Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I9794da718805b40cc922e4f3cf316255398029a9
Type: improvement
Signed-off-by: Damjan Marion <damarion@cisco.com>
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
Type: improvement
Signed-off-by: Saravanan Murugesan <sarmurug@cisco.com>
Change-Id: I90e90678ae6586019cc842f9d504d53991cfabe4
|
|
- support of in-place growth of vectors (if there is available space next to
existing alloc)
- drops the need for alloc_aligned_at_offset from memory allocator,
which allows easier swap to different memory allocator and reduces
malloc overhead
- rework of pool and vec macros to inline functions to improve debuggability
- fix alignment - in many cases macros were not using native alignment
of the particular datatype. Explicitly setting alignment with XXX_aligned()
versions of the macro is not needed anymore in > 99% of cases
- fix ASAN usage
- avoid use of vector of voids, this was root cause of several bugs
found in vec_* and pool_* function where sizeof() was used on voids
instead of real vector data type
- introduce minimal alignment which is currently 8 bytes, vectors will
be always aligned at least to that value (underlay allocator actually always
provide 16-byte aligned allocs)
Type: improvement
Change-Id: I20f4b081bb13bbf7bc0ace85cc4e301787f12fdf
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: refactor
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I30cc31c438d357d48576fba84e54809455960eaa
|
|
Type: improvement
Signed-off-by: sarmurug <sarmurug@cisco.com>
Change-Id: Ibbfe827b9c4c603a6fe7cc49970a46bd683194ce
|
|
Type: fix
Change-Id: I857403b9d93ee4c17f2dd5ac8e6dafd66260a252
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Creating tap interface / sub interface causes allocation of a new
software interface with possible sw interface pool reallocation.
In such case accessing L3 MTU and interface flags by obsolete sw
pointer is UAF.
Instead, keep desired tap interface MTU value before sw intreface
creation and refetch sw pointer right before sw flags inheritance.
Type: fix
Fixes: b89c1ddcb3b4f9138ca3ebefb2115f896ff3e1bd
Signed-off-by: Vladislav Grishenko <themiron@yandex-team.ru>
Change-Id: I21ea46d146d11060bb9bedc77377ab17ae9e22e8
|
|
Type: fix
Fixes: 91f17dc
Change-Id: I860b6d5d5e9cf47d84fde0a2c92be43125038694
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Currently, the plugin can monitor link state changes on hardware
interfaces for which a linux-cp pair exists. When the link goes down on
one of the hardware interfaces, the plugin processes IPv4 routes that
resolve through that interface according to the configurations:
del-static-on-link-down and del-dynamic-on-link-down.
The problem is that link state changes are not signaled for
subinterfaces and the code that handles IPv4 routes is not triggered.
When the link on a hardware interface goes down, it implies
that subinterfaces added to that interface also will have the link in
the down state.
With this change, when the link goes down on a hardware interface,
iterate over subinterfaces added to the interface and apply the same
logic of routes processing as for hardware interfaces.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I97337d2e328437c73f2d99a00737768778f197a1
|
|
Type: improvement
lcp_router_table_flush() is used to remove routes from the given route
table if they are resolved through the given interface with specified
FIB source. Currently, if you need to remove routes from a route table
that are resolved through one of the interfaces from a vector, the
function has to be executed for every interface from the vector. Every
execution walks the route table again.
With this change, lcp_router_table_flush() accepts a vector of
interfaces. Thus, the function can walk a single route table once and
delete routes that resolve through one of the interfaces from the
vector.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I6d99384064d279dea24bb4dc1790a1af943be41c
|
|
Fix ordering in CMakeLists.txt
Type: refactor
Signed-off-by: Ray Kinsella <mdr@ashroe.eu>
Change-Id: I8e71e4fbc048a80c4b250c2a66cfd8a522bde5f4
|
|
format() expects a NULL-terminated C-string as format string.
