Age | Commit message (Collapse) | Author | Files | Lines |
|
In tap tx routine, virtio_interface_tx_inline, there used to be an
interface spinlock to ensure packets are processed in an orderly fashion
clib_spinlock_lock_if_init (&vif->lockp);
When virtio code was introduced in 19.04, that line is changed to
clib_spinlock_lock_if_init (&vring->lockp);
to accommodate multi-queues.
Unfortunately, althrough the spinlock exists in the vring, it was never
initialized for tap, only for virtio. As a result, many nasty things can
happen when running tap interface in multi-thread environment. Crash is
inevitable.
The fix is to initialize vring->lockp for tap and remove vif->lockp as it
is not used anymore.
Change-Id: Ibc8f5c8192af550e3940597c06992dfdaccb4c49
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
Indirect buffers are used to store indirect descriptors
to xmit big packets.
This patch moves the indirect buffer allocation from
interface creation to device node. Now it allocates
or deallocates buffers during tx for chained buffers.
Change-Id: I55cec208a2a7432e12fe9254a7f8ef84a9302bd5
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: I73f76c25754f6fb14a49ae47b6404f3cbabbeeb5
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Crash will happen when someone will try to setup a tap interface
in host namespace without providing the host side of tap interface
custom name. This patch fixes the problem by using the default name
in this case.
Change-Id: Ic1eaea5abd01bc6c766d0e0fcacae29ab7a7ec45
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
(cherry picked from commit 2a6861ff7bc90c1518b68459210830052774d429)
|
|
Some API action handlers called vl_msg_ai_send_shmem()
directly. That breaks Unix domain socket API transport.
A couple (bond / vhost) also tried to send a sw_interface_event
directly, but did not send the message to all that had
registred interest. That scheme never worked correctly.
Refactored and improved the interface event code.
Change-Id: Idb90edfd8703c6ae593b36b4eeb4d3ed7da5c808
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
Change-Id: I215e1e0208a073db80ec6f87695d734cf40fabe3
Signed-off-by: Jim Thompson <jim@netgate.com>
|
|
Change-Id: I7c6e4bf2abf08193e54a736510c07eeacd6aebe7
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: I9282a838738d0ba54255bef347abf4735be29820
Signed-off-by: Jim Thompson <jim@netgate.com>
|
|
Change-Id: I550313a36ae02eb3faa2f1a5e3614f55275a00cf
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Change-Id: Id71ffa77e977651f219ac09d1feef334851209e1
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
RDMA ibverb is a userspace API to efficiently rx/tx packets. This is an
initial, unoptimized driver targeting Mellanox cards.
Next steps should include batching, multiqueue and additional cards.
Change-Id: I0309c7a543f75f2f9317eaf63ca502ac7a093ef9
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Change-Id: I53ab8d17914e6563110354e4052109ac02bf8f3b
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
This reverts commit e63325e3ca03c847963863446345e6c80a2c0cfd.
Allow time for CSIT to accommodate.
Change-Id: I59435e4ab5e05e36a2796c3bf44889b5d4823cc2
Signed-off-by: ot@cisco.com
|
|
Use of consistent API types for interface.api
Change-Id: Ieb54cebb4ac96b432a3f0b41596718aa2f34885b
Signed-off-by: Jakub Grajciar <jgrajcia@cisco.com>
|
|
Fix a typo in vhost_user_rx_discard_packet which may cause
txvq->last_avail_idx to go wild.
Change-Id: Ifaeb58835dff9b7ea82c061442722f1dcaa5d9a4
Signed-off-by: Steven Luong <sluong@cisco.com>
(cherry picked from commit 39382976701926c1f34191c1311829c15a53cb01)
|
|
Change-Id: I8819bcb9e228e7a432f4a7b67b6107f984927cd4
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I911fb3f1c6351b37580c5dbde6939a549431a92d
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
map_guest_mem may be called from worker-thread/dataplane. It has a call
to vlib_log and may crash inside vlib_log's ASSERT statement
/* make sure we are running on the main thread to avoid use in dataplane
code, for dataplane logging consider use of event-logger */
ASSERT (vlib_get_thread_index () == 0);
The fix is to convert the vlib_log call in map_guest_map to event logger
Change-Id: Iaaf6d86782aa8a18d25e0209f22dc31f04668d56
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
while https://gerrit.fd.io/r/#/c/16590/ fixed the leaked fd which coverity
reported at that time, new coverity run reports simailar leaked fd in a
different goto punt path. It would be nice if coverity reported both of them
at the same time. Or perhaps it did and I just missed it. Anyway, the new fix
is to put the close (fd) statement prior to the return of tap_create_if routine
which should catch all goto's.
