Age | Commit message (Collapse) | Author | Files | Lines |
|
Inspired by a real-life conundrum: scenario X involves a vpp crash in
ip4-load-balance because vnet_buffer(b)->ip.adj_index[VLIB_TX] is
(still) set to ~0.
The problem takes most of a day to occur, and we need to see the
broken packet's graph trajectory, metadata, etc. to understand the
problem.
Fix a signed/unsigned ASSERT bug in vlib_get_trace_count().
Rename elog_post_mortem_dump() -> vlib_post_mortem_dump(), add
dispatch trace post-mortem dump.
Add FILTER_FLAG_POST_MORTEM so we can (putatively) capture a ludicrous
number of buffer traces, without actually using more than one dispatch
cycle's worth of memory.
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: If093202ef071df46e290370bd9b33bf6560d30e6
|
|
Type: improvement
Change-Id: Iafcaeb88fa3569e042fb9cb270cf062b145416e2
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Signed-off-by: jiangxiaoming <jiangxiaoming@outlook.com>
Change-Id: Id61966048875ca56a416caa520125c8c2366bf7b
|
|
This change mitigates software faults issued by some versions of the
linux kernel vfio-pci driver when VF PCI BARs are used without setting
the memory enable bit in the PCI configuration. This problem is
mentionned in https://lkml.org/lkml/2020/6/25/628
Change-Id: Idc177be4a5adb6ee467b4dd8f055f133ff267fe1
Type: improvement
Signed-off-by: Mohammed Hawari <mohammed@hawari.fr>
|
|
logging {
class avf {
rate-limit 200
level debug
syslog-level debug
}
class dpdk/cryptodev {
rate-limit 200
level disabled
syslog-level disabled
}
}
Type: improvement
Change-Id: Iebe49b3ed4dfbfd589e80fcd2f9a63a8fe66b83b
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
If VPP is started in interactive mode, instead of sending logs to syslog
server we print them directly to stderr.
Output is colorized, but that can be turned off with unix { nocolor }
Type: improvement
Change-Id: I9a0f0803e4cba2849a6efa0b6a86b9614ed33ced
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
create tap
delete tap tap0
create loopback interface
delete loopback interface intfc loop0
create tap <- here I hit a segmentation fault
Type: fix
Fixes: 148c7b768721231325a349fa82db693190513b53
Signed-off-by: Ole Troan <ot@cisco.com>
Change-Id: If270cc57610e2ded7422e7030d8c1f9fadd4d233
|
|
Type: improvement
Change-Id: I2a176fe2871d2e54b010bffc4f1f7a3616f0c455
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Allows us to declare plugin registrations in a non-disgusting way:
var plugin_reg = vpp.PluginRegistration{
Description: "The CGO plugin",
Version: "My Version",
Overrides: "sample_plugin.so",
}
It turns out that the specific compiler setup generates (.data section
offset, length) pairs in the .vlib_plugin_r2 section:
Contents of section .vlib_plugin_r2:
1ba9d0 00000000 00000000 50a81800 00000000 ........P.......
1ba9e0 0a000000 00000000 00000000 00000000 ................
1ba9f0 00000000 00000000 00000000 00000000 ................
1baa00 00000000 00000000 00000000 00000000 ................
1baa10 00000000 00000000 70a81800 00000000 ........p.......
1baa20 0e000000 00000000 ........
Contents of section .data:
18a800 00a81800 00000000 00000000 00000000 ................
18a810 00000000 00000000 00000000 00000000 ................
18a820 00000000 00000000 00000000 00000000 ................
18a830 00000000 00000000 00000000 00000000 ................
18a840 00000000 00000000 14000000 00000000 ................
18a850 4d792056 65727369 6f6e0000 00000000 My Version......
18a860 00000000 00000000 14000000 00000000 ................
18a870 54686520 45474f20 706c7567 696e0000 The CGO plugin..
18a880 00000000 00000000 0c000000 00000000 ................
<etc>
Unfortunately, it seems impossible to torture clang / gcc into
producing anything like this. This patch fabricates a plausible
vlib_plugin_registration_t from the so-called vlib_plugin_r2_t.