Type: fix
Change-Id: Ib428cf2debbf98850eed512907175f8ae8ba3c04
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
* Correct endianness issue
* lip_namespace is a vector not a string
* Provide null termination to avoid unpack() failures in the client
Responses in the python API now look like:
lcp_itf_pair_details(_0=328, context=3, phy_sw_if_index=1,
host_sw_if_index=3, vif_index=19, host_if_name='ice0',
host_if_type=<vl_api_lcp_itf_host_type_t.LCP_API_ITF_HOST_TAP: 0>,
namespace='dataplane')
lcp_itf_pair_details(_0=328, context=3, phy_sw_if_index=2,
host_sw_if_index=4, vif_index=20, host_if_name='ice1',
host_if_type=<vl_api_lcp_itf_host_type_t.LCP_API_ITF_HOST_TAP: 0>,
namespace='dataplane')
lcp_itf_pair_details(_0=328, context=3, phy_sw_if_index=5,
host_sw_if_index=6, vif_index=21, host_if_name='ice0.1234',
host_if_type=<vl_api_lcp_itf_host_type_t.LCP_API_ITF_HOST_TAP: 0>,
namespace='dataplane')
Type: fix
Signed-off-by: Pim van Pelt <pim@ipng.nl>
Change-Id: If4bf06a8b70977676ec7f5b1413cee6cc9d1714a
|
|
Type: improvement
Change-Id: I5317afa02fa1525a7d8df595b56eb6546ccded57
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Format host interface name as vector rather than c-string.
Otherwise non-null-terminated vector overrun triggers ASAN.
Type: fix
Fixes: 1705a6baefe205bb6792b547c7376eee3f328a71
Signed-off-by: Dmitry Valter <d-valter@yandex-team.ru>
Change-Id: Ib204e57ee17c7ed3bfeb568dcdd834d7d7519102
|
|
Do connects in an session layer rpc instead of doing cli process sleeps.
Performance with 4 workers goes from ~120k/180k first run/warmed up to
~135k/240k
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Id184913e9898f8db099e29a605f3a9b1fc67be63
|
|
Type: improvement
Change-Id: I946d91e67c332ecac0b09d50980529b9d4e74f9c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Currently, when the link goes down on an interface, routes that resolve
through that interface and created with Nexthop API are removed by the
kernel. However, IPv4 routes remain in the FIB because the kernel
doesn't send any notifications about that.
And for the plugin working with user-space applications that create
routes in the kernel using Nexthop API there should be a mechanism to
synchronize the FIB and the kernel in this case.
With this change, add two new startup configuration options to the
plugin to be able to control what should happen with static and dynamic
routes managed by the plugin on link down:
- del-static-on-link-down (disabled by default, delete routes created
with the linux-cp static FIB source on link down),
- del-dynamic-on-link-down (disabled by default, delete routes created
with the linux-cp dynamic FIB source on link down).
Then, monitor link state changes on interfaces for which a linux-cp pair
exists. If the link goes down on one of the interfaces, process routes
that resolve through that interface according to the new configurations.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I0fbaeeca3f3d1fcd22e8eebb08a0a4a3d0dfe5b8
|
|
Type: fix
Change-Id: Id7ae0d4c53cbca3785964f0bef92e9cd589d4ce9
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
The way of specifying which interface supports what has changed.
Re-add support for adaptive mode in virtio.
Type: fix
Signed-off-by: Paul Atkins <patkins@graphiant.com>
Change-Id: I2f8ffa6311a4081b93fb08a7e92408b8bffbae64
|
|
The support for interrupt mode is not being respected as it is not
copied into the driver conf, which is what is checked when trying to
enable it.
Type: fix
Signed-off-by: Paul Atkins <patkins@graphiant.com>
Change-Id: I2e502306e27bd98f8037d1a0a396201e099b50b5
|
|
Use of clib_mem_is_heap_object is not reliable enough for production use
as it relies on just few bytes of memory allocator chunk header.
Type: improvement
Change-Id: I48c8adde8b6348b15477e3a015ba515eb7ee7ec2
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
This allows specifying both c string and vector for node name
and removes need for crafting temporary string.
