Age | Commit message (Collapse) | Author | Files | Lines |
|
Type: improvement
virtio interfaces support packet coalescing and buffering which
depends on timer expiry to flush the stored packets periodically.
virtio input node checks timer expiry and schedules tx queue
accordingly. In poll mode, timer expiry is handled naturally,
as input node runs periodically. In interrupt mode, virtio
input node depends on the interrupts send from backend.
Stored packets could starve, if there would not be interrupts
to input node.
This patch implements a process node which periodically
sends interrupt to virtio input node given coalescing or buffering
feature is enabled on an interface.
Change-Id: Ic38f749f74b001073d4d0579dca149d0a4cea039
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
- Avoid feature arc lookup per packet, as all packets belong to the
same interface
- don't validate enqueue for l2 packets, as they all have same
destionation
Type: improvement
Change-Id: I32ddd801f219eab9281736abafdc91a4ff9e0989
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Change-Id: I4fdcb40142e8cc9f5f1d0035116659543d5d92fe
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
There is no need to inline 3 times virtio_refill_vring()...
Type: fix
Change-Id: Ic26e24fd7911af743fedd0e2282784b715e86c3c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Change-Id: I478b6fc54c47f0e77a1470ed29fdd56774671441
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
clang-11 complains:
error: field 'buffer_template' with variable sized type 'vlib_buffer_t' not at the end of a struct or class is a GNU extension [-Werror,-Wgnu-variable-sized-type-not-at-end]
Type: improvement
Change-Id: I2cb6b4fde723a05b42cf33dd8130df074f0362ab
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
This is part of bigger refactor.
Type: refactor
Change-Id: I6fc2c0a1e2d217a70952901bcf775b8485bd3c20
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
This patch fixes the commit e347acbc31111504c015531e8ad764a86d489309
Change-Id: Icee7a6e250c94ae93e606d7869acc55c7a5806b1
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
QUEUE_SELECT and QUEUE_NOTIFY_OFF registers are shared between all
workers operating on the same device and operations are not atomic
Type: fix
Change-Id: Ie017b1bfc7e3b6b4e59029f45db78eeffd9f3aeb
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: feature
This patch adds packet buffering on tx for
slow backend which have some jitter/delays
in freeing the vrings.
There are some limitations to the current design:
1) It only works in poll mode.
2) Atleast 1 rx queue of an interface (with buffering
enabled) should be placed on each worker and main thread.
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Ib93c350298b228e80426e58ac77f3bbc93b8be27
|
|
Type: fix
Change-Id: I96c30baaf34fe7b0cd899966a507501e58cde934
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: refactor
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I37fef2c32d0561b944b201a1012b87c7ac315e73
|
|
sanitizer complains about strlen on hi->name in tap_dump_ifs.
hi->name is a vector which is not null-terminated, so use vec_len.
Type: fix
Signed-off-by: Vladimir Isaev <visaev@netgate.com>
Change-Id: Icdd5f65369bb51b0c4a9cd86c24899e6febd837c
|
|
Initialize txq lock only if some txq are shared and check if another
worker is already operating on the txq before processing gro timeouts
in input node.
Type: fix
Change-Id: I89dab6c0e6eb6a7aa621fa1548b0a2c76e6c7581
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: style
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Id86e16677564669b0295bbb9fc1303fe39a21b6f
|
|
The syntax of the deprecated flag has evolved.
Clean up usage to be "option deprecated;".
Type: fix
Signed-off-by: Ole Troan <ot@cisco.com>
Change-Id: If2b639f275eb8db58b36c457f9245fe35a4d8cb1
|
|
Type: fix
Change-Id: Ic07d0ae313b32e420ba93693cb75960a86f752a9
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
This patch adds the RSS steering queues set interface, and it's
implementation in DPDK device:
/* Interface to set rss queues of the interface */
typedef clib_error_t *(vnet_interface_rss_queues_set_t)
(struct vnet_main_t * vnm, struct vnet_hw_interface_t * hi,
clib_bitmap_t *bitmap);
This patch also introduces a command line to set the RSS queues:
set interface rss queues <interface> <list <queue-list>>
To display the rss queues, use "show hardware-interfaces"
Below is the example to configure rss queues for interface Gig0:
vpp# set interface rss queues Gig0 list 0,2,4-7
vpp# show hardware-interfaces brief
Name Idx Link Hardware
VirtualFunctionEthernet18/1/0 1 down VirtualFunctionEthernet18/1/0
Link speed: unknown
RSS queues: 0 2 4 5 6 7
local0 0 down local0
Link speed: unknown
vpp#
Users can also configure the rss queues on a dpdk interface in
startup.conf:
dpdk {
dev 0000:18:01.0 {
rss-queues 0,2,5-7
}
}
Type: feature
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
Change-Id: I1835595a1c54016a84eabee9fd62ce137935385d
|
|
Type: refactor
tap, virtio and vhost use virtio/vhost header files from linux
kernel. Different features are supported on different kernel
versions, making it difficult to use those in VPP. This patch
removes virtio/vhost based header dependencies to local header
files.
