Age | Commit message (Collapse) | Author | Files | Lines |
|
Queue everything to ethernet_input to not break pcap rx trace.
Fix a minor typo in vmxnet3_api.c
Type: fix
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I7df7b3e20b525e3d6f7421c8e213b6541cdd15b6
|
|
Type: fix
Change-Id: I9f8a8da31624637c060a9757256589c8f372f192
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Type: fix
Fixes: 1404698df397bc4d3007daea41f52ad75ed4486c
Change-Id: Ibb77d95bb5e862cf5d1cb65843c526fd2e225c62
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
1. Add option '[gso-enabled]' in cli 'create interface virtio'
2. Add gso information in virtio_show()
Change-Id: I4eb58f4421325ef54a6a68c8341b3a6d3d68136a
Signed-off-by: Chenmin Sun <chenmin.sun@intel.com>
|
|
Change-Id: Ibd1012c2c1382de442ed294dd9e8a0ec89349fa1
Type: fix
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
|
|
Provide default packet_to_capture value. Display interface name
correctly for "pcap tx/rx trace status".
Type: fix
Signed-off-by: John Lo <loj@cisco.com>
Change-Id: I7064d0dbea236a9aff68bba7fbaf2c4a73b16c6f
Signed-off-by: John Lo <loj@cisco.com>
|
|
Type: feature
Change-Id: Ia9a5b1a6d85c7f3f2e6db583b33b5b20029a1e88
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Type: fix
Change-Id: I67b72b5ad03b972198c27bc0d927867f41b0e20b
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Type: fix
Change-Id: I1fa8c5326d6f22cfb8dd40e97d8a22d11a716922
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
* deactivtates the TODOs on doxygen (were empty)
* This move punt.md to readthedocs (should be the new
place for dev doc ?)
* Makes Handoff queue demo plugin a child of dev doc
in doxygen
Type: fix
Change-Id: I1f0476a911b35208212af8dd608bc76160efd22a
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Previous implementation of clib_rwlock_t used two spinlocks: one
writer lock, and one to guard the counter for the number of readers.
This implementation uses a single condition variable rw_cnt which
has the following properties:
if a writer has the rwlock, rw_cnt = -1
if the rwlock is free, rw_cnt = 0
otherwise, rw_cnt > 0 and rw_cnt = number of readers
rw_cnt will never be less than -1
Benchmarking:
The results below are the cycle counts from test_rwlock.c, configured so
that for 10000 iterations, 6 reader and 6 writer threads on separate cores
are spawned such that each writer thread increments a global counter
10000 times in each iteration. For Taishan, 4 reader and 4 writer
threads are spawned in each test.
x86 Xeon old rwlock: 12.473e8, 11.655e8, 13.201e8, 11.347e8, 13.182e8
x86 Xeon new rwlock: 5.881e8, 5.796e8, 6.536e8, 5.540e8, 5.890e8
Aarch64 ThX2* old rwlock: 9.263e7, 8.933e7, 9.074e7, 8.979e7, 9.378e7
Aarch64 ThX2* new rwlock: 7.221e7, 8.107e7, 7.515e7, 7.672e7, 7.386e7
A72 old rwlock: 3.268e6, 3.200e6, 3.086e6, 3.176e6, 3.170e6
A72 new rwlock: 1.261e6, 1.288e6, 1.251e6, 1.229e6, 1.234e6
*ThunderX2 used additional gcc options "-march=armv8.1-a+crc+crypto+lse"
Type: refactor
Change-Id: I7c347d3037b36205ab532cbcb52a374c846eb275
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
"timer.[ch]" used a signal handler to deliver timer callbacks. Without
indulging in a set of sigprocmask(...) system calls, it would be
unsafe to use the mechanism.
Rather than wait for another developer to accidentally open this
particular can of worms, best to remove the code. It's nothing more
than an attractive nuisance at this point.
Type: refactor
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Ia3e7b00a389c302b466605dff0c1bf3566b8dbbd
|
|
Type: fix
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Ie37ff66faba79e3b8f46c7a704137f9ef2acc773
|
|
Type: fix
Change-Id: I898de67c017c3a45bed123d81041b32b43f749d0
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: fix
Fixes: fe2fff37
this improves the tracing for dropped ARP packets
Change-Id: Iefd0391e349fc54f1beebda403b2349534b20c48
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Change-Id: Ib9164d8f6c681e8900e645306f3a2dc0ac0e40a8
Signed-off-by: Filip Varga <filipvarga89@gmail.com>
|
|
Tested performance of a CAS implementation (using __atomic_compare_exchange)
against a TAS implementation (using __atomic_exchange) using test_spinlock.c
and found some performance improvement.
Generated assembly for CAS and TAS implementations show that TAS always
executes with a load-store dependency, but CAS moves a branch condition
between the load and store so that only a load occurs when the lock is free.