Change-Id: I0a51ed3710e32d5d74c9cd9b5066a667153e2f9d
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
Change-Id: I01c4f5755d579282773ac227b0bc24f8ddbb2bd1
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Symptom
-------
With NDR traffic blasting at VPP, bringing up a new VM with vhost
connection to VPP causes packet drops. I am able to recreate this
problem easily using a simple setup like this.
TREX-------------- switch ---- VPP
|---------------| |-------|
Cause
-----
The reason for the packet drops is due to vhost holding onto the worker
barrier lock for too long in vhost_user_socket_read(). There are quite a
few of system calls inside the routine. At the end of the routine, it
unconditionally calls vhost_user_update_iface_state() for all message
types. vhost_user_update_iface_state() also unconditionally calls
vhost_user_rx_thread_placement() and vhost_user_tx_thread_placement().
vhost_user_rx_thread_placement scraps out all existing cpu/queue mappings
for the interface and creates brand new cpu/queue mappings for the
interface. This process is very disruptive and very expensive. In my
opinion, this area of code needs a makeover.
Fixes
-----
* vhost_user_socket_read() is rewritten that it should not hold
onto the worker barrier lock for system calls, or at least minimize the
need for doing it.
* Remove the call to vhost_user_update_iface_state as a default route at
the end of vhost_user_socket_read(). There is only a couple of message
types which really need to call vhost_user_update_iface_state(). We put
the call to those message types which need it.
* Remove vhost_user_rx_thread_placement() and
vhost_user_tx_thread_placement from vhost_user_update_iface_state().
There is no need to repetatively change the cpu/queue mappings.
* vhost_user_rx_thread_placement() is actually quite expensive. It should
be called only once per queue for the interface. There is no need to
scrap the existing cpu/queue mappings and create new cpu/queue mappings
when the additional queues becomes active/enable.
* Change to create the cpu/queue mappings for the first RX when the
interface is created. Dont remove the cpu/queue mapping when the
interface is disconnected. Remove the cpu/queue mapping only when the
interface is deleted.
The create vhost user interface CLI also has some very expensive system
calls if the command is entered with the optional keyword "server"
As a bonus, This patch makes the create vhost user interface binary-api and
CLI thread safe. Do the protection for the small amount of code which is
thread unsafe.
Change-Id: I4a19cbf7e9cc37ea01286169882e5603e6d7eb77
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
This commit adds a "gso" parameter to existing "create tap..." CLI,
and a "no-gso" parameter for the compatibility with the future,
when/if defaults change.
It makes use of the lowest bit of the "tap_flags" field in the API call
in order to allow creation of GSO interfaces via API as well.
It does the necessary syscalls to enable the GSO
and checksum offload support on the kernel side and sets two flags
on the interface: virtio-specific virtio_if_t.gso_enabled,
and vnet_hw_interface_t.flags & VNET_HW_INTERFACE_FLAG_SUPPORTS_GSO.
The first one, if enabled, triggers the marking of the GSO-encapsulated
packets on ingress with VNET_BUFFER_F_GSO flag, and
setting vnet_buffer2(b)->gso_size to the desired L4 payload size.
VNET_HW_INTERFACE_FLAG_SUPPORTS_GSO determines the egress packet
processing in interface-output for such packets:
When the flag is set, they are sent out almost as usual (just taking
care to set the vnet header for virtio).
When the flag is not enabled (the case for most interfaces),
the egress path performs the re-segmentation such that
the L4 payload of the transmitted packets equals gso_size.
The operations in the datapath are enabled only when there is at least
one GSO-compatible interface in the system - this is done by tracking
the count in interface_main.gso_interface_count. This way the impact
of conditional checks for the setups that do not use GSO is minimized.
"show tap" CLI shows the state of the GSO flag on the interface, and
the total count of GSO-enabled interfaces (which is used to enable
the GSO-related processing in the packet path).
This commit lacks IPv6 extension header traversal support of any kind -
the L4 payload is assumed to follow the IPv6 header. Also it performs
the offloads only for TCP (TSO - TCP segmentation offload).
The UDP fragmentation offload (UFO) is not part of it.
For debug purposes it also adds the debug CLI:
"set tap gso {<interface> | sw_if_index <sw_idx>} <enable|disable>"
Change-Id: Ifd562db89adcc2208094b3d1032cee8c307aaef9
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
-fno-common makes sure we do not have multiple declarations of the same
global symbol across compilation units. It helps debug nasty linkage
bugs by guaranteeing that all reference to a global symbol use the same
underlying object.
