None of the routine names in the backtrace exist in master/latest – it’s your
code - so it will be challenging for the community to help you.
See if you can repro the problem with a TAG=vpp_debug images (aka “make build”
not “make build-release”). If you’re lucky, one of the numerous ASSERTs
8 12:19 AM
To: 'Hu, Xuekun' ; Dave Barach (dbarach)
; vpp-dev@lists.fd.io
Subject: RE: [vpp-dev] About in-band telnet/ssh support of VPP
Partially completed work will also be welcome if you agree to share.
Or you can just push your temporary work to sandbox gerrit so that anyone could
get some idea
Please do not use the vpp host stack to listen to port 23 (telnet) on a
network-facing interface. You could do that, but please don't do that.
All you would need to add is a well-known default password, and you would have
created a super-trivial attack surface for your product.
Florin and I
You just demonstrated one of the basic properties of vector packet processing:
as the offered load increases, the cost per vector element decreases. Although
you didn’t explicitly report the vector sizes involved, the vector size
necessarily increases as the offered load increases. Anyhow, it’s
, see vlib_node_state_t
*/
void vlib_node_set_state (vlib_main_t * vm, u32 node_index, vlib_node_state_t
new_state)
-Original Message-
From: Prashant Upadhyaya
Sent: Thursday, November 22, 2018 8:53 AM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Difference
Use VLIB_NODE_TYPE_INPUT. Pre-input nodes - of which there is one - exist to
make sure that a certain epoll(...) call happens at the top of the loop.
D.
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Thursday, November 22, 2018 7:41 AM
To:
See
https://wiki.fd.io/view/VPP/Pulling,_Building,_Running,_Hacking_and_Pushing_VPP_Code#Pushing_Code_with_git_review
From: chetan bhasin
Sent: Tuesday, November 20, 2018 11:43 AM
To: Dave Barach (dbarach)
Subject: Re: [vpp-dev] Getting crash while running load on VPP18.01 for 6 hours
Thanks
See previous email about 18.01 support. Much has changed in the code base in
the last year. You would be well-advised to move to 18.10 at your earliest
possible convenience.
To start working out what’s wrong: grab one of the worker threads in gdb, and
see what it’s doing. Possibilities include
Several suggestions:
* Try a debug image (PLATFORM=vpp TAG=vpp_debug) so the crash will be more
enlightening
* Switch to 18.10. 18.01 is no longer supported. We don’t use the mheap.c
memory allocator anymore, and so on and so forth.
* See https://wiki.fd.io/view/VPP/BugReports
See https://wiki.fd.io/view/VPP/BugReports "Core Files" section.
From: vpp-dev@lists.fd.io On Behalf Of Rubina Bianchi
Sent: Monday, November 19, 2018 8:05 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] VPP coredump question
Hi Dear VPP
I have two question about vpp coredump.
How can I use
+1, "vpp# show hardware" and "$ cat /proc/cpuinfo" to make sure that the
selected cores are on the right NUMA socket...
HTH... Dave
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Vratko Polak -X
(vrpolak - PANTHEON TECHNOLOGIES at Cisco) via Lists.Fd.Io
Sent: Wednesday,
Try fixing this problem, reported in the Jenkins job log:
Identified problems
virl nfs mount failure
virl simulated node not mounting nfs directory properly
Indication
Please submit this as a draft patch, and add me as a reviewer. Thanks... Dave
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Monday, November 12, 2018 1:21 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Regarding dead client scan in shared memory
Yes, it should be easy to cherry-pick into any release where it's needed.
HTH... Dave
-Original Message-
From: Prashant Upadhyaya
Sent: Sunday, November 11, 2018 3:05 PM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Regarding communication from main thread
Check out src/plugins/perfmon/perfmon_periodic.c for one take on that problem...
Dave
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Sunday, November 11, 2018 2:37 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Regarding communication from main thread
Dear Ewan,
Please consider moving to vpp 18.10. 17.04 is not supported anymore.
Thanks... Dave
From: vpp-dev@lists.fd.io On Behalf Of xulang
Sent: Saturday, November 10, 2018 4:37 AM
To: xulang
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] vlib_buffer_alloc error
My version is 17.04.