Type: improvement
Change-Id: I0b016cd70aeda0f68eb6f9171c5152f303be7369
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Fixes: 65105c95fe03 ("dpdk: improve logging")
Signed-off-by: Tianyu Li <tianyu.li@arm.com>
Change-Id: I69d616c7e6e7b5395ebf083b1ac5c3e85f99bbdd
|
|
Type: refactor
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ibb6d19de053c306e9758dbfa827ab7bcab5de856
|
|
- move memory and server specific vl_msg_api_handler_with_vm_node()
to memory server code only
- keep api_global_main static
Apart from being cleaner, this also helps avoiding symbols conflict
when both client and server libs are loaded in the same process, as is
done by the prom plugin.
Those symbols conflict confuse ASan and can be nasty to debug.
Type: improvement
Change-Id: Iaf58596cc753ad8d3fedd8d65c4bf480ac129c2c
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: fix
Change-Id: I43ebb2c2922f3b8b8eddf26ccdf044f31d7b7a10
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Change-Id: Ib025d8c1bc9bd651b6448d6a41fd1efa6f0b7362
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I313c916d268c4b2b448b93e90bc67da341b803e3
|
|
Type: fix
With multi-txq in VPP, user should be able to create
more txqs than vpp threads. MEMIF_MAX_M2S_RING should
be defined to 256 instead of number of vpp threads.
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I337c3a5ea691470815653ff2dbfa862bb324b240
|
|
Type: refactor
Change-Id: I26a2a410f8f4070d10696f796d5887842cad3916
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Just like commit 252647482b24bb3474e8f13bc86100718176832f
did for Wireguard interface keys, prevent stack smashing
the peer keys.
Integer math on 32 bytes of base64 data might yield 33 bytes
of data in some poorly formed user input of private key values.
Rather than smashing the stack (detected) and aborting, simply
allow for the possible yet irrelevant 33-rd byte of data.
Type: fix
Fixes: edca1325cf296bd0f5ff422fc12de2ce7a7bad88
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: I9f77b3faaaa01d3123b356c958db60c87238db9c
|
|
- Add peer_index, table_id, and keep-alive.
- Fix some lingering cut-n-paste issues in the API file.
Type: improvement
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Change-Id: I6fc0729ba0268bbcda0248f680979c44e68b6b0c
|
|
Type: feature
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
Change-Id: If24a43b7f79e05092306562c192de03994dec550
|
|
When dNAT-ing to a VIP, it can happen
that the return session conflicts with
another forward session than the one
we own.
This patchs adds a rsession_flags
CNAT_SESSION_RETRY_SNAT that makes cnat_session_create
search for a free src port to use for the
resulting return session.
It also makes forward & return session
share their fate in the session scanner.
Type: fix
Change-Id: Id0edf59abf8e5bc0c0d8941ba289c4563c77dee0
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: fix
Calling vnet_hw_if_register_tx_queue should
be done with the worker barrier held, as
virtio-pre-input might be grabbing a queue
while a memif connect event is triggered.
Change-Id: Ie1272cdfd2477faf7a4e10f30778279872f04916
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Add missing translation of ICMP inner IP layer.
Change responsible test so that it actually tests something.
Type: fix
Fixes: 4881cb4c6f
Signed-off-by: Klement Sekera <klement.sekera@gmail.com>
Change-Id: Id3a6f12a7308d81b1cdf9815f857221fab2f24d9
|
|
In order to support the requirement of RSS and packet steering of new
protocols, such as GTPU PDU-type and QFI, for 5G UPF, a generic pattern
is introduced in vnet flow. The generic flow pattern is based on DDP
(Dynamic Device Personalization) function and Parser Library module in
DPDK. Using generic flow pattern, we do not need to create new packet
and field type and offset in API parser for every new protocols. We can
create flows for any protocol immediately as long as supported by DDP.
The generic flow can be used to support 5G related protocols in
different scenarios.
The input of this generic pattern are two binary strings for spec and
mask. Spec is the binary presentation of the target packet type, and
mask is used to mark the target fields.
In this patch DPDK plugins is enabled for POC. Next step we will enable
generic flow in native IAVF, which is the main target.
