Re: CPU hangs on Celeron

Even if it the tests are just glimpses to check if it works, the results seem OK.
At the moment, for the latency test, I do not understand how a max overall latency of  15206 ns, against a period of 100000 ns, 
could have been seen as an overrun. To have an overrun the latency should exceed 100000 ns.
What happens if you relaunch it consecutively a few times after the very first one?
Can you run latency for a significantly longer time, say 2 hours at least, intense Linux activities, even of a simply settable kind, e.g.,  since you are running two CPUs,
a couple of "ping -f somewhere" , Linux make (while "true"; do make clean; make -j2; done). 

Furthermore I do not see why 5.0 does not stall and 4.1 does. There have been improvements but none is a the bare level of solving a stall a stall of the type you reported. The very core stuff is the same for both.

Paolo
________________________________________
From: [email protected] [[email protected]]
Sent: Saturday, December 5, 2015 8:08 PM
To: Paolo Mantegazza; [email protected]
Subject: Re: [Rtai]  CPU hangs on Celeron

Hello!

Thanks for your advise about the new version.
I tested it and now it does not stall anymore (kernel 3.16.7 with
rtai-5.0-test1 and gcc 4.9.2).

Can you please have a quick look at the results of the testsuit.
I am not sure if 27 overruns is ok or there is still a problem.

Thanks!

regards,
Christoph



root@AD502:/usr/realtime/testsuite/user/latency# ./run
*
*
* Type ^C to stop this application.
*
*

## RTAI latency calibration tool ##
# period = 100000 (ns)
# average time = 1 (s)
# use the FPU
# start the timer
# timer_mode is oneshot

RTAI Testsuite - USER latency (all data in nanoseconds)
2015/12/5 20:05:03
RTH|    lat min|    ovl min|    lat avg|    lat max|    ovl max| overruns
RTD|      -1377|      -1377|      -1009|      13831| 13831|         27
RTD|       -831|      -1377|       -754|      15206| 15206|         27
RTD|       -780|      -1377|       -758|        677| 15206|         27
RTD|       -789|      -1377|       -754|       2138| 15206|         27
RTD|       -780|      -1377|       -758|        711| 15206|         27
RTD|       -779|      -1377|       -760|        -70| 15206|         27
RTD|       -777|      -1377|       -751|       8512| 15206|         27
RTD|       -801|      -1377|       -751|       9202| 15206|         27
RTD|       -800|      -1377|       -764|        -51| 15206|         27
RTD|       -786|      -1377|       -764|        -36| 15206|         27
RTD|       -820|      -1377|       -764|        -82| 15206|         27
RTD|       -845|      -1377|       -794|       -293| 15206|         27
RTD|       -795|      -1377|       -764|        -85| 15206|         27
RTD|       -778|      -1377|       -760|        -53| 15206|         27
RTD|       -781|      -1377|       -764|       -149| 15206|         27
^CRTD|       -781|      -1377|       -764|       -149| 15206|         27

 >>> S = 98.696, EXECTIME = 0.0185549

root@AD502:/usr/realtime/testsuite/user/latency# cd ..
root@AD502:/usr/realtime/testsuite/user# ls
latency  preempt  switches
root@AD502:/usr/realtime/testsuite/user# cd preempt/
root@AD502:/usr/realtime/testsuite/user/preempt# ls
display  preempt  run
root@AD502:/usr/realtime/testsuite/user/preempt# ./run
*
*
* Type ^C to stop this application.
*
*
RTAI Testsuite - LXRT preempt (all data in nanoseconds)
RTH|     lat min|     lat avg|     lat max|    jit fast|    jit slow
RTD|       -1357|        -871|       10105|        7638|        8907
RTD|       -1357|        -827|       10105|        7962|        8907
RTD|       -1357|        -862|       14684|       15479|        8907
RTD|       -1357|        -859|       14684|       15479|        8907
RTD|       -1357|        -855|       14684|       15479|        8907
RTD|       -1357|        -861|       14684|       15479|        8907
RTD|       -1357|        -858|       14684|       15479|        8907
RTD|       -1357|        -862|       14684|       15479|        8907
^CRTD|       -1357|        -862|       14684|       15479| 8907