Benchmarking:
The results below are the cycle counts from test_spinlock.c, configured so
that for 10000 iterations, 12 threads on separate cores are spawned, each of
which increments a global counter 10000 times in each iteration. For
A72, 8 threads are spawned in each test.
x86 Xeon TAS: 7.333e8, 7.605e8, 7.535e8, 7.485e8, 7.321e8
x86 Xeon CAS: 5.842e8, 5.433e8, 5.389e8, 5.983e8, 5.552e8
Aarch64 ThX2* TAS: 9.852e7, 10.209e7, 9.190e7, 9.600e7, 9.224e7
Aarch64 ThX2* CAS: 7.640e7, 7.486e7, 7.425e7, 7.269e7, 7.534e7
A72 TAS: 7.289e6, 6.963e6, 7.208e6, 6.976e6, 7.200e6
A72 CAS: 1.695e6, 1.608e6, 1.600e6, 1.634e6, 1.746e6
*ThunderX2 used additional gcc options "-march=armv8.1-a+crc+crypto+lse"
Type: refactor
Change-Id: Ic5cd97991804f6b012707fad1a5d1a6edb96cd3d
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Type:fix
Change-Id: I2f4224a53360f533c086ebde6a3056e60431da87
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Type: fix
Fixes: 83832e7ced8be8b7de394415feaba70c32e3c38d
Change-Id: I63130c442f71d1aef0d389ab00bac8092224bec2
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
The previous change seems to affect all CRCs.
Type: fix
Fixes: b504777e7f1c9728e65b874284b4dfd39359c8a8.
Change-Id: I65e27ce22020e5ebc764b5a51b5fc84992ddb40a
Signed-off-by: Vratko Polak <vrpolak@cisco.com>
|
|
- lb_vip_dump/lb_vip_details - get all vip.
- lb_as_dump/lb_as_details - get all as list per vip.
- adds api unit test.
- adds vpp_lb to test framework.
Ticket:
Type: feature
Change-Id: I24be50d62c5234f3535cc840603ddd9df7eb3f07
Signed-off-by: Hongjun Ni <hongjun.ni@intel.com>
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
Signed-off-by: Hongjun Ni <hongjun.ni@intel.com>
|
|
There is a need to be able to specifiy whether a prefix in
a request is to match exactly or if more specific
prefixes are also desired.
Todo: Uncomment defaults, once supported in vppapigen.
Type: feature
Change-Id: I74fdef0e89e3aefda822c7c0a477e22479297a90
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
Type: fix
Change-Id: I04f136a04bc022d223e4bcb5c59920bd1f1fd560
Signed-off-by: Filip Varga <filipvarga89@gmail.com>
|
|
Type: fix
Fixes: 22e9cfd760be613f33a4135e9247729b64619cc6
Change-Id: Idb198642e439dc3d54c04d8bde9d3e8382ebf830
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Add gso option in create vhost interface to support gso and checksum
offload.
Tested with the following startup options in qemu:
csum=on,gso=on,guest_csum=on,guest_tso4=on,guest_tso6=on,guest_ufo=on,
host_tso4=on,host_tso6=on,host_ufo=on
Type: feature
Change-Id: I9ba1ee33677a694c4a0dfe66e745b098995902b8
Signed-off-by: Steven Luong <sluong@cisco.com>
|
|
Type: feature
store: write a QoS value into the buffer meta-data
record: Extract a QoS value from a packet header and store it.
mark: Make a change to the content of a packet header by writing a stored
QoS value
Change-Id: I07d1e87dd1ca90d40ac1ae1774fee1b272cab83f
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: feature
Change-Id: I92757f041cde399229c42e34515ace0fcd37908b
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
add_vpp_packaging maccro was circumventing scripts/version causing:
1. build number not being appended to the package
2. invalid package name format (due to cmake cpack module)
Change-Id: I2b9a985357a3f3bd501be6db6ca638d4430f4bbb
Type: make
Fixes: def35a2352c9a54f748d301ffa47a446d25a83e0
Signed-off-by: YohanPipereau <ypiperea@cisco.com>
|
|
Enabling capture on pg with the file already existing
results in a misleading error message. Fix the text.
Change-Id: I1aea49cfeda3b4bfe6ed7b18fd543948a078508a
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Error index calculation is error_code + error_node->error_heap_index.
Type: fix
Fixes: gerrit 20802
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I66cf05a29b3cfd9ef9c5468e399290e862b784af
|
|
Spawns a uniform number of writer and reader threads across a number of
cores where each writer thread increments a global variable a specified
number of times, and the reader threads continually poll the global's
value until the writers complete.
Type: test
Change-Id: I979c3734c6d03139d0802bff1846875d226f6fbb
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Spinlock performance improved when implemented with compare_and_exchange
instead of test_and_set. All instances of test_and_set locks were refactored
to use clib_spinlock_t when possible. Some locks e.g. ssvm synchronize
between processes rather than threads, so they cannot directly use
clib_spinlock_t.
