Age | Commit message (Collapse) | Author | Files | Lines |
|
Move from synchronous flushing of io and ctrl events from transports to
applications to an async model via a new session_input input node that
runs in interrupt mode. Events are coalesced per application worker.
On the one hand, this helps by minimizing message queue locking churn.
And on the other, it opens the possibility for further optimizations of
event message generation, obviates need for rx rescheduling rpcs and is
a first step towards a fully async data/io rx path.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Id6bebcb65fc9feef8aa02ddf1af6d9ba6f6745ce
|
|
Type: improvement
Change-Id: I830f7a2ea3ac0aff5185698b9fa7a278c45116b0
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: improvement
Change-Id: Ie0bad9e03ac2e29da23af01ee7f63cb44489ad9c
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
load-balance and replicate dpos both store their number of buckets as
u16, which can overflow if too many paths are configured. For
load-balance it can happens quite quickly because of weights
normalization.
Type: fix
Change-Id: I0c78c39fc3d40626dfc58b49e7d99d71f9852b50
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
IV requirements vary wildly with the selected mode of operation. For
example, for AES-CBC the IV must be unpredictable whereas for AES
counter mode (CTR or GCM), it can be predictable but reusing an IV with
the same key material is catastrophic.
Because of that, it is hard to generate IV in a generic way, and it is
better left to the crypto user (eg. IPsec).
Type: improvement
Change-Id: I32689c591d8c6572b8d37c4d24f175ea6132d3ec
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Pool's pool_put_will_expand() calls clib_bitmap_will_expand(),
so every put except ones that leads to free_bitmap reallocation
will get false positive results and vice versa.
Unfortunatelly there's no related test and existing bitmap
tests are failing silently with false positive result as well.
Fortunatelly neither clib_bitmap_will_expand() nor
pool_put_will_expand() are being used by current vpp codebase.
Type: fix
Signed-off-by: Vladislav Grishenko <themiron@yandex-team.ru>
Change-Id: Id5bb900cf6a1b1002d37670f5c415c74165b5421
|
|
In clib_bitmap_set_region and clib_bitmap_set_multiple the index of
the last bit to set was off by 1. If this index was pointing to the
last bit of the bitmap, another uword would have been allocated,
even though it was unnecessary.
Moreover, in clib_bitmap_set_region, bits in the last word were not
properly set. Indeed, the n_bits_left value is wrong since n_bits
is not decreased by the number of already set bits.
Type: fix
Signed-off-by: Maxime Peim <mpeim@cisco.com>
Change-Id: I8d7ef6f47abb9f1f64f38297da2c59509d74dd72
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I035e3fdbb52eca010ad7b2c20ca2930cb1645978
|
|
otherwise the next time the counter is validated this is dangling.
Type: fix
Fixes: 58fd481d73
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: Ifa8d5ff27175cf6dfb30cbf023fa3251fe5c780e
|
|
Free up the vapi context in case of failure.
Type: fix
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
Change-Id: I4f64e8718014d714f1b82877e69d2354b5fa44fb
|
|
This patch adds performacne and functional tests for ip4
outbound traffic policy matching.
Test setup is configurable in startup.conf and though the test
parameters. Cache, fast path, fast path burst mode can be enabled
and disabled,
and performance for different lookup setup can be measured.
Type: feature
Signed-off-by: Piotr Bronowski <piotrx.bronowski@intel.com>
Change-Id: I1d04d196e412f47f43b7e5cbd46607bf6a9cc40e
|
|
Add vapi_connect_from_vpp() and vapi_disconnect_from_vpp()
calls to allow API clients from within VPP process.
Add a new memclnt_create version that gives the user a
knob to enable or disable dead client scans (keepalive).
Type: feature
Signed-off-by: Ole Troan <ot@cisco.com>
Change-Id: Id0b7bb89308db3a3aed2d3fcbedf4e1282dcd03f
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
In test_bihash_threads, if a test thread fails to be created,
it is still counted towards the total thread count, which could
lead to never-ending test loop. This patch fixes the issue.
Type: fix
Signed-off-by: Jing Peng <pj.hades@gmail.com>
Change-Id: Ic0f1d4dde9c5ea672b52f0e2e49f16d42f982b77
|
|
This patch fixes test_bihash_unittest in two ways:
1. The number of searches, namely tm->search_iter, defaults to 0,
thus disabling the test. This patch changes the default to 1.
2. Test errors are reported by clib_warning() instead of being
returned, thus the caller test/test_bihash.py is never aware of them.