At
In view of significant code contributions to the vpp project - see below - I'm
pleased to nominate Andrew Yourtchenko as a vpp project committer. I have high
confidence that he'll be a major asset to the project in a committer role.
Andrew has contributed 181 merged patches, including
+1.
If the vpp service is running - and you're using the distributed version of
/etc/vpp/startup.conf - you'll see this stanza:
api-segment {
gid vpp
}
Make sure that the user running vppctl is either in the vpp group, or is root.
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of
Sure: go look at .../src/vppinfra/{dlmalloc.[ch], mheap.[ch]}
From: vpp-dev@lists.fd.io On Behalf Of tianye@sina
Sent: Monday, November 5, 2018 9:23 PM
To: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] About memory debug
So, everybody here debug the memory issue only with code review? :)
That does
optimized by complier.
I am still wondering how different between using xxx_node.index and
node->node_index. Anyway,thanks, now I understand it is a performance
consideration.
Best Regards,
Kingwel
原始邮件
主题： RE: Incrementing node counters
来自： "Dave Barach (dbarach)"
发
See also
https://fdio-vpp.readthedocs.io/en/latest/gettingstarted/developers/vnet.html#creating-packets-from-scratch
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Friday, November 2, 2018 12:54 PM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev]
Yes, you missed something. This pattern is used in inline functions called with
compile-time constant values for is_ip6:
always_inline uword
ah_encrypt_inline (vlib_main_t * vm,
vlib_node_runtime_t * node, vlib_frame_t * from_frame,
int is_ip6)
VLIB_NODE_FN
# rm /dev/shm/* # and try it again.
From: vpp-dev@lists.fd.io On Behalf Of ???
Sent: Wednesday, October 31, 2018 11:29 PM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Run failed on 1807 version vpp code on centos7.
Hi,
make run didn't seem to work,
Logs :
[root@localhost vpp]# ls
When you register a feature node on the "device-input" arc, that node will
receive EVERY packet on all interfaces which have the feature enabled...
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Monday, October 29, 2018 10:11 PM
To: Damjan Marion
Please send the exact config you used, and the relevant packet tracer output.
From: wangchuan...@163.com
Sent: Tuesday, October 30, 2018 1:13 AM
To: Dave Barach (dbarach) ; vpp-dev
Subject: Re: RE: [vpp-dev] NAT44 && vxlan tunnel conflict
The attempt failed！ Adding static mappin
Cool, glad you found what you needed...
-Original Message-
From: Prashant Upadhyaya
Sent: Monday, October 29, 2018 11:48 AM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Regarding vlib_buffer_t construction for tx of a
self-made packet
Thanks a bunch Dave
The NAT plugin is tossing vxlan out-to-in packets. You’ll get different results
if you add a static mapping for (UDP, 4789) packets - maybe not the desired
result, but at least a different result...
D.
From: vpp-dev@lists.fd.io On Behalf Of
wangchuan...@163.com
Sent: Monday, October 29, 2018
Look at .../src/vnet/ipfix-export/flow_report.c : send_template_packet(...) for
a decent example of most of the mechanics involved. Also look at
.../src/vlib/buffer.c : vlib_buffer_add_data(...).
If you're going to send lots of such packets, it may well improve performance
to allocate a fair
Here’s one simple way to do it:
$ cat /etc/vpp/startup.conf
unix {
nodaemon
log /var/log/vpp/vpp.log
full-coredump
cli-listen /run/vpp/cli.sock
startup-config /etc/setup.gate
gid vpp
}
From: vpp-dev@lists.fd.io On Behalf Of tianye@sina
Sent: Friday, October 26, 2018 6:29 AM
To:
“# rm /dev/shm/*” and try again. The shared VM arena is persistent.
From: vpp-dev@lists.fd.io On Behalf Of siddarth rai
Sent: Thursday, October 25, 2018 8:44 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] VPP cores out of vl_client_connect()
Hi all,
I am facing VPP crash from vl_client_connect .
There is no such thing as a free lunch.
If you force a fresh allocation measured in gigabytes, the memory allocator
will mmap(...) a bunch of (4k) pages which will incur (expensive) pagefaults as
they’re populated. Vec_validate(...) copies data when necessary. When
structures grow to
“show memory” looks at every object in the heap, with packet processing
disabled for the duration.
