This patch will temporary add the "dpaa2" driver support in VPP in-built DPDK v16.07
Change-Id: Ieb7005c73bdb72e5eac88e17c21474479ad087f0
Signed-off-by: Sachin <sachin.saxena@nxp.com>
There are now two 0007... patch files. So renaming one to 0008...
Change-Id: I82ab883db3436b1fce6b169065b257df052fe63c
Signed-off-by: John Lo <loj@cisco.com>
This crash happens because the device still has MSI configured,
the fix is to free the IRQ. Adding pending further investigation.
Change-Id: Ifcd3404fcc25fe27dcccc67d0dcf9d8f0ec18582
Signed-off-by: Todd Foggoa (tfoggoa) <tfoggoa@cisco.com>
Signed-off-by: Ray Kinsella <ray.kinsella@intel.com>
Patch to prevent the unthrottled message from spilling to syslog and
fill up disk space.
Change-Id: I239ca27a7568182217a675692e0be966fc3486f5
Signed-off-by: John Lo <loj@cisco.com>
Some applications may wish to define their own implentation of
usec delay other than the existing blocking one. The default
behavior remains unchanged.
Change-Id: I6cb626e7a5d9d985928aa266e522a849d264a2b2
Signed-off-by: Todd Foggoa (tfoggoa) <tfoggoa@cisco.com>
Signed-off-by: Ray Kinsella <ray.kinsella@intel.com>
Ensure the packet_type is reset before the buffer is used.
This can cause packets to be mishandled in systems with more than one type of driver in use.
Change-Id: I14ef9bd6be834d66d19bae36122b648450a9f7e6
Signed-off-by: Ray Kinsella <ray.kinsella@intel.com>
Signed-off-by: Todd Foggoa (tfoggoa) <tfoggoa@cisco.com>
Commit d9a2009a81089093645fea2e04b51dd37edf3e6f on DPDK.org, reverted the ptype functionality
from the ixgbe vPMD. Reverting this change pending a deeper analysis.
Change-Id: I2515f3d700b2a4922495ba9e63f202d679543d01
Signed-off-by: Ray <ray.kinsella@intel.com>
The PKT_RX_VLAN_PKT flag will be totally obsolete by 16.11. We still need
this ENIC driver patch for 16.07.
Change-Id: I287035525d2d1b5313e88e9e74b6310536de2c32
Signed-off-by: John Lo <loj@cisco.com>
Change the default DPDK version to 16.07, and rename the indicated
patch directory. Use the native vhost-user driver.
Change-Id: Ie3d17e90e363ce86f0233b58c152de683b5d9456
Signed-off-by: Dave Barach <dave@barachs.net>
Author: Nelson Escobar <neescoba@cisco.com>
Date: Mon Jun 13 17:24:41 2016 -0700
We did not properly set the rq pointers we pass up to dpdk when
rx scatter was introduced. This resulted in segfaults whenever
more than one rq was being used.
Change-Id: I77a9793ea9035baa1ada77d3af7bc83baf58d097
Signed-off-by: Shesha Sreenivasamurthy <shesha@cisco.com>
DPDK vhost-user support is disabled due to significan changes
in the DPDK vhost-user code which are not compatible with current
VPP code.
Change-Id: I3f0d28cb75f6370282ec7e33d57cbfb77e1a3ce1
Signed-off-by: Damjan Marion <damarion@cisco.com>
When ENIC ports are used as slaves to a bonded interface, these patches
enable traffic on bonded interface to resume properly on admin down and
then up of the bonded interface.
Change-Id: I339926010da614068417c72009b7af2f892890ce
Signed-off-by: John Lo <loj@cisco.com>
Also clear x-bit for old patch files 0013-xxx and 0014-xxx.
Change-Id: I217fdbfb32cef1ae575c668270d3baf593e688c6
Signed-off-by: John Lo <loj@cisco.com>
This change should improve DPDK performance on 40G+ NICs
with small packets.
It also removes ITR setting for i40e devices, allowing driver to set default.
Change-Id: I70761b155e48fb0281f7c231516d83027bd16ca2
Signed-off-by: Damjan Marion <damarion@cisco.com>
Upstreaming of DPAA2 driver changes is in progress.This patch will
temporary add the support in VPP in built DPDK.
Two types of changes:
1. Driver specfic independent files. No impact on any other functionality.
2. Changes in common EAL framework. These changes are done in compile time DPAA2
specific flag, so no impact is expected on other existing features if not
compiling for DPAA2.
