wiles@intel.com> Acked-by: Declan Doherty <declan.doherty@intel.com>
Alejandro Lucero [Tue, 21 Mar 2017 10:43:20 +0000 (10:43 +0000)]
net/nfp: fix packet/data length conversion
Chained mbufs hold data_len as the length of that particular mbuf
and pkt_len as the full packet length including all the chained
mbufs. It is not clear from the mbuf definition if pkt_len should
be set for all the mbufs in a chain, but code there for handling
mbufs suggests just the first mbuf requires to have pkt_len set.
NFP PMD was assuming pkt_len is set in all the chained mbufs and
unit tests for gather dma were building mbufs with pkt_len always
set. This patch gets rid of that assumption.
git.droids-corp.org / dpdk.git / log