From: vpp-dev@lists.fd.io On Behalf Of
david.leitch@gmail.com
Sent: Wednesday, October 24, 2018 4:58 AM
To: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Memory Performance issue #vpp
hi Matus
Use vlib_node_add_next(...) to create the graph arc at your convenience.
Memorize the arc index when you create it, so you can set e.g. next0 to the
correct value in MyNode.
HTH... Dave
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of Prashant Upadhyaya
Sent: Wednesday,
Read at face value, the vpp binary API memory segment (VA range 0x3000 and
above) is out of memory. The failing allocation is a request for 60 bytes,
which is not an unreasonable ask.
Before doing anything else, try increasing the size of the API segment and see
if the problem recurs.
We
See https://wiki.fd.io/view/VPP/VPP_Home_Gateway for a NAT-based example.
From: vpp-dev@lists.fd.io On Behalf Of
wangchuan...@163.com
Sent: Tuesday, October 23, 2018 2:55 AM
To: vpp-dev
Subject: [vpp-dev] How translate some ctrl data to OS ?
Hi all，
PCA have running a VPP that have only 1
See …/src/vlib/main.c, and
https://fdio-vpp.readthedocs.io/en/latest/gettingstarted/developers/vlib.html?highlight=dispatch#graph-node-dispatcher
At this level, we need to make a distinction between a vlib_pending_frame_t – a
vector of buffer indices to be processed by a certain node – and
Thanks for the report… This is the sort of thing we need to clean up. We do
very little negative testing of the form: send the kitchen sink with nothing
configured.
From: on behalf of JB
Date: Sunday, October 21, 2018 at 9:59 AM
To: "vpp-dev@lists.fd.io"
Subject: Re: [vpp-dev] VPP crash
Dear Ray,
As you probably suspect: plugins are neither unloaded nor reloaded.
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of Rui Cai via
Lists.Fd.Io
Sent: Thursday, October 18, 2018 9:43 PM
To: vpp-dev@lists.fd.io
Cc: vpp-dev@lists.fd.io
Subject: [SUSPICIOUS] [vpp-dev] Question about VPP
#include and try again?
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of carlito nueno
Sent: Wednesday, October 17, 2018 11:16 PM
To: hongjun...@intel.com
Cc: Mehran Memarnejad ; vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Unable to build router plugin
Hi Hongjun,
Thanks for
There’s no reason to economize when sizing the main heap for a vppinfra
application. It’s virtual memory, only pages which are in use are backed by
physical pages, etc.
From: on behalf of Aleksander Djuric
Date: Wednesday, October 10, 2018 at 3:29 AM
To: "vpp-dev@lists.fd.io"
Subject: Re:
Debug CLI commands are executed in “process” [aka cooperating multitasking
thread] context. As such, the stack has a fixed maximum size of 65K. See
…/src/vlib/unix/cli.c, search for “.process_log2_n_stack_bytes = 16”…
HTH… Dave
From: on behalf of Alok Makhariya
Date: Tuesday, October 9,
Note the vector size. Vpp isn’t even working hard enough to run efficiently.
The tx side of the interface you’re using appears to be throttled.
D.
From: Santosh Kumar Dornal
Sent: Friday, October 5, 2018 6:30 PM
To: j...@netgate.com
Cc: vpp-dev@lists.fd.io; Dave Barach (dbarach)
Subject: Re
I’ll let others jump in on the tx [and rx] drops. At some level, it looks the
vfio device simply isn’t disposing of traffic rapidly enough. Perhaps too few
buffers have been configured, but I don’t think that’s the most likely
explanation.
The “show run” stats are not helpful because they
accident might be involved, but I don’t see the problem.
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of Dave Barach via
Lists.Fd.Io
Sent: Wednesday, October 3, 2018 6:10 PM
To: Brendan Kelly
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] WRED
OK, let me see if I can repro it...
From
OK, let me see if I can repro it...
