Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I0d9af8e022cc25b8b4503f808f79028190475f8f
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I239143f920285050ad85244ece6893236195534f
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
This is preparation work for moving DPDK
EAL and mempool initialization to vlib.
Change-Id: I2dc35aa53edec0e07fc2650d630aa625831154c3
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
* compiler -march= parameter is changed from native to corei7
so code is always genereted with instructions which are available
on the Nehalem microarchitecture (up to SSE4.2)
* compiler -mtune= parameter is added so code is optimized for
corei7-avx which equals to Sandy Bridge microarchitecture
* set of macros is added which allows run-time detection of available
cpu instructions (e.g. clib_cpu_supports_avx())
* set of macros is added which allows us to clone graph node funcitons
where cloned function is optmized for different microarchitecture
Those macros are using following attributes:
__attribute__((flatten))
__attribute__((target("arch=core-avx2)))
I.e. If applied to foo_node_fn() macro will generate cloned
functions foo_node_fn_avx2() and foo_node_fn_avx512() (future)
It will also generate function void * foo_node_fn_multiarch_select()
which detects available instruction set and returns pointer to the
best matching function clone.
Change-Id: I2dce0ac92a5ede95fcb56f47f3d1f3c4c040bac0
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
vpp_restart solves the problem of restarting vpp and not exiting until
the new vpp instance is sufficiently established to allow immediate
binary API connections. The point is to avoid using arbitrary "sleep
N" commands e.g. in CSIT shell scripts.
We send SIGTERM to the current vpp process, and expect / depend on the
process-monitor du jour to remove the vpp-api shared-memory segment.
vpp_restart exits w/ status 0 if all is well, non-zero upon failure.
In trying to make vpp_restart reliable, we discovered a
recently-introduced heap corruption bug in .../dpdk/init.c, which
caused vpp to crash on startup once every 20 times on a particular
rig.
If possible, we should check /var/log/syslog for evidence of
unexpected multiple restarts during regression-testing.
Change-Id: Ic48c74b1a94a4368066ba2e326d4b12d51192088
Signed-off-by: Dave Barach <dbarach@cisco.com>
|
|
Change-Id: Id7ea0410d6a789000c747c36a7e50076e31dc29a
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: I88a8038a36e7e5060f354cd66a2532ef8c996b67
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Ib0a05f9d1b08bacef09f6d7c101391737031ee0d
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: I14693b81ecc0358683fd4dbfea41d0afc479ad51
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: If95b45c87f6d2f1d6dad9fdaee9430129b2f4962
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Change-Id: Ife9d074c83df739a39f363203fac149bfbc2c49b
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: I90983f3df94a3b28199908b29ffd8f827ab0c379
Signed-off-by: Damjan Marion <damarion@cisco.com>
|