Messages

151

Hardware and Performance / Re: Hyperthreading is vailable in 19.1.x?

« on: July 12, 2019, 11:09:55 am »

@mimugmail: sorry, but I think it does make sense to raise this topic in the hardware section for opnsense.

@tillsense: What you said, is a little bit unclear to me. I do know what Intel HTT does. I do understand, that HTT is a vendor-branding name for the concept called Symmetric Multithreading, SMT. What I dont really understand, how AMD implemented their own version of SMT.

Does that mean that 1 single core of the GX-412TC can execute 2 threads in parallel, but its not called as Hyperthreading, so the logical CPU count actually equals the physical CPU core count, and the SMT magic is hidden from the OS schedulers eyes?

I am curious whether machdep.hyperthreading_allowed="0" should be set, if HTT is not supported by this CPU, but multithreading is?

Update: according to wikipedia, Jaguar class CPUs (like the GX-412TC) do not feature Clustered MultiThread (CMT), the type of SMT we are speculating as incorrectly marked as HTT capability

Jaguar does not feature clustered multi-thread (CMT), meaning that execution resources are not shared between cores
Source:
https://en.m.wikipedia.org/wiki/Jaguar_(microarchitecture)

152

Hardware and Performance / Re: Hyperthreading is vailable in 19.1.x?

« on: July 11, 2019, 02:46:57 pm »

This has to be checked with the hardware vendor I'd guess

Actually, I just checked this strange behavior on my 11 yr old Intel Core 2 Duo machine. No Core 2 duo (or Quad, for the record) ever supported Hyperthreading. Thats for sure. Still, that CPU also reports "HTT" in the dmesg under Freebsd 12.0. But it reports only 2x CPU cores afterwards. So somebody is lying here... either BSD detects it totally wrong, or HTT doesnt mean HyperThreading in the CPU capability flag list.

153

Hardware and Performance / Re: Hyperthreading is vailable in 19.1.x?

« on: July 11, 2019, 10:27:31 am »

Isn't this Intel-only .. ?

Yes HT Technology is Intel and the equivalent for AMD it is SMT (i.e. multiple threads) but the SOC model above appears to have 4 cores on the chip and that's just physical cores not SMT.

I dont know, if HTT is restricted to be present only in intel CPUs or not, but this AMD CPU clearly adcertises that CPU feature flag. If it in fact does not support HTT, it should not advertise such capability.

154

Hardware and Performance / Hyperthreading is available in 19.1.x?

« on: July 10, 2019, 02:05:20 pm »

Hi all,

is hyperthreading (HTT) available in opnsense 19.1.x? The APU2 I have seems to have HTT capability as per the dmesg CPU feature flags:

CPU: AMD GX-412TC SOC                                (998.15-MHz K8-class CPU)
  Origin="AuthenticAMD"  Id=0x730f01  Family=0x16  Model=0x30  Stepping=1
  Features=0x178bfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,MMX,FXSR,SSE,SSE2,HTT>  -> see at the very end

FreeBSD/SMP: Multiprocessor System Detected: 4 CPUs
FreeBSD/SMP: 1 package(s) x 4 core(s)

 and sysctl machdep.hyperthreading_allowed: 1 allows it.

Still, I only see 4 physical CPU cores according to:

hw.ncpu: 4

Are there any settings in the OS that prohibit the usage of HTT? Or is this a coreboot BIOS defect?

155

19.1 Legacy Series / Re: CRASH in 19.1.8 when PPPOE refreshes

« on: July 06, 2019, 10:38:41 am »

Well, in that case you are safe. My cisco did not allow PPPoE frame to pass-through, so the opnsense router constantly tried to connect, but the modem never received those packets at all.

I know its totally out of scope for this community, but should I try to open a new thread for this tricky subject. Or all people will just ignore it here?

156

19.1 Legacy Series / Re: CRASH in 19.1.8 when PPPOE refreshes

« on: July 05, 2019, 11:18:56 am »

Be careful what kind of switch you put between the router NIC and the modem NIC. I used a cheap Cisco managed switch from ebay (2940/2950/2960 models, not the very latest models though), and it seems for some reason the PPPoE frames are not going through this setup. I created a new VLAN between the router WAN NIC and the modem, turned off Spanning tree, CDP, VTP etc. all the unnecessary junk, the interface config was basically empty apart from being set to static L2 access port in that specific VLAN. Not even PPPoE connect could be established. As soon I removed the UTP cable from the switch, and plugged it from the router straight to the modem, PPPoE session was established. All the glorious cisco CCIE experts colleagues around me couldnt figure out why a static L2 access port does not send through PPPoE frames (or at least they did not try it hard enough).
The LAN side of the router connecting to the same switch in another VLAN was working correctly for the rest of the  LAN clients

157

Hardware and Performance / Re: PC Engines APU2 1Gbit traffic not achievable

« on: February 25, 2019, 09:55:15 am »

Hi,
I've just found this blog entry: https://teklager.se/en/knowledge-base/apu2-1-gigabit-throughput-pfsense/
So the APU2 series should be able to achieve 1 gbit with pfsense. 

best regards
Dirk

IF(!!!) the wan type is NOT pppoe! That fact is not revealed in that blog. Can cause giant speed decrease, thanks to Freebsd pppoe handling defect.
I can only achieve 160-200 mbit, and that is fluctuating heavily between test runs. A cheap Asus RT-AC66U B1 can easily reach 800+ mbit on the very same modem/subscription.