From: Brendan Kelly
Sent: Wednesday, October 3, 2018 6:03 PM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] WRED
This is what I gathered. I need to figure out how to get on Jira.
vpp# show version verbose
Version
, October 3, 2018 10:39 AM
To: brendanake...@yahoo.com
Cc: Dave Barach (dbarach) ; vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] WRED
Hello Dave, if you change the nsim delay values on the fly it appears vpp
crashes and all the state is lost.
Is this supported? Should I file a bug?
If so what logs
It would be easy enough for the delay simulator to enable/disable promiscuous
mode. But it’s handy to bridge the pipes: traffic flows whether or not the
delay simulation is enabled.
D.
From: vpp-dev@lists.fd.io On Behalf Of Dave Barach via
Lists.Fd.Io
Sent: Wednesday, October 3, 2018 9:09 AM
bin nsim_configure delay 10 ms bandwidth 10.1 gbit packet-size 128
bin nsim_enable_disable FortyGigabitEthernet86/0/1 TenGigabitEthernetd/0/0
HTH... Dave
From: Brendan Kelly
Sent: Tuesday, October 2, 2018 10:03 PM
To: vpp-dev@lists.fd.io; Dave Barach (dbarach)
Subject: Re: [vpp-dev] WRED
Hello
With respect to delay / loss impairments, please have a look at the “nsim”
plugin: src/plugins/nsim.
From: on behalf of "brendanakelly via Lists.Fd.Io"
Reply-To: "brendanake...@yahoo.com"
Date: Monday, October 1, 2018 at 6:18 PM
To: "vpp-dev@lists.fd.io"
Cc: "vpp-dev@lists.fd.io"
Subject:
).
D.
From: Burt Silverman
Sent: Sunday, September 30, 2018 9:36 PM
To: Dave Barach (dbarach)
Cc: ru...@microsoft.com; vpp-dev
Subject: Re: [vpp-dev] Question about ip4(6)-glean node behavior and packet
drops
I read that a traditional Cisco CEF router behaves as follows:
"When the
This is deliberate, traditional router behavior.
Alternatives involving queueing packets for reinjection after glean resolution
events give rise to resource exhaustion attacks. Mitigating that kind of attack
burns clock cycles, which in turn gives rise to a different kind of attack.
HTH...
Just to ask: a supported (64-bit ()) PCI-ID table, that plus vendor strings,
or something else?
Makes sense but as Damjan wrote we're swamped right now. If you have time, go
for it. Stick the results in the stats segment, maybe?
D.
-Original Message-
From: vpp-dev@lists.fd.io On
put inline and instantiating it.
Beyond that, Damjan is right: it's not a good idea, please don't go there.
HTH... Dave
-Original Message-
From: Damjan Marion
Sent: Saturday, September 22, 2018 10:31 AM
To: Nitin Saxena
Cc: Dave Barach (dbarach) ; vpp-dev
Subject: Re: [vpp-dev] Skipping
You would have to pull a bunch of work back into the device driver: counting IP
packet errors, head of ip4/6 unicast/multicast input feature arcs, and
mandatory input checks (expired TTL, ip header length, fragment).
Hardware offload will NOT accomplish all of these tasks. If by chance you can
Any address above the first 4gb would be fine. At this point, it’s a bit like
Avogadro’s number: when I see it, I know instantaneously which heap it came
from.
If e.g. clib_mem_is_heap_object(...) ASSERTs on such an address, I know that
some bit of code forgot “clib_mem_set_heap(...),” and so
+1. Note the vnet_buffer(b) and vnet_buffer2(b) macros. Track down the
definitions in .../src/vlib/buffer.h and .../src/vnet/buffer.h and you'll
understand how these opaque buffer metadata spaces are to be used.
Depending on where your graph nodes land, you must be careful not to smash
At that PPS rate, you don't need two worker threads. The worker threads burn a
bunch of cycles - poll-wait or not - doing next-to-nothing. Try running the
main thread all by itself...
D.
-Original Message-
From: Heckman, Steve
Sent: Tuesday, September 18, 2018 5:15 PM
To: Dave Barach
"show run" please...
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of sheckman
Sent: Tuesday, September 18, 2018 4:58 PM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Increase in main core CPU usage between 17.10 and 18.04
I've been seeing a dramatic increase in CPU usage by the
The typical coding pattern to use when you don't know how many messages will
come back: send a dump request, followed by an echo-ping.