This patch returns the errors constructed by clib_error_return().
Type: fix
Signed-off-by: Jing Peng <pj.hades@gmail.com>
Change-Id: I60e99a829ebe6aa2a56e7a9332cf973afa100311
|
|
Type: refactor
Change-Id: I5ca142ec1557d5b5c3806b43553ad9d3b5ea1112
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Use of _vec_len() to set vector length breaks address sanitizer.
Users should use vec_set_len(), vec_inc_len(), vec_dec_len () instead.
Type: improvement
Change-Id: I441ae948771eb21c23a61f3ff9163bdad74a2cb8
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
- 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>
|
|
- 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: refactor
Change-Id: Ifd533a095d979dc55bfbe5fac7e0b7510a4d900c
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
When the adj subsystem is notified of a BFD session, it attempts to find the appropriate adjacency from the session's key.
This could lead to a mismatch between the adj used by BFD and that of FIB. The BFD session stores the adj it is using, so FIB uses that instead.
Since adj is now using the same adj as BFD, it does not need to maintain its own locks.
In BFD it is necessary to initialise the adj index used in INVALID and ensure it is not unlock before listeners are notified of the session delete.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I9630867b10bb18969475299a0c754942a8df0f44
|
|
Type: test
The FIB UT fails in debug mode because there is no string associated woth its fib_node_type_t.
Change the tests to register their own type, which will give it a name.
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I36e546718faa7241c088494cbae10939aca51d5a
|
|
Prep for supporting multiple callbacks, optional args, etc.
Type: improvement
Change-Id: I96244c098712e8213374678623f12527b0e7f387
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: test
Change-Id: Ic9fddc9fedd5140984c5901c4cac53dec022dcec
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
On interface delete we were not removing
the lock taken by a previous ip_table_bind()
call thus preventing the VRFs to be removed.
Type: fix
Change-Id: I11abbb51a09b45cd3390b23d5d601d029c5ea485
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: improvement
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: Iec9dacde170533ca16e8117787e62da8af69ae96
|
|
Due to multiarch nature of that code, those tests doesn't bring much
value. New tests will be addes as part of refactor.
Type: refactor
Change-Id: I41056dc99d08cd6ca38f9e00e8cf6a465c90edb7
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
clib_strncat
There are 3 versions of the string functions. For example, for strcpy,
they are
1. strcpy(dst, src) -- the legacy unsafe version
2. strcpy_s(dst, dmax, src) -- C11 safeC version which has an addition argument
named dmax.
3. clib_strcpy(dst,src) -- clib version to enable legacy code that uses strcpy
to make use of strcpy_s without adding the additional argument, dmax, which is
required by the C11 safeC version.
The implementation for the clib version is to artificially provide dmax to
strcpy_s. In this case, it uses 4096 which assumes that if the legacy code
works without blowing up, it is likely to work with the clib version without
problem.
gcc-11 is getting smarter by checking if dmax is within the object's boundary.
When the object is declared as static array, it will flag a warning/error
if dmax is out of bound for the object since the real size of dst can be
determined at compile time.
There is no way to find the real size of dst if the object is dynamically
allocated at compile time. For this reason, we simply can't provide support
for the clib version of the function anymore. If any code is using the clib
version, the choice is to migrate to the safeC version.
Type: fix
Fixes: b0598497afde60146fe8480331c9f96e7a79475a
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: I99fa59c878331f995b734588cca3906a1d4782f5
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Id58ac44d45fb8b8a0d803f02e0242ec6f4b3db05
|
|
Type: fix
Change-Id: I2f7579980f081d3b0a8d19baade0a6599d55cf80
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
When an mfib entry was created with both paths and entry_flags then
the entry flags were being ignored. If there are no paths then the
flags were passed into mfib_table_entry_update, but in the case where
the entry didn't exist and there were paths and flags, the entry was
created within mfib_table_entry_paths_update() which used a default
of MFIB_ENTRY_FLAG_NONE.
Pass the flags through into the mfib_table_entry_paths_update fn. All
existing callers other than the create case will now pass in
MFIB_ENTRY_FLAG_NONE.
Type: fix
Signed-off-by: Paul Atkins <patkins@graphiant.com>
Change-Id: I256375ba2fa863a62a88474ce1ea6bf2accdd456
|
|
Type: test
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I5d44a6ea24e4aa0842024a0961f1fb22c6e6419a
|
|
Type: fix
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Id41e22345be3ec401813ba43ddc7d92666784eb4
|
|
This patches fixes an issue that could cause
fib locks to underflow: if an API user deletes
a fib and quickly recreates it, the fib may not
have been actually deleted. As a result, the
lock would not be incremented on the create call
leading to the fib potentially disappearing
afterwards - or to the lock to underflow when
the fib is deleted again.