Change-Id: I064a8adb5cd9753c986b6f224bb075200b3856af
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: improvement
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Ia1276d00dded36ee28b4b2e93b4cc7c1df6b1eef
|
|
Type: feature
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I95d7fc1cc8db5199570c66535f45e867a7cae676
|
|
Type: feature
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I3e00deb94943c545d1649865b2efdf7d51b90f4d
|
|
Type: fix
Change-Id: Ia28161b583ea26ab820a494332a79b64add7004d
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: fix
Change-Id: I8d55c2bfdd3c4607044370ebabf40cbac78b4996
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: fix
Change-Id: I12b08333f3f69aaa882e8801f4f69bca2d7bd558
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: fix
Change-Id: I6c6d66ad8aa158be8d2b9d111de7d46473b9dc02
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: feature
Change-Id: I205f7c146a213d603d9d1e46fcf5195a876608dc
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: refactor
Change-Id: I7342178f9ab9adb99b91a4f984bc22bef2ce8021
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: feature
Change-Id: I5868dd267aa26aa97aec5fd70e70c5956ac52277
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: improvement
Change-Id: I991e32d531719693c387db4ef93d04d4b562789d
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: fix
Change-Id: Ie0cff37b474f8d85a3ae376e0f547a347fb1ad8a
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: fix
From kernel 4.20, xdp support has been added in tun_sendmsg.
If sndbuf == INT_MAX, kernel executes xdp data path
for tun driver which assumes packets are ethernet frames.
This patch is avoiding the xdp data path in kernel by setting
the sendbuf value < INT_MAX.
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Ia4aa54b177b96d56a2d513d18d26ca01d5b88929
|
|
Type: fix
Change-Id: I3bcc8ff1cf0a828ce3ba112694d38e3287d38d8d
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: improvement
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I0b1a1ce085ceb75fe139092e40878566682a7986
|
|
Type: fix
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Ifeae641ec0aa7de74e33e582234505bf6e28ca87
|
|
Type: fix
Change-Id: I92e8e30e4175229e24dd92471016e9095b5cc83e
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
The filter sw_if_index was in the API sw_interface_vhost_user_dump. But
it was never implemented in the backend. This patch is to add the backend,
vat, and custom dump support for the filter.
Type: feature
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: Iaa41a7e11bfbcbb4c60092375e4b0dcf0950077b
|
|
Type: improvement
Change-Id: I220ea6ab609e3c1628f5210be441d0d5e825a32c
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: improvement
Change-Id: I5b9d5ea192776f14a45bf909acc4bef7793521e8
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: docs
Signed-off-by: John DeNisco <jdenisco@cisco.com>
Change-Id: Iba106d33d34766b91e46980e7237fbdfc3710b8b
|
|
Type: fix
Change-Id: I5601bdeb47d08118476ff7bd29435d2c1dba34b9
Signed-off-by: Mohsin Kazmi <sykazmi@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
|
|
Type: feature
Change-Id: I37752af8496e0042a1da91124f3d94216b39ff11
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Add a hook to src/vlibapi/api_shared.c to fuzz (screw up) binary API
messages, e.g. by xoring random data into them before processing. We
specifically exempt client connection messages, and inband debug CLI
messages. We step over msg_id, client index, client context, and
sw_if_index. Otherwise, "make test" vectors fail too rapidly to learn
anything.
The goal is to reduce the number of crashes caused to zero. We're
fairly close with this patch.
Add vl_msg_api_max_length(void *mp), which returns the maximum
plausible length for a binary API message.
Use it to hardern vl_api_from_api_to_new_vec(...) which takes an
additional argument - message pointer - so it can verify that
astr->length is sane. If it's not sane, return a u8 *vector of the
form "insane astr->length nnnn\0".
Verify array lengths in vl_api_dhcp6_send_client_message_t_handler(...)
and vl_api_dhcp6_pd_send_client_message_t_handler(...).
Add a fairly effective binary API fuzz hook to the unittest plugin,
and modify the "make test" framework.py to pass "api-fuzz { on|off }"
to enable API fuzzing: "make API_FUZZ=on TEST=xxx test-debug" or similar
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I0157267652a163c01553d5267620f719cc6c3bde
|
|
Type: refactor
This patch refactor the existing flags and also add a new
flag for packet coalescing.
Change-Id: Ic826e4c81313f26d87c475cdf666b06cbed60a3a
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Type: feature
Change-Id: I6f607f383dc77a71e8712124f7613b38b4ac065a
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
Type: improvement
tap_delete_if() returns early if the interface type is not
VIRTIO_IF_TYPE_TAP. Allow VIRTIO_IF_TYPE_TUN also and take
appropriate action for those interfaces.
Change-Id: I196b6d6f3f5e1543a14d6be76fd879d44c9794fd
Signed-off-by: Matthew Smith <mgsmith@netgate.com>
|
|
GCC-10 increases string truncations warnings. Refactor string copies
confusing it.
Type: refactor
Change-Id: I9720a0539059de00ab212ff2fc73055f04f5af1d
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: fix
Change-Id: Ib09c7cebb6978b3adc09ac36cb32f7947b143e51
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
This matches vhost queues to linux netdev queues and avoids random
packet shuffling across vhost queues on rx.
Change-Id: I9901689d361e440fb0b91c9fbaf8124ce525b316
Type: fix
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
|