When the ping reply shows up, the dump is complete...
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of
wangchuan...@163.com
Sent: Tuesday, September 11,
Please write up issues as described at: https://wiki.fd.io/view/VPP/BugReports.
Without even a hint as to which version of vpp is involved, this report is not
actionable.
Here’s an excerpt from the wiki page:
Although every situation is different, this page describes data which will help
In future, please refer to https://wiki.fd.io/view/VPP/BugReports. The issue
you're reporting is not trivially reproducible.
At a minimum: we'll need a backtrace from a debug image, the vpp version
string, the vpp startup config, and the connection method you used.
D.
From:
Folks,
I've run into this annoyance myself. Attempting to authenticate to wiki.fd.io
other than on the top-level page tends to produce "suspicious request, please
try again" complaints.
This seems like a configuration error. Please either fix it, or explain why it
cannot or should not be
Folks,
I've run into this annoyance myself. Attempting to authenticate to wiki.fd.io
other than on the top-level page tends to produce "suspicious request, please
try again" complaints.
This seems like a configuration error. Please either fix it, or explain why it
cannot or should not be
No, there isn’t. There isn’t much of a point in deleting graph arcs. If the
feature is reenabled, the existing arc from A to B will be reused.
D.
From: vpp-dev@lists.fd.io On Behalf Of Andrew Yourtchenko
Sent: Tuesday, September 4, 2018 2:50 PM
To: .ılı.D'p@k.ılı.
Cc: vpp-dev@lists.fd.io
If possible, repeat the exercise with a debug image. The code involved is well
tested, and does not crash without “help.” I wouldn’t be surprised to run into
an ASSERT elsewhere which explains the problem.
For the specific packet in question: see if
vnet_buffer(b)->sw_if_index[VLIB_TX] is set
Cool... Please let us know if you run into trouble...
Thanks... Dave
-Original Message-
From: Carlito Nueno
Sent: Monday, August 27, 2018 4:09 PM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] received signal SIGSEGV and vpp restarts continuously
Hi Dave,
I
Add graph node to the IP lookup feature arc. See attached code, from a
known-to-work example. Program src routes into a non-default FIB.
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of
georgi.mel...@gmail.com
Sent: Monday, August 27, 2018 1:28 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev]
Looks like a NULL pointer - please decode the pc. It would help if you could
repro the problem with a debug image and send a full backtrace.
vnet[10192]: received signal SIGSEGV, PC 0x7fa62aecade3, faulting address 0x0
D.
-Original Message-
From: vpp-dev@lists.fd.io On Behalf Of
Please refer to https://wiki.fd.io/view/VPP/BugReports for bug reporting, also
https://wiki.fd.io/view/VPP/Configuration_Tool
If you expect folks to volunteer to help you: write up your system
configuration, vpp version and configuration, and so forth to a reasonable
standard.
From:
Almost no matter what, vpp shouldn’t crash. Please at share the backtrace. See
https://wiki.fd.io/view/VPP/BugReports and also
https://wiki.fd.io/view/VPP/VPP_Home_Gateway for a known-to-work similar BVI /
IRB configuration.
D.
From: vpp-dev@lists.fd.io On Behalf Of Aleksander Djuric
Sent:
Dear Michael,
It is completely normal for a router data plane to forbid this scenario. Given
a packet to route to 13.13.13.114, which interface should the data plane choose?
>From what I can tell, it looks like you’re trying to route untagged as well as
>VLAN 11 and VLAN 12 packets. This
Try this in mycmd:
u8 * line;
if (unformat (input, "%U", unformat_line, ))
process_line;
Note that line will be a true u8 * vector: no null-termination. If you need
null termination: vec_add1 (line, 0);
Remember to vec_free(...) it unless you're planning to keep it.
HTH... Dave
The tw_timer implementation is extremely fast. We routinely test it to 60
million concurrent timers, the number required to bring up 10M TCP sessions.
Given a 500ms granularity, we have: max range = 86400 (secs/day) * 2 (clock
ticks/second) = 172800 slots. A 2-ring, 512 slots/ring geometry
Although I could explain how to fix that specific problem, please do not to use
17.10. We don't support 17.10 anymore.
The tcp stack worked better in 18.04, and much better than that in 18.07. I'd
strongly suggest that you switch to 18.07...