Type: refactor
Change-Id: Ia16b5d4cd49209b2b57b8df6c94615c28b11bb60
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Spawns a uniform number of threads across a number of cores where each
thread increments a global variable a specified number of times.
Type: test
Change-Id: I12b3a37708a199c297d022348d99dbb0e8349a9f
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Type:fix
Change-Id: I91dfe7e0ae2e632022fbf639ca16c93f570849de
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Now that we have support for f64:
- create explicit types for timestamp(datetime)/timedelta
- update log_details to use timestamp and remove redundant string representation.
If you need the string representation, in python do str(timestamp).
If you prefer the raw f64 value, the client can pass in the
_no_type_conversion option.
Type: feature
Change-Id: I547b5fa7122d2afa12628b7db0192c23babbbae8
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
Type: fix
Change-Id: I35838baea21ead4a3f45d998ff225a513781d7ee
Signed-off-by: Paul Vinciguerra <pvinci@vinciconsulting.com>
|
|
In a lot of places within the unit tests pg_start() starts
the capture with an already existing capture running
for the same test.
If the pcap file already exists, then it is renamed and there
is no problem.
However, there is a potential for race if the previous
pg_start() has enabled the capture, but the check for
renaming it happened just slightly earlier than the first
packet has arrived.
Then a second call to pg_start() will hit a check that
a file exists, and will cause an error. This is especially
visible when running the tests in parallel due to increased
load.
Solution: disable the capture before enabling it.
This will flush the aready running capture and eliminate the race.
The additional delay that flushing of the pcap creates has exposed
several other race conditions:
NAT tests: Some of the NAT reassembly tests
verify that the entries were added to the reassembly data structures,
but do so by comparing the quantities of entries. With the default
timeout being 2s, some of the entries might timeout,
resulting in a bogus test failure.
Solution: Bump the timeout to 20s for the affected tests.
Punt tests: nr_packets == 3 makes test intermittently fail,
nr_packets > 3 make it reliably fail, and nr_packets = 2 works
Solution: set nr_packets == 2 for the time being
IGMP tests: the leave-group calls get a spurious packet
from the time the new groups were configured
Solution: add 1 second delay before starting to delete the groups
Type: test
Change-Id: I931182a7b2860cf670e030ee7da8038f6e87356d
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Type: feature
Change-Id: Icb1bd3fce768aebf8919c63a104f771ca7fa1d6f
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Change-Id: I01eeedf8d5015b07b9422c65afe78bfe8177c22c
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Ticket: 1729
The flags that are permanently set on a path-list should form part of
its key in the path-list DB. Otherwise, if shared, they will not behave
as expected.
Change-Id: I0aa7c7c5d270c97b08014e4a47ddbdcee2358706
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Ticket: 1728
Change-Id: I679c2b8c5b0f751c9476db3669ab3f6c26dcdd28
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Change-Id: I0c9353598d3c9b7ea587ea8a2b6e1faa5454843d
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Change-Id: I29d24c8ec7b8e0613d4fbf5eedc72384326dc284
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Change-Id: I031a60ac010b55110f32f0a08e19b1156aeda268
Type: refactor
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
|
|
All instances of test_and_set locks used the following sequence
to release the locks:
CLIB_MEMORY_BARRIER ();
p->lock = 0; // p is a generic struct with a TAS lock
Use clib_atomic_release to generate more efficient assembly code.
Type: refactor
Change-Id: Idca3a38b1cf43578108bdd1afe83b6ebc17a4c68
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
Type: fix
Fixes: 99536f4
Change-Id: Ica230ec9fa7f6fd36e2754e8b0b9db555460ca55
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
Type: fix
Fixes: c30d87e6139c64eceade54972715b402c625763d
Change-Id: I86b606b18ff6a30709b7aff089fd5dd00103bd7f
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Modified test-and-set spin locks to call CLIB_PAUSE () when spinning
for code consistency. Decreases the memory bandwidth consumed.
Type: fix
Change-Id: I1cca4f87f44f23f257c7a35466cd2e7767072f51
Signed-off-by: Jason Zhang <jason.zhang2@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Lijian Zhang <Lijian.Zhang@arm.com>
|
|
To the sample plugin. We should probably suggest that folks use the
pipeline.h coding model more often. It's really easy, and these days
the performance results are similar to quad-single loop coding.
Type: refactor
Change-Id: Ie2caa087972737e6d9c31c4ac79355f3d8ced282
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Type:fix
Freeing mq messages in vpp (producer), if enqueueing fails, invalidates
consumer assumption that messages can be freed without a lock.
Change-Id: I748a33b8846597bdad865945d8e899346d482434
Signed-off-by: Florin Coras <fcoras@cisco.com>
Signed-off-by: Tal Saiag <tal.saiag@gmail.com>
|