root@AD502:/usr/realtime/testsuite/user/preempt# cd ..
root@AD502:/usr/realtime/testsuite/user# cd switches/
root@AD502:/usr/realtime/testsuite/user/switches# ls
run  switches
root@AD502:/usr/realtime/testsuite/user/switches# ./run
*
*
* Type ^C to stop this application.
*
*


Wait for it ...


FOR 10 TASKS: TIME 19 (ms), SUSP/RES SWITCHES 20000, SWITCH TIME 953 (ns)

FOR 10 TASKS: TIME 19 (ms), SEM SIG/WAIT SWITCHES 20000, SWITCH TIME 989
(ns)

FOR 10 TASKS: TIME 23 (ms), RPC/RCV-RET SWITCHES 20000, SWITCH TIME 1175
(ns)

root@AD502:/usr/realtime/testsuite/user/switches#
root@AD502:/usr/realtime/testsuite/user/switches# dmesg
[ 1014.107240] I-pipe: head domain RTAI registered.
[ 1014.107247] RTAI[hal]: mounted. ISOL_CPUS_MASK: 0, LINUX CPU ISOLATED
MAP: 0).
[ 1014.107253] SYSINFO - # CPUs: 2, TIMER NAME: 'lapic', TIMER IRQ:
4353, TIMER FREQ: 68590998, CLOCK NAME: 'tsc', CLOCK FREQ: 2194923000,
CPU FREQ: 2194923000, LINUX TIMER IRQ: 4353.
[ 1041.325030] RTAI[malloc]: global heap size = 2097152 bytes, <BSD>.
[ 1041.325097] , kstacks pool size = 524288 bytes.
[ 1041.325102] RTAI[sched]: hard timer type/freq = lapic/68590998(Hz);
timing: oneshot; linear timed lists.
[ 1041.325105] RTAI[sched]: Linux timer freq = 250 (Hz), TimeBase freq =
2194923000 hz.
[ 1041.325107] RTAI[sched]: timer setup = 183 ns, resched latency = 3816 ns.
[ 1058.464561] SCHED releases registered named TASK THRSRV
[ 1058.464706] RTAI[malloc]: unloaded.
[ 1058.561068] RTAI[sched]: unloaded (forced hard/soft/hard transitions:
traps 0, syscalls 0).
[ 1064.623410] RTAI[malloc]: global heap size = 2097152 bytes, <BSD>.
[ 1064.623469] , kstacks pool size = 524288 bytes.
[ 1064.623472] RTAI[sched]: hard timer type/freq = lapic/68590998(Hz);
timing: oneshot; linear timed lists.
[ 1064.623475] RTAI[sched]: Linux timer freq = 250 (Hz), TimeBase freq =
2194923000 hz.
[ 1064.623477] RTAI[sched]: timer setup = 160 ns, resched latency = 3838 ns.
[ 1068.145817] LXRT releases PID 22475 (ID: display).
[ 1069.200090] SCHED releases registered named TASK THRSRV
[ 1069.200271] RTAI[malloc]: unloaded.
[ 1069.300292] RTAI[sched]: unloaded (forced hard/soft/hard transitions:
traps 0, syscalls 0).
[ 1073.287142] RTAI[malloc]: global heap size = 2097152 bytes, <BSD>.
[ 1073.287203] , kstacks pool size = 524288 bytes.
[ 1073.287206] RTAI[sched]: hard timer type/freq = lapic/68590998(Hz);
timing: oneshot; linear timed lists.
[ 1073.287209] RTAI[sched]: Linux timer freq = 250 (Hz), TimeBase freq =
2194923000 hz.
[ 1073.287212] RTAI[sched]: timer setup = 170 ns, resched latency = 3828 ns.
[ 1073.537716] SCHED releases registered named TASK THRSRV
[ 1073.537883] RTAI[malloc]: unloaded.
[ 1073.636819] RTAI[sched]: unloaded (forced hard/soft/hard transitions:
traps 0, syscalls 0).