HTH... Dave
-Original Message-
From:
)
Sent: Saturday, August 11, 2018 5:24 AM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] format_bihash not printing all the elements
Hi Dave,
Could you please let me know what params do you use for testing the
“test_bihash_template” executable?
vec64 options is passing
Which version of vpp are you running? This should work fine in 18.07 or
master/latest...
From: vpp-dev@lists.fd.io On Behalf Of Richard Throne
Sent: Thursday, August 9, 2018 6:35 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Running VPP as non-root user
Hi,
I need to run VPP as non-root user.
You might consider sending e-mail to the dpdk mailing list, or e.g. Ray
Kinsella @ Intel. Pretty clearly a [dpdk] driver / HW issue.
From: vpp-dev@lists.fd.io On Behalf Of j...@yeager1.com
Sent: Wednesday, August 8, 2018 6:46 PM
To: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] Performance and CLI
What is the maximum value of j at the start of the inner loop? Note that kv.key
= i. The test replaces many, many keys on purpose.
Set TESTS += test_bihash_template in vppinfra.am
From: Vijay Katamreddy (vkatamre)
Sent: Wednesday, August 8, 2018 3:47 PM
To: Dave Barach (dbarach)
Cc: vpp-dev
No known issues at that level. Since this sounds like a test code, can you
share it?
Thanks... Dave
From: vpp-dev@lists.fd.io On Behalf Of Vijayabhaskar
Katamreddy via Lists.Fd.Io
Sent: Wednesday, August 8, 2018 3:14 PM
To: vpp-dev@lists.fd.io
Cc: vpp-dev@lists.fd.io
Subject: [vpp-dev]
Dear Gulakh,
As you move forward, please be careful. An arbitrary 10g NIC may or may not
have adequate PCI bus bandwidth to handle 10gb line-rate, full-duplex @ 64 byte
pkts. Depopulated memory channels, incorrect NUMA placement, and a host of
other configuration errors may yield awful
Trusty Tahr, aka Ubuntu 14.04, is no longer supported.
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#10059): https://lists.fd.io/g/vpp-dev/message/10059
Mute This Topic: https://lists.fd.io/mt/24220396/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Would it be possible to bump the priority of this ticket?
Thanks... Dave
From: Luke, Chris
Sent: Monday, August 6, 2018 12:30 PM
To: Dave Barach (dbarach)
Cc: vpp-dev@lists.fd.io
Subject: RE: git.fd.io expired certificate error
FWIW, this is why Coverity scans have stopped, too.
Chris (back
Please fix AYEC:
Chrome is complaining: "This server could not prove that it is git.fd.io; its
security certificate expired 3 days ago. This may be caused by a
misconfiguration or an attacker intercepting your connection. Your computer's
clock is currently set to Monday, August 6, 2018. Does
+1, the aggregate RX rate seems to be around 12 KPPS, the vector rate is small.
Absent I/O silliness, one core should handle this load with no problem.
D.
From: vpp-dev@lists.fd.io On Behalf Of Damjan Marion via
Lists.Fd.Io
Sent: Wednesday, August 1, 2018 4:27 PM
To:
Seems like you may have forgotten to add my_main->msg_id_base to mp->_vl_msg_id
when sending the reply message from your data plane plugin.
From: vpp-dev@lists.fd.io On Behalf Of Rubina Bianchi
Sent: Wednesday, August 1, 2018 6:42 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] Error in VPP DPDK
At least in C, it’s perfectly possible: use
vl_client_connect_to_vlib_no_rx_pthread(...).
Follow the sketch in the default rx_thread_fn(..) pretty carefully. You’ll need
to manually implement a non-while(1) version of vl_msg_api_queue_handler(...).
Spin-waiting for replies will completely
Before doing anything else: please revert to the previous DPDK version and see
if the issue vanishes.
From: vpp-dev@lists.fd.io On Behalf Of Peter Mikus via
Lists.Fd.Io
Sent: Monday, July 30, 2018 3:02 AM
To: vpp-dev@lists.fd.io
Cc: vpp-dev@lists.fd.io
Subject: [vpp-dev] CSIT -
I don't understand what you mean by "reserve symbols." Please explain what
you're trying to do in more detail.
