Type: improvement
Some tests i.e. ipsec see performance regression when offload flags
are moved to 2nd cacheline. This patch moves them back to 1st cacheline.
Change-Id: I6ead45ff6d2c467b0d248f409e27c2ba31758741
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
trajectory trace has been broken for a while because we used to save the
buffer trajectory in a vector pointed to in opaque2. This does not work
well when opaque2 is copied (eg. because of a clone) as 2 buffers end up
sharing the same vector.
This dedicates a full cacheline in the buffer metadata instead when
trajectory is compiled in. No dynamic allocation, no sharing, no tears.
Type: refactor
Change-Id: I6a028ca1b48d38f393a36979e5e452c2dd48ad3f
Signed-off-by: Benoît Ganne <bganne@cisco.com>
This patch fixes a vxlan flow type checking bug which is introduced in
commit 7a016e87e4.
It is the native avf instead of dpdk that doesn't support vxlan flow offloading.
So this patch removes Vxlan flow supporting in native avf_plugin and adds
the vxlan flow back to dpdk_plugin.
Type: fix
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
Change-Id: I1a72629cee462b890327c589c846f6d532b2d61d
This patch fixes the below bugs in avf flow:
1. remove the VXLAN flow as it is not yet supported
2. fix a typo in IPv6 flow
3. fix the wrong type of the AH flows
Also did some cleanup in this patch
Type: fix
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
Change-Id: I5eae131f1e89071364a41b7006a00e8425089a4f
Need to add msg_id_base with VL_API_AVF_DELETE to properly set the
correct is_mp_safe variable.
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I20f6e9441727fb7ffc132965b93c0a90482e1a52
When deleting an avf interface which is not the very first interface
that was created (dev_instance == 0), VPP crashes.
The reason is every avf interface delete always removes the very first
device instance due to ad->dev_instance was wiped out prior to the
statement pool_put_index (am->devices, ad->dev_instance)
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I69bd1588aab9a176f8eef46be7aa5063f5d29482
This patch adds the flow support to VPP avf_plugin.
Currently only IP4-N-TUPLE and GTPU flow are supported as
a reference. More flows will be added later.
Type: feature
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
Change-Id: I563dbfacecd3eba31e8c3c7769142ea6adba54b6
This patch introduces a 'program flow' virtchannel event.
The parameter is_add is used to identify adding or deleting the rule from avf
Type: improvement
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
Change-Id: I1f5545ddc23e4494a25afa36ff601607405a8c3d
Type: improvement
This patch adds flags to represent the modern NICs capabilities.
Change-Id: I96d38d9ab7eac55974d72795cd100d8337168e1e
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Type: refactor
This patch refactors the offload flags in vlib_buffer_t.
There are two main reasons behind this refactoring.
First, offload flags are insufficient to represent outer
and inner headers offloads. Second, room for these flags
in first cacheline of vlib_buffer_t is also limited.
This patch introduces a generic offload flag in first
cacheline. And detailed offload flags in 2nd cacheline
of the structure for performance optimization.
Change-Id: Icc363a142fb9208ec7113ab5bbfc8230181f6004
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Performance improvement: on supported platforms, currently only Intel Tremont,
use a write combining store to update the tail pointers.
Also, Tremont node variant is added for all.
Type: improvement
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Change-Id: Ie9606e403b7d9655184f778e3ffee3027c8c9edd
Use vlib_buffer_t::current_data instead of
vnet_buffer_opaque_t::l2_hdr_offset to compute l2_len for checksum
offload (l2_hdr_offset might be invalid if packet originates from an L3
interface)
Change-Id: I2031ea6fd6a7af4b6e186751e119ebd6161641b5
Type: fix
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
Initializing struct avf_ip6_psh by {0} using gcc with O2 optimize option
will trigger the -Werror=maybe-uninitialized compiling warning on Arm
because gcc compiler will think some members of the struct avf_ip6_psh
may not be initialized, which probably is a false positive in this case.
The compiling error log is shown as below. Avoid this compiling warning
by explicitly declaring the IPv6 src and dst ip in avf_ip6_psh as
ip6_address_t.