Change-Id: I02abe7189313835b51ff654b4d7e566bc0fb8327
Signed-off-by: Sachin <sachin.saxena@nxp.com>
The tarball repository fast.dpdk.org is fast and reliable enough
to be used in VPP build process.
Change-Id: Ifaae57d6f8308127b93fc51b2a2a863da5766cd2
Signed-off-by: Thomas Monjalon <thomas.monjalon@6wind.com>
Check packet for presence of VLAN tag in the MAC header and set
the PKT_RX_VLAN_PKT flag to trigger sub-interface lookup in VPP
ethernet-input node.
Change-Id: Iadbfc59d08ef85efb936c88323a90194f3cd656f
Signed-off-by: John Lo <loj@cisco.com>
* compiler -march= parameter is changed from native to corei7
so code is always genereted with instructions which are available
on the Nehalem microarchitecture (up to SSE4.2)
* compiler -mtune= parameter is added so code is optimized for
corei7-avx which equals to Sandy Bridge microarchitecture
* set of macros is added which allows run-time detection of available
cpu instructions (e.g. clib_cpu_supports_avx())
* set of macros is added which allows us to clone graph node funcitons
where cloned function is optmized for different microarchitecture
Those macros are using following attributes:
__attribute__((flatten))
__attribute__((target("arch=core-avx2)))
I.e. If applied to foo_node_fn() macro will generate cloned
functions foo_node_fn_avx2() and foo_node_fn_avx512() (future)
It will also generate function void * foo_node_fn_multiarch_select()
which detects available instruction set and returns pointer to the
best matching function clone.
Change-Id: I2dce0ac92a5ede95fcb56f47f3d1f3c4c040bac0
Signed-off-by: Damjan Marion <damarion@cisco.com>
VPP still needs to use packet type recognition in vRX path of the
ixgbe driver so ip packet can bypass ethernet-input node processing
for better performance. The shortcoming of packet type recognition
in the vRX path does not affect VPP.
Change-Id: Ibc96021c50402a75a8bc4f80125c664714276cef
Signed-off-by: John Lo <loj@cisco.com>
For using clang as a compiler it is enough
to specify CC=clang in the make command line
Change-Id: I06f1c1d418b68768f8119de5bdc8748c51f90c02
Signed-off-by: Damjan Marion <damarion@cisco.com>
DPDK 16.04 introduced new config parameter RTE_ARCH_X86
which was missing in the dpdk/Makefile
That caused issues with virtio devices on older kernels
(observed on 3.13.0).
This patch changes the source of generated config so
it will also include all defines in config/defconfig_TARGET
file.
Change-Id: Idc6ccbef1d01da84235281ea1078c081846446de
Signed-off-by: Damjan Marion <damarion@cisco.com>
Can be used by specifying DPDK_VERSION=16.04-rc2 in the make command line
Change-Id: I657b44d7ca22f1ef57756e7703088020fab12bc6
Signed-off-by: Damjan Marion <damarion@cisco.com>
This includes all patches in ENIC driver which are up-streamed to DPDK
to improve RX performance, fix buffer/error handling and interoperation
with link bonding PMD library.
Change-Id: Id4c71a350d5234834951f9261c69db5476ba396b
Signed-off-by: John Lo <loj@cisco.com>
poll mode driver library. The bonded interfaces to be created on
VPP startup is specified in the dpdk section of startup.conf
or qn.conf, using DPDK EAL command. Following is an example of a
dpdk section white listing PCI addressses of 4 ethernet interfacess
to be under VPP control plus two bonded interface and the PCI addresses
of the slaves in each:
dpdk { socket-mem 1024,1024
dev 0000:0f:00.0 dev 0000:10:00.0 dev 0000:11:00.0 dev 0000:12:00.0
vdev eth_bond0,mode=2,slave=0000:0f:00.0,slave=0000:11:00.0,xmit_policy=l34
vdev eth_bond1,mode=2,slave=0000:10:00.0,slave=0000:12:00.0,xmit_policy=l34
}
Note that only balance XOR (mode 2) is supported and "xmit_policy=l34"
specifies to use layer 3 SIP/DIP and layer 4 Sport/Dport for load
balance. Using "xmit_policy=l2" for SMAC/DMAC or "xmit_policy=l23" for
SMAC/DMAC and SIP/DIP should also work.
Change-Id: Iaf6438686fa20cce893cb5a823b76e2886b4360b
Signed-off-by: John Lo <loj@cisco.com>