On 2015-12-05 17:50, Paolo Mantegazza wrote:
> Thanks for the 32 bit test, but it must work also 64 bits. Nowadays I'm developing RTAI mostly using a 64 bit PC, checking it afterward on a 2 CPU SMP 32 bits PIII, gcc 4.3.
> Therefore there a, extended soft/hard compatibility time lapse should be still be maintained by RTAI, the only thing needed being a TSC.
>
> Paolo
> ________________________________________
> From: [email protected] [[email protected]]
> Sent: Saturday, December 5, 2015 5:31 PM
> To: Paolo Mantegazza; [email protected]
> Subject: Re: [Rtai]  CPU hangs on Celeron
>
> Hello!
>
> 1. Thanks for your help!
>
> 2. I use x86_64 kernel setting.
> Linux runs smooth so I think its good for the Celeron 1020E.
>
> I will try 32bit Celeron the next days and also try RTAI 5 if it is
> released.
>
> Thanks again!
>
> Best regards,
> Christoph
>
>
> On 2015-12-05 15:07, Paolo Mantegazza wrote:
>> Can you try with VULCANO CVS, or wait a few days till the release of RTAI-5.0-test1?
>> My be it is just a deception, but I'd prefer fixing any possible bug on the most recent stuff,
>> with a cleaner rtai_hal-ko unified for 32/64 bits and without any old legacy stuff.
>> Notice that you'll need an original 3.16.7 though.
>> Paolo
>> ________________________________________
>> From: Rtai [[email protected]] on behalf of [email protected] [[email protected]]
>> Sent: Saturday, December 5, 2015 12:03 PM
>> To: [email protected]
>> Subject: [Rtai]  CPU hangs on Celeron
>>
>> Hi!
>>
>> Here I like to add some output to my thread yesterday.
>>
>>
>> Here it stalls...
>> #/usr/realtime/testsuite/user/latency# ./run
>> *
>> *
>> * Type ^C to stop this application.
>> *
>> *
>>
>> ## RTAI latency calibration tool ##
>> # period = 100000 (ns)
>> # average time = 1 (s)
>> # use the FPU
>> # start the timer
>> # timer_mode is oneshot
>>
>> No output lines anymore.
>>
>> If I kill it then it gives me this.
>>
>>
>> rmmod: ERROR: Module rtai_msg is not currently loaded
>> rmmod: ERROR: Module rtai_mbx is not currently loaded
>> rmmod: ERROR: Module rtai_sem is not currently loaded
>> rmmod: ERROR: Module rtai_sched is not currently loaded
>>
>> rmmod: ERROR: Module rtai_hal is not currently loaded
>> RTAI Testsuite - USER latency (all data in nanoseconds)
>> 2015/12/5 12:02:54
>> RTH|    lat min|    ovl min|    lat avg|    lat max|    ovl max| overruns
>> RTD|         79|         79|         -1|    4195876| 4195876|          0
>>
>>
>>
>>
>>
>> Here the modules loaded after start.
>>
>> #lsmod
>> Module                  Size  Used by
>> rtai_msg               29319  0
>> rtai_mbx               11715  0
>> rtai_sem               28493  1 rtai_mbx
>> rtai_sched             97600  3 rtai_mbx,rtai_msg,rtai_sem