It also helps avoiding benign mistakes such as declaring enum as global
objects instead of types in headers (hence the minor fixes scattered
across the source).
Change-Id: I55c16406dc54ff8a6860238b90ca990fa6b179f1
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Change-Id: I4e836244409c98739a13092ee252542a2c5fe259
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Example:
buffers {
default data-size 1536
}
Change-Id: I5b4436850ca18025c9fdcfc7ed648c2c2732d660
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Idd560f3afde1dd03bc3d6fbb2070096146865f50
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: Ifc98373371b967c49a75989eac415ddda1dcf15f
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: I60f88d50f062b004e6dea487bd627d303d0a5e75
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: Ib1316482dd7b1ae3c27c7eeb55839ed8af9ca162
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: I2e5fd45abcd07e9eda6184587889bdcd9613a159
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: Ieadf0a97379ed8b17241e454895c4e5e195dc52f
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: Id7fccf2f805e578fb05032aeb2b649a74c3c0e56
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Change-Id: I25b2a28513821bc5eab9ac6890a3964d412b0399
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I26f99d95f52c9fe107d17dcbbf5c6185523beade
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
* u32/u64/uword mismatches
* pointer-to-int fixes
* printf formatting issues
* issues with incorrect "ULL" and related suffixes
* structure alignment and padding issues
Change-Id: I70b989007758755fe8211c074f651150680f60b4
Signed-off-by: David Johnson <davijoh3@cisco.com>
|
|
Change-Id: If2bbfbc52994f5de0879763e0b7a7864498debb6
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
coverity complains about fd leaking inside the if statement because there is
a goto which bypasses the statement close (fd).
The fix is to close (fd) immediately after it is no longer used.
Change-Id: Ic5035b07ec1f179ff3db77744843e47aa8067a3c
Signed-off-by: Steven <sluong@cisco.com>
|
|
Verify that last node in the computed feature order matches
reality. This check doesn't make sense in all cases, so we skip it if
the newly-added vnet_feature_arc_registration_t ".last_in_arc" datum
is a NULL pointer.
Change-Id: Ia99c3e2b2da2e4780a7d5bc71670c5742a66fef2
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Should be less expensive...
Change-Id: I678a39e42a054bf5f6ef9c59d0fb93ff9719b964
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I1ed39c4ee084b26faac8286d9729413311ba9508
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I9dbeff51d3ede6db3cd5a097623aa580e5e25042
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I6e6963882825e83d8da3a460be35c7349e107777
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Ia495f8f50c43baf0d6eeb8e9ba04314ce277286f
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: If96f5a7c7e4b511cab3d57e5b57796aa516aff11
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Idda43d1236889ef91d8c37faf98ae23a19de688c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Ibf68423e9514b8e85cdf0a3e57ababd55dd4fcc4
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
To facilitate dispatch trajectory tracing, vlib_buffer_t decoding, etc.
through Wireshark
Change-Id: I31356b9fa1f40cba8830aaf10a86a9fbb7546438
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: Id4f37f5d4a03160572954a416efa1ef9b3d79ad1
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Typically we have scalar_size == 0, so it doesn't matter
but vlib_frame_args was providing pointer to scalar frame
data, not vector data. To avoid future confusion function
is renamed to vlib_frame_scalar_args(...)
Change-Id: I48b75523b46d487feea24f3f3cb10c528dde516f
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
(gdb) bt
bt
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
(gdb) frame 5
frame 5
293 if (PREDICT_FALSE (rxvq->last_avail_idx == rxvq->avail->idx))
(gdb) p *rxvq
p *rxvq
$3 = {cacheline0 = 0x7f290bcadd80 "\377\003", qsz_mask = 1023, last_avail_idx = 0, last_used_idx = 0, n_since_last_int = 0, desc = 0x0, avail = 0x0, used = 0x0, int_deadline = 0, started = 1 '\001', enabled = 1 '\001', log_used = 0 '\000', cacheline1 = 0x7f290bcaddc0 "\377\377\377\377\016", errfd = -1, callfd_idx = 14, kickfd_idx = 19, log_guest_addr = 5151049792, mode = 0}
The crash is because we access the null pointer rxvq->avail,
which is supposed to be derived from the mmap informed by the driver.
We fixed a similar issue before in
https://gerrit.fd.io/r/#/c/14545/
The reason was the driver ummaps the memory without doing the disconnect in
SR-IOV environment. The fixed was applied to the RX path. Now it happens in the
TX path. We just need to apply the same check in the TX path.
Change-Id: I7b1dfc96797cb5b52845bc6cec09a8c5d4325280
Signed-off-by: Steven <sluong@cisco.com>
|