158

19.1 Legacy Series / Netflow + insight local cache and CPU utilization

« on: February 14, 2019, 01:48:11 pm »

Hello all,

I am trying to find any documentation about the storage requirement for Netflow + the Insight collector, and the CPU load it generates to deal with this type of workload.

If I recall correctly, I read somewhere (unfortunately I am unable to find it anymore, where exactly) that the max storage that the /var/netflow/ files occupy should be max. around 100 MB. In contrast, for the moment it is about total of 700 MB after a week of run. There is 1 file (I cant recall its name) which is 500+ MB, the rest are much smaller.
I use tmpfs on /var, and also tried to disable the backup of netflow data during reboot, as it takes ages to backup this 700+ MB before shutdown and restore again when it boots, plus it wears my tiny 16 GB SSD.

The second part of my question: there is a periodic load spike, when a python process eats 1 of the 4 cores for 5-6 seconds, then goes down to 0. After 10-12 seconds, it is repeated. Happens during low traffic periods as well. I assume that python process has to do something with this Netflow/Insight workload. Is it taking such huge CPU load due to the large 700 MB+ database, and should decrease if the database is much smaller (e.g. after a fresh reboot with 0 database content)?

159

Development and Code Review / Re: APU LEDs Plugin

« on: February 07, 2019, 07:53:04 pm »

I would prefer a script to keep it simple, yes.


Cheers,
Franco

Some hints:
-  WAN UP/DOWN trigger LED-changes
- system load over x / load below y trigger LED-changes
- WAN interface traffic in the past minute above X bps / below Y bps trigger LED-changes (or define 10 ranges: 1) at max. 10% bandwidth utilization 1 blink/sec, 2) at max  20% bw util 2 blinks/sec, etc. and finally over 90% bw util, 9 blinks/sec)
- free disk space is low trigger LED-changes

My idea is to have the scripts selected from the above list, 1 script
assigned to each LED device, max of 3 scripts active at any time.

160

Hardware and Performance / Re: Poor NIC performance on APU2c4 board

« on: February 07, 2019, 07:18:25 pm »

THANKS! That was the problem!

Checking a LAN Client on the other side of the APU board:

$ iperf -c 172.20.1.1 -p 4712 -u -t 60 -i 10 -b 1000M
------------------------------------------------------------
Client connecting to 172.20.1.1, UDP port 4712
Sending 1470 byte datagrams, IPG target: 11.22 us (kalman adjust)
UDP buffer size: 9.00 KByte (default)
------------------------------------------------------------
[  4] local 10.0.0.10 port 63781 connected with 172.20.1.1 port 4712
[ ID] Interval       Transfer     Bandwidth
[  4]  0.0-10.0 sec  1.22 GBytes  1.05 Gbits/sec
[  4] 10.0-20.0 sec  1.22 GBytes  1.05 Gbits/sec
[  4] 20.0-30.0 sec  1.22 GBytes  1.05 Gbits/sec
[  4] 30.0-40.0 sec  1.22 GBytes  1.05 Gbits/sec
[  4] 40.0-50.0 sec  1.22 GBytes  1.05 Gbits/sec
[  4] 50.0-60.0 sec  1.22 GBytes  1.05 Gbits/sec

 

Even with some IDS rules enabled 

Your results look highly unlikely!

Are you performing network address translation (a.k.a NAT) in your firewall config (the builtin "pf" firewall by default does NAT,  unless you switched it off explicitely). Otherwise you are just routing traffic. There is no problem routing @1Gbit speed on this platform. Its the NAT, that kills the performance miserably!

161

Development and Code Review / Re: APU LEDs Plugin

« on: February 04, 2019, 12:40:11 pm »

The driver is there now by default and can be used, but the plugin path needs to be evaluated still... it's a bit heavy for my taste with a C++ daemon required.


Cheers,
Franco

At least can it be tested via CLI, how?

162

Development and Code Review / Re: APU LEDs Plugin

« on: February 04, 2019, 11:48:45 am »

I am also very interested, eagerly waiting for this feature to be shipped since 18.1.

Is there any description what is possible / what has been implemented as part of "APU LED" ?

LED sequence During startup phase / LED sequence normal working phase / LED sequence during shutdown phase? 

163

Documentation and Translation / Re: cron log

« on: December 17, 2018, 11:32:14 am »

We do not have a cron log. For all the activities the log locations are System or Backend log.


Cheers,
Franco

Is it intentional to not have log about cron activity? How can custom cron jobs be debugged, as I could not see any cron-related entry neither in System nor in Backend?

164

Documentation and Translation / Re: cron log

« on: December 10, 2018, 09:33:56 pm »

Am I missing something obvious, or that question is really that impossible to answer?

/var/log/<something with the name "cron" in it> is indeed a valid file in normal freebsd, so my question cannot be that stupid.

165

Hardware and Performance / Re: APU2 Bios

« on: December 01, 2018, 04:21:14 pm »

"Not only in 4.8.0.6.
At least the whole 4.8.x releases have these bugs, they just weren't added to the previous releases after they were found."

Exactly.
There may be other open issues in 4.8 that are already known, but not yet published. So I dont recommend 4.8.x, personally I stick to 4.0 as long as necessary. Dont want to risk my firewall crash because of these BIOS issues, it is rock solid (at least the BIOS) on 4.0.19.