>> rtai_hal             3731229  4 rtai_mbx,rtai_msg,rtai_sem,rtai_sched
>>
>>
>>
>>
>> #dmesg
>>
>> [  225.975757] I-pipe: head domain RTAI registered.
>> [  225.975768] RTAI[hal]: compiled with gcc version 4.9.2 (Debian
>> 4.9.2-10) .
>> [  225.975772] ------------[ cut here ]------------
>> [  225.975780] WARNING: CPU: 0 PID: 13635 at kernel/ipipe/core.c:920
>> ipipe_set_hooks+0x216/0x250()
>> [  225.975783] Modules linked in: rtai_hal(O+)
>> [  225.975793] CPU: 0 PID: 13635 Comm: insmod Tainted: G O  3.16.1_ch_7 #11
>> [  225.975796] Hardware name: Neousys Technology Inc. Nuvo-4000
>> Series/NVS-3000, BIOS NV4KA003.Build150129 2014/12/22
>> [  225.975800]  0000000000000009 ffffffff8183a602 0000000000000000
>> ffffffff810cf8a7
>> [  225.975809]  0000000000000003 ffffffff82200b80 ffffffffa0002350
>> 0000000000000000
>> [  225.975817]  ffffffffa0003a10 ffffffff8115c986 ffffffff8115bf2d
>> ffffffff82200b80
>> [  225.975825] Call Trace:
>> [  225.975836]  [<ffffffff8183a602>] ? dump_stack+0x41/0x51
>> [  225.975843]  [<ffffffff810cf8a7>] ? warn_slowpath_common+0x77/0x90
>> [  225.975853]  [<ffffffffa0002350>] ? ack_bad_irq+0x40/0x40 [rtai_hal]
>> [  225.975859]  [<ffffffff8115c986>] ? ipipe_set_hooks+0x216/0x250
>> [  225.975868]  [<ffffffff8115bf2d>] ? __ipipe_do_sync_stage+0xfd/0x210
>> [  225.975873]  [<ffffffff8115df10>] ? ipipe_get_ptd+0x30/0x30
>> [  225.975881]  [<ffffffffa0002350>] ? ack_bad_irq+0x40/0x40 [rtai_hal]
>> [  225.975887]  [<ffffffff8115e0ad>] ? ipipe_catch_event+0xad/0x110
>> [  225.975894]  [<ffffffffa0002350>] ? ack_bad_irq+0x40/0x40 [rtai_hal]
>> [  225.975902]  [<ffffffffa00024e7>] ? __rtai_hal_init+0x197/0x340
>> [rtai_hal]
>> [  225.975907]  [<ffffffff811eb763>] ? kfree+0x163/0x490
>> [  225.975914]  [<ffffffff813b517e>] ? kasprintf+0x3e/0x40
>> [  225.975923]  [<ffffffffa0000a90>] ? rt_printk+0xb0/0xb0 [rtai_hal]
>> [  225.975930]  [<ffffffffa0002350>] ? ack_bad_irq+0x40/0x40 [rtai_hal]
>> [  225.975936]  [<ffffffff81000318>] ? do_one_initcall+0xd8/0x220
>> [  225.975941]  [<ffffffff811eb763>] ? kfree+0x163/0x490
>> [  225.975947]  [<ffffffff81144e06>] ? load_module+0x1eb6/0x2490
>> [  225.975954]  [<ffffffff81140e30>] ? store_uevent+0x40/0x40
>> [  225.975961]  [<ffffffff8114552d>] ? SyS_finit_module+0x7d/0xa0
>> [  225.975967]  [<ffffffff818428f0>] ? system_call_fastpath+0x1a/0x1f
>> [  225.975971] ---[ end trace 0386303a0229cbe7 ]---
>>
>> Thanks for help!
>>
>> Cheers,
>> Christoph
>>
>> _______________________________________________
>> Rtai mailing list
>> [email protected]
>> https://mail.rtai.org/cgi-bin/mailman/listinfo/rtai
>>
>

_______________________________________________
Rtai mailing list
[email protected]
https://mail.rtai.org/cgi-bin/mailman/listinfo/rtai