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I8c0c5a24f3b7bfea07d5181a7250b3d9685e8446
|
|
Type: improvement
Signed-off-by: Florin Coras <fcoras@cisco.com>
Change-Id: Ia884d745a11565918c7bf89435ceb0e17b6aef59
|
|
This adds a new data model for counters.
Specifying the errors severity and unit.
A later patch will update vpp_get_stats to take advantage of this.
Only the map plugin is updates as an example.
New .api language:
A new "counters" keyword to define counter sets.
counters map {
none {
severity info;
type counter64;
units "packets";
description "valid MAP packets";
};
bad_protocol {
severity error;
type counter64;
units "packets";
description "bad protocol";
};
};
Each counter has 4 keywords. severity, which is one of error, info or warn.
A type, which is one of counter64 or gauge64.
units, which is a text field using units from YANG.
paths {
"/err/ip4-map" "map";
"/err/ip6-map" "map";
"/err/ip4-t-map" "map";
"/err/ip6-t-map" "map";
};
A new paths keyword that maps the counter-set to a path in the stats segment KV store.
Updated VPP CLI to include severity so user can see error counter severity.
DBGvpp# show errors
Count Node Reason Severity
13 ethernet-input no error error
Type: feature
Signed-off-by: Ole Troan <ot@cisco.com>
Change-Id: Ib2177543f49d4c3aef4d7fa72476cff2068f7771
Signed-off-by: Ole Troan <ot@cisco.com>
|
|
Turns out that clang is a bit too enthusiastic about mapping static
variables to registers, which makes it hard to extract (especially)
the faulting VA from an optimized core file.
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I836e5d9695aeb4c5ee4a27f9565acf08ee4eaac0
|
|
Type: improvement
Change-Id: Ibc15aa6f75e989b2b7caa9a57e3665b1ffc7cc18
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
This patch adds smal header in front of dlmalloc space, and it stores
some additional information about the heap.
Immediate benefit of this patch is that we know the underlying page size
si we can display heap page statistics / real memory usage.
Type: improvement
Change-Id: Ibd6989cc2f2f64630ab08734c9552e15029c5f3f
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
vlib_main_t *vlib_get_main_not_inline(void)
vlib_thread_main_t *vlib_get_thread_main_not_inline(void)
elog_main_t *vlib_get_elog_main_not_inline()
Type: refactor
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I6de306d567283ad28ef34c9be0cf27452aecbf6c
|
|
Also spiffed up the vpp_api_test plugin loader so it executes
VLIB_INIT_FUNCTIONs and VLIB_API_INIT_FUNCTIONs.
Type: feature
Change-Id: Id9a4f455d73738c41bcfea220df2112bb9679681
Signed-off-by: Jon Loeliger <jdl@netgate.com>
Signed-off-by: Ole Troan <ot@cisco.com>
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Fixes an obscure issue when spinning up pure pthreads which don't have
vlib_main_t clones.
VLIB_REGISTER_THREAD () = {
<snip>
.use_pthreads = 1,
.no_data_structure_clone = 1,
}
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Ifae5a2a35865de057eea7782ee71b1011a21c88e
|
|
Type: improvement
Change-Id: Ie04302c576869bc7bfaa9f13ed2ea8a403a393d4
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Change-Id: Iaf60ec93ee9750074f52ba4dbe955355de5dbccc
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
- it is confusing from end consumer perspective that some thing
is somewhere called heap and somewhere mspace
- this is base for additional work where heap pointer is not the same
thing like mspace
Type: improvement
Change-Id: I644d5a0de17690d65d164d8cec3c5654571629ef
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Signed-off-by: Mohsin Kazmi <sykazmi@cisco.com>
Change-Id: I788678d85dfa1d5b99497013806047440028eb97
|
|
Currently there is no way to decrease simple counter. To extend
counters API new method allows to decrement counter by given value.
This might be useful in implementation of statistics stored in stats
segment.