In order to keep the existing API semantics,
we use the locks with API and CLI source as flags.
This means we need to use a different counter
for the interface-related locks.
This also prevents an issue where an interface being
bound to a vrf via API and released via CLI could
mess up the lock counter.
Finally, this will help with cleaning up the
interface-related locks on interface deletion
in a later patch.
Type: fix
Change-Id: I93030a7660646d6dd179ddf27fe4e708aa11b90e
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
|
|
_pool_init_fixed uses mmap to initialize a fixed-size and preallocated
pool, whose size is the sum of vector_size and free_index_size with
alignment to the CLIB_CACHE_LINE_BYTES and page size. In this way
vector_size equals to pool_header_t + vec_header_t + elt_size * max_elts
so moving to the end of the pool space should be pool_header_t pointer +
vector_size, instead of vec_header_t pointer + vector_size.
Simple code to reproduce this error:
u64 *pool;
pool_init_fixed(pool, 2042);
Improve unit test to cover this case
Type: fix
Signed-off-by: Jieqiang Wang <jieqiang.wang@arm.com>
Reviewed-by: Lijian Zhang <lijian.zhang@arm.com>
Reviewed-by: Tianyu Li <tianyu.li@arm.com>
Change-Id: If088ef89b3dcb2d874ee837ae9da60983b14615c
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Type: improvement
Change-Id: Ib7e2f5f314144064de7b6be0fade3db2f9c943fe
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: improvement
Linux uses pseudo header checksum when checksum of l4 is offloaded.
This patch adds similar support in virtual interfaces.
Change-Id: I6a94d1104e59356f95057e7c122e3be9cd8659a3
Signed-off-by: Aloys Augustin <aloaugus@cisco.com>
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
|
|
Exported entries are tracked only when the prefix found in the export
FIB is really attached, exporter tracker is not set if the export
entry is not valid for export, ex. for special FIB entries - default
route, zeronet, mcast and broadcast prefixes.
When imported entries need to be purged, such unset exporter tracker is
being removed by non-initialized index with absent delegate entries,
causing corresponding assert and crash.
Type: fix
Signed-off-by: Vladislav Grishenko <themiron@yandex-team.ru>
Change-Id: Ib24a2e7853a03a960577872480213e1e8097da5a
|
|
Type: refactor
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I4525bc669d1c42c41dbc398cf239a093b6853298
|
|
Type: feature
Change-Id: If0edbb21a0283d66c648a9e190d238c8cfa56353
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Type: feature
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I3652ae275385d9b1eb1b11f418e3a7e5fef2f556
|
|
Type: make
Change-Id: I2958e9eddadee6434766ecd3cdb3b9cea742ed64
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
The init of fib_index_by_sw_if_index wasn't
setting default value to 0. Which we now
need for setting interfaces unnumbered
Change-Id: Ie5be9b5e5373ef055557a871ad4d1c45fbfc1dee
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
vlib_buffer_chain_linearize() truncates partial data in chained buffers
in corner cases when current_data is negative.
Strengthen test cases to reproduce the errors and fix it.
Type: fix
Change-Id: Ida621923711c5755508224bdc3842b31003c6c0b
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: refactor
Change-Id: Id10cbf52e8f2dd809080a228d8fa282308be84ac
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: test
UT for patch: 4fc68ee
Signed-off-by: Neale Ranns <neale@graphiant.com>
Change-Id: I96fac9a6407a7f3ebeecc5e68a7683e541063de7
|
|
Test added to the unittest plugin / test_vlib.py
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I73445e57918347c102ff6f5e8c9ddb9bd96f1407
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Ife79871c6530d2cd485928fee465baf2c8957e11
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: I7f9c4b9b6e523ab549087ad21724f34f08fca793
|
|
By convention, connects segment manager will be first. Therefore it will
be the one with the first segment wherein lies the app's message queue.
Saves us the trouble of allocating it on first connect, if app started
by listening, and we no longer need to track if it's assignable to a
listener or if it can be removed.
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Iba9a8ffaab618eeb41ec2144dcfee62d006dc7a2
|
|
Type: feature
Change-Id: I36041fe5c5f0ff129aee42516189807e96f62123
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|