Here is an example. If we want to create a flow for GTPU QFI,
spec is:
00000000000100000000000208004500003C00000000001100000101010102020202000
008680028000034FF001C00000000000000850100010045000014000000000000000001
01010102020202
mask is:
00000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000007F0000000000000000000000000000
00000000000000
A naming API POC is created via VAPI to help create the rule with
the target packet format similar to Scapy. It is based on a function
module called PacketForge. In this way, the user no need to create
binary string spec and mask by themselves.
Type: feature
Signed-off-by: Ting Xu <ting.xu@intel.com>
Change-Id: Id3444f95c158bdcdfeeee19d795cd9ecbeeec07c
|
|
The index i was incremented in the wrong place, and the check on the
presence of a next buffer in the chain was actually done for the
next desc_status rather than the current one.
Type: fix
Signed-off-by: Mauro Sardara <msardara@cisco.com>
Change-Id: I74a64a34fea497900b7969cd96e1aeeb570a1bba
|
|
With DPDK plugin, VPP does the DMA page map in IOMMU, only when
DPDK supported ethernet devices are present. As a result, Mellanox NIC
and QAT combo doesn't work. As part of this fix, DPDK supported
crypto device check is added to do the DMA page map.
Type: fix
Signed-off-by: mgovind <govindarajan.mohandoss@arm.com>
Change-Id: I02de4588c5b021e0c9c62612137f28ed8784bea6
|
|
Fixes a minor issue that causes printing an error message when there is no error.
Type: fix
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
Change-Id: I59f5c6af7c5aeae3e812b4cf0c75a47894bb8bbd
|
|
Type: improvement
This patch improves the per dpdk-input loop number of packets
received from the port. The change mimics how packets rx happened
before VPP 22.02/DPDK 21.11: instead of trying to rx huge number
of packets (256) in one go, rx more times with up to 32 packets
max each time.
Signed-off-by: Fan Zhang <roy.fan.zhang@intel.com>
Change-Id: I804dce6d9121ab21b02e53dd0328dc52ac49d80f
|
|
Thanks to Ben McKeegan <ben@netservers.co.uk> for the report.
Type: fix
Change-Id: I8170dda572c326b6b1823fd330dbd5e961fdad74
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: refactor
Change-Id: Ieb7a595e40d801af5349c83b128fa92c7698a346
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Recent "dpdk: refactor device setup" have broken vlans programming for IXGBE_VF.
Type: fix
Signed-off-by: Dzmitry Sautsa <dzmitry.sautsa@nokia.com>
Change-Id: Idacda33a473f6b10dbe002d9926661a19d0f3f97
|
|
Type: refactor
Change-Id: Ifd533a095d979dc55bfbe5fac7e0b7510a4d900c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Currently, when an interface is brought down administratively, IPv4
routes that resolve through that interface remain in the FIB. However,
the kernel removes those routes but doesn't send any notifications about
that. Desynchronization between the kernel and VPP happens.
With this change, when a notification received from the kernel
indicating that an interface was brought down, in addition to bringing
the VPP interface down, walk the IPv4 FIB bound to that interface and
remove any entries that resolve through that interface and were added
with one of the linux-cp FIB sources.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I0cd14bb63c9e6616ae1c5739b17c3bf33b186bc2
|
|
Type: fix
Primarily fix an issue reported by Coverity in
lcp_nl_open_sync_socket() that close() could possibly be run with
negative fd. Also, add more checks and error logging there.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I9a88520d068392977a6eba0766451e5652fe512c
|
|
Type: improvement
Currently, read event signal is sent on every notification message
received and added in the queue.
With this change, signal read event only when all currently available
notification messages are received.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: Ib86d189311ce01f50167e4e97feb99df0292ad96
|
|
Type: improvement
Currently, while reading notifications, ENOBUFS error is ignored and
reading continues. This was done to minimize the number of notifications
that are lost due to reopening the socket.
Now that synchronization is implemented to recover from socket errors,
ignoring ENOBUFS and reading as much notifications as possible is not
actual. Before synchronization, all currently enqueued notification are
discarded in any case.
With this change, stop reading notifications if any error occurs.
Signed-off-by: Alexander Chernavin <achernavin@netgate.com>
Change-Id: I1184d9a3aa99df63ef59bc2a67be2b1e5e0e9329
|