ccache /usr/lib/ccache/gcc-10 -DHAVE_FCNTL64 -DHAVE_GETCPU -DHAVE_MEMFD_CREATE -I/home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src -I. -Iinclude -I/home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/plugins -Iplugins -Iplugins/avf -Wno-address-of-packed-member -g -fPIC -Werror -Wall -march=armv8-a+crc -O2 -fstack-protector -DFORTIFY_SOURCE=2 -fno-common -fPIC -DCLIB_MARCH_VARIANT=cortexa72 -march=armv8-a+crc+crypto -mtune=cortex-a72 -DCLIB_N_PREFETCHES=6 -MD -MT plugins/avf/CMakeFiles/avf_plugin_cortexa72.dir/output.c.o -MF plugins/avf/CMakeFiles/avf_plugin_cortexa72.dir/output.c.o.d -o plugins/avf/CMakeFiles/avf_plugin_cortexa72.dir/output.c.o -c /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/plugins/avf/output.c
In file included from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/vector_funcs.h:41,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/vector.h:196,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/string.h:48,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/mem.h:49,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/vec.h:42,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/format.h:44,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/elf.h:41,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/elf_clib.h:41,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vlib/vlib.h:44,
from /home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/plugins/avf/output.c:18:
/home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/plugins/avf/output.c: In function ‘avf_device_class_tx_fn_cortexa72’:
/home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/vppinfra/byte_order.h:59:10: error: ‘*((void *)&psh+32)’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
59 | return __builtin_bswap16 (x);
| ^~~~~~~~~~~~~~~~~~~~~
/home/snowball/tasks/test_vpp_build/test-patch-9/vpp/src/plugins/avf/output.c:115:23: note: ‘*((void *)&psh+32)’ was declared here
115 | struct avf_ip6_psh psh = { 0 };
| ^~~
Type: fix
Change-Id: I2684b101b07823dfacc4a56cc29d152828d0cf37
Signed-off-by: Jieqiang Wang <jieqiang.wang@arm.com>
When using classifier to filter traces, not all packets will be traced.
In that case, we should only count traced packets.
Type: fix
Change-Id: I87d1e217b580ebff8c6ade7860eb43950420ae78
Signed-off-by: Benoît Ganne <bganne@cisco.com>
Entering duplicate pci address when creating an avf interface causes
crash in register_node
(gdb) f 4
145 error_exit (1);
(gdb) up
354 clib_error ("more than one node named `%v'", n->name);
(gdb) p n
$1 = (vlib_node_t *) 0x7fffbbe55de4
(gdb) p n->name
$2 = (u8 *) 0x7fffbc410b10 "avf-0/4/a/0-tx"
(gdb)
The fix is to loop through the pci addresses in the avf interface pool
and to reject the duplicate.
Type: improvement
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I4ed6fb630fb11982d85c5bb325d9f0d6beeaf023
Check CLI queue size is within the range of 64 and 4096
Enhance show hardware to display queue size and number of queues.
Type: improvement
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I360e3cdb2e69e4ea7380ed924e71a5ae84ed4b64
Avoid situations where promisc mode is chaged while avf process
is suspended in the middle of adminq operation.
Type: fix
Change-Id: Ia1fc6551e83218b5938630ad3a15d4f3f0ceceff
Signed-off-by: Damjan Marion <damarion@cisco.com>
device pool my grow during suspemd which will cause crash in avf process
after it exits from suspend.
Type: fix
Change-Id: I51fec90088c909cfbaaca6c245272a28c0827ca0
Signed-off-by: Damjan Marion <damarion@cisco.com>
It may happen that process node is suspended while it waits for response
from adminq and during that time CLI or API process can call
avf_delete_if. When avf process node resumes, it may happen that device
is not there anymeore.
This patch delegates interface deletion to process node, so CLI/API
process just sends signal instead of deleting device instance itself.
Type: fix
Change-Id: I7f12e12df3071650f6e60ad7eb5af23b7acfe335
Signed-off-by: Damjan Marion <damarion@cisco.com>
Fix and optimize DMAC check in ethernet-input node to utilize NIC or
driver which support L3 DMAC-filtering mode so that DMAC check can be
bypassed safely for interfaces/sub-interfaces in L3 mode.
Checking of interface in L3-DMAC-filtering state to avoid DMAC check
require the following:
a) Fix interface driver init sequence for devices which supports L3
DMAC-filtering to indicate its capability and initialize interface
to L3 DMAC-filtering state.
b) Fix ethernet_set_flags() function and its associated callback
flags_change() functions registered by various drivers in interface
infra to provide proper L3 DMAC filtering status.
Maintain interface/sub-interface L3 config count so DMAC checks can be
bypassed if L3 forwarding is not setup on any main/sub-interfaces.
Type: fix
Ticket: VPP-1868
Signed-off-by: John Lo <loj@cisco.com>
Change-Id: I204d90459c13e9e486cfcba4e64e3d479bc9f2ae
Support VIRTCHNL_VF_CAP_ADV_LINK_SPEED.
Type: improvement
Change-Id: Ie45c9d557d0369464cbb6881ee1259ef3807d7d5
Signed-off-by: Damjan Marion <damarion@cisco.com>
If buffer alloc fails, it may happend that rx queue will be stuck
as old code only refills if at least one packet is received.
Type: fix
Change-Id: I388c4f8a9fb2c208bdc222e31b443cbe6b94af82
Signed-off-by: Damjan Marion <damarion@cisco.com>