From: vpp-dev@lists.fd.io On Behalf Of xulang
Sent: Sunday, July 29, 2018 2:43 AM
To: vpp-dev@lists.fd.io
Subject: [vpp-dev] target symbol
Hi all,
How can I reserve symbols in the
If you want to attempt a manual cherry-pick: https://gerrit.fd.io/r/#/c/13745 -
merged a couple of days ago - would be the right place to start.
D.
From: vpp-dev@lists.fd.io On Behalf Of Damjan Marion via
Lists.Fd.Io
Sent: Saturday, July 28, 2018 7:20 AM
To: Prashant Upadhyaya
Cc:
Guys,
The current behavior is probably not what the code author intended. It’s easy
to change. I’ve already offered under separate cover to push a draft patch
which won’t do this under any circumstances, so folks can kick the tires.
In its final form, a command-line argument of the form “...
Thanks very much... Dave
From: Tina Tsou
Sent: Friday, July 27, 2018 1:51 PM
To: Dave Barach (dbarach)
Cc: Brian Brooks ; vpp-dev@lists.fd.io
Subject: Re: Manual patch verify request: gerrit 13769
Dear Dave,
Looking into it...
Thank you,
Tina
On Jul 27, 2018, at 7:23 AM, Dave Barach
Folks,
Would it be possible for someone to download and manually verify that
https://gerrit.fd.io/r/#/c/13769 is functionally correct, and that the proposed
change isn't a performance disaster on aarch64?
Thanks... Dave
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this
If you send the vpp_api_test script sequence and the traffic profile - what the
packets which should hit but do not actually look like - I'll take a look. It
might be pretty simple...
D.
From: vpp-dev@lists.fd.io On Behalf Of Jan Gelety via
Lists.Fd.Io
Sent: Tuesday, July 24, 2018 10:43 AM
Please don’t remove extern declarations from header files...
Damjan is exactly right. It’s oh so easy to end up with N separate, independent
but identically-named data structure instances. It’s super-confusing when that
happens.
Thanks... Dave
From: vpp-dev@lists.fd.io On Behalf Of
I would expect very little if any measurable effect on forwarding performance.
At least for the counters that I moved to the stats segment: vpp was already
maintaining them. Not much changed.
HTH... Dave
From: vpp-dev@lists.fd.io On Behalf Of Dave Cornejo
Sent: Thursday, July 19, 2018 5:22 PM
We missed a step when pulling the release throttle which causes the JVPP tests
to fail.
As soon as we merge https://gerrit.fd.io/r/#/c/13522, order should be restored.
We need to wait until that one passes validation, for obvious reasons.
Thanks... Dave
-=-=-=-=-=-=-=-=-=-=-=-
Links: You
,I actually merged the patch. However, it was indeed Verified+1
:(
I have gone now to see the actual CLANG build failure and wondering: is it
possible that job is not configured as voting/gating job in Jenkins so the
verification process succeeds anyway?
- Marco
On Thu, 2018-07-12 at 22:49 +
Revert complete... HTH... Dave
From: Dave Barach (dbarach)
Sent: Thursday, July 12, 2018 5:01 PM
To: Damjan Marion (damarion) ; Florin Coras (fcoras)
; 'Marco Varlese' ; Ed Kern (ejk)
Cc: vpp-dev@lists.fd.io
Subject: Revert of gerrit 13408 / "vppinfra: AVX2 blend" in progress..
Folks,
Unfortunately, the AVX2 blend patch causes 100% clang validation failures on
unrelated patches. The clang validation job actually failed on the original
patch, but somehow fd.io JJB voted +1 anyhow.
See https://gerrit.fd.io/r/#/c/13457 - revert, https://gerrit.fd.io/r/#/c/13408
-
What you wrote is not correct.
DPDK drivers run in user-mode, as part of the vpp executable. To a first
approximation: device drivers use the uio_pci_generic kernel module to map
device registers into user-space. Once that happens, user-mode device drivers –
either dpdk or vpp native –
501 - 600 of 920 matches
Mail list logo