Type: improvement
Change-Id: I7c08c62bffa6a2d50e9e2cf884f7b2d48538f34b
Signed-off-by: Sergey Matov <sergey.matov@travelping.com>
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Type: improvement
Change-Id: I04752c011e4ca58f56aa53f6ae27bae93a5c4590
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Heap may use different page sizes so we will not be able to create
stack protection page.
Type: improvement
Change-Id: Ibb35c9f0a151c464ee0167d17f2bd773ef6f530b
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Change-Id: I381fc3dec8580208d0e24637d791af69011aa83b
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Change-Id: If2e907967c9b75997b581ff0c058bd5c15e823f5
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
To hold more data later...
Type: improvement
Change-Id: I4006d22dcacd788988c4907f2c263fd4e4a9d398
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: improvement
Change-Id: Ib64c9b8207776986656e5a26c13a221edc6cc950
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Change-Id: I1b4f52e186165b04db5bd5f11058dc77b647bc94
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: improvement
Change-Id: I13517bff4129497644e0efc3495d0ee7897fe5a7
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Calling vlib_get_node_by_name via the VPE api
doesn't work due to hash weirdness. Haven't
gotten around the real cause of this. But this
fixes it.
Change-Id: I89f95dba2bcd9573b8f1f435e063e9dd57f9ca93
Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
|
|
Adjust buffer allocation so it always have odd number of cache lines.
That should result in better distribution of cachelines among cache sets.
Type: improvement
Change-Id: I0d39d4cf01cff36ad6f70a700730823a96448c22
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Scenarios where SIGHUP is sent would include the user closing an xterm
while in interactive/nodaemon mode, or similarly when running vpp in the
same mode during testing (e.g., using ssh to run VPP on a DUT). VPP
should exit in these cases; however, generating a core is unwanted.
Type: fix
Signed-off-by: Christian Hopps <chopps@labn.net>
Change-Id: Ibccfe5e676547e913c8a205ff16ab56d9abb1c82
|
|
Type: fix
adding routes should be MP safe. When new prefixes with differrent
prefix lengths are added, adjust the sorted list in an MP safe way.
Change-Id: Ib73a3c84d01eb86d17f8e79ea2bd2505dd9afb3d
Signed-off-by: Neale Ranns <nranns@cisco.com>
|
|
There is a number of TERMs with big length, such as
'screen.konsole-256color' (23). These TERMs can not
be processed properly by vpp because maximum telnet
byte stream supported is 24 and we need 6 more service
bytes to send TTYPE.
So I extended maximum depth guard to 32.
Type: fix
Signed-off-by: Vladimir Isaev <visaev@netgate.com>
Change-Id: I9ca506996a97e9567d06483c5f020d6cc394329c
|
|
Make it shorter and same format when converting to biggor or smaller
types.
Type: refactor
Change-Id: I443d67e18ae65d779b4d9a0dce5406f7d9f0e4ac
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Ibad744788e200ce012ad88ff59c2c34920742454
|
|
Type: fix
Change-Id: Iab512ba8c72c9e20aeba2d4265276bcabf095d46
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Callbacks for monitoring and performance measurement:
- Add new callback list type, with context
- Add callbacks for API, CLI, and barrier sync
- Modify node dispatch callback to pass plugin-specific context
- Modify perfmon plugin to keep PMC samples local to the plugin
- Include process nodes in dispatch callback
- Pass dispatch function return value to callback
Type: refactor
Signed-off-by: Tom Seidenberg <tseidenb@cisco.com>
Change-Id: I28b06c58490611e08d76ff5b01b2347ba2109b22
|
|
clang tends to force alignment of all sections when compiling for
address sanitizer, confusing VPP plugin infra.
On the contrary, GCC does not support this attribute on sections.
Selectively enable it depending upon the compiler.
Type: fix
Fixes: cea46522e79637f6ec37c03ec3fbeb87b160a378
Change-Id: I2dd6e52e70e6b4d89c101171cafc813e175ec472
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Type: feature
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I586547508003b95eaa74e18e4a5ac6f72986822c
|
|
Type: fix
Change-Id: If8dbbcb46193fd057fe8d704058609a3a8787d6c
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
Deal with circular macro definitions instead of crashing due to stack
overflow.
Separate macro tables, per CLI session
Add documentation to the Sphinx docs
Type: improvement
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I55fc9152bd37ad0c15fa3959f38b07b63100e634
|
|
All of the pieces have been sitting around for years.
Added several debug CLI commands:
"define <variable-name> <value>"
"undefine <variable-name>"
"show macro [noeval]"
"echo <whatever>"
Macros may refer to other macros. To defer evaluation:
"define foo \$(bar)" or some such.
The macro evaluator is not smart about "define foo \$(foo)" or more
complicated circular definitions, so don't do that.
Environment variables are available, simply use $<name-of-environment-vbl>
The macro expander has a table of (overrideable) builtin names, which
are evaluated by calling functions.
Simple example:
echo $USER
define ip1 192.168.1.1/24
define ip2 192.168.2.1/24
loop create
loop create
set int ip address loop0 $ip1
set int ip address loop1 $ip2
show int addr
show macro
undefine ip1
undefine ip2
Type: feature
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I08a800647bac573d8ae3cfd75c40061d41c5f976
|
|
This patch documents the DNS plugin, vlib packet tracing, vnet feature
arcs, and the vppinfra library.
Type: docs
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Idb16ec0593d428ce78fe8404374a823a56b4b420
|
|
Type: improvement
Change-Id: I81df4b61d1f0b8c1df77c1ee9bebcb491e155b69
Signed-off-by: Damjan Marion <damarion@cisco.com>
|
|
Type: fix
Change-Id: I4a93e1d9936414c514cb237a22624986b3ef5b3d
Signed-off-by: Benoît Ganne <bganne@cisco.com>
|
|
When recycling a graph node vnet_register_interface, it is missing an
explicit call to vlib_worker_thread_node_runtime_update(). However,
there is an implicit call to vlib_worker_thread_node_runtime_update()
via vnet_sw_interface_set_flags_helper() if it enables a new feature on
the interface for the first time. But that implicit call is not
guaranteed. For example, if an interface is created, deleted, and
created, then it may skip the implicit call to
vlib_worker_thread_node_runtime_update(). When that happens, the graph
nodes on thread 0 are not sync'ed to the worker threads. So the worker
thread's graph nodes are out of sync momentarily with the main thread's
graph nodes until some other event happens which calls for a sync is
needed. During this window, the worker thread's graph node is
vulnerable and may experience a crash.
When deleting a graph node, we never trigger a sync to the worker
thread. A patch was committed 3 years ago via
https://gerrit.fd.io/r/c/vpp/+/7523 to fix a show run crash. In
hindsight, the approach taken by 7523 is not orthogonal. While at it,
let's fix it right for both issues with a call to
vlib_worker_thread_node_runtime_update() in the appropriate place and
remove 7523.
Type: fix
Ticket: VPPSUPP-86
Fixes: gerrit 7523 / 19e9d954bd9eb4f04d48640d6540198e84ef65d7
Signed-off-by: Steven Luong <sluong@cisco.com>
Change-Id: Ic9472bd2d3a212dbfeceb526506ed0400983a142
|
|
Packet tracing performance doesn't justify inlining
vlib_add_trace(...) over 500 times.
It makes a 15% text-segment size difference in a representative use-case:
Inline:
$ size .../vnet_skx.dir/ipsec/ipsec_input.c.o
text data bss dec hex filename
6831 80 0 6911 1aff .../vnet_skx.dir/ipsec/ipsec_input.c.o
Not inline:
$ size .../vnet_skx.dir/ipsec/ipsec_input.c.o
text data bss dec hex filename
5776 80 0 5856 16e0 .../vnet_skx.dir/ipsec/ipsec_input.c.o
Retain the original code as vlib_add_trace_inline, instantiate once as
vlib_add_trace.
Type: refactor
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: Iaf431dbf00c4aad03663d86f9dd1322e84d03962
|
|
Fix a nit warning: we're not likely to create a vlib process with more
than 4gb of stack.
Type: fix
Ticket: VPP-1888
Signed-off-by: Dave Barach <dave@barachs.net>
Change-Id: I8bc7f64287c2802b0c286ce3d04443ac723a9a33
|