From: pavel@denx.de (Pavel Machek)
To: cip-dev@lists.cip-project.org
Subject: [cip-dev] RT testing
Date: Mon, 15 Jul 2019 10:41:30 +0200 [thread overview]
Message-ID: <20190715084130.GA27894@amd> (raw)
In-Reply-To: <TYAPR01MB2285F72DD7C51D8D1E2C38B3B7F10@TYAPR01MB2285.jpnprd01.prod.outlook.com>
Hi!
(List added to the cc).
> Here are the links I was pasting into the Zoom chat regarding RT testing.
Thanks for all the help!
I got -cip-rt kernel to work on EBV Socrates, and it seems to be
suitable machine for testing. Latencies are under 70 usec under normal
loads. They go up-to 150 usec in longer tests and under very heavy
loads... These results are also consistent with osdal data:
https://www.osadl.org/Latency-plot-of-system-in-rack-7-slot.qa-latencyplot-r7s5.0.html?shadow=1
On iwg20m, I was getting 299 usec on cyclictest alone (job/1726) and
some mysterious 2000usec+ latencies on 4.19.50-cip3-rt1 (job/1797). I verified
Thinkpad X60 is not suitable due to hwlat. Thinkpad T40p is somehow
more promising (nothing detected by hwlat), but latencies cyclictest
is often in 500 usec range, and I have seen 755 usec with
4.4.176-cip31-rt23.
So.. EBV Socrates seems like best board to test on. Fortunately we
have very similar board in the lab:
https://lava.ciplatform.org/scheduler/device_type/Altera-Terasic-Deo-Nano
.
I believe we should do the following steps:
1) Add -rt configuration for socfpga to cip-kernel-config repositories
(me)
2) Enable CONFIG_HIGH_RES_TIMERS=y in all configs (realtime or not) so
we have comparison (me)
3) Prepare rt and non-rt kernels for socfpga. (I can do it one time,
but it would be good to do it automatically as part of checkin tests).
4) Tune cyclictest (length of test, background load) so that it
reproduces cca 100 usec latency on socfpga. If we set threshold to
cca 200 usec, it should consistently fail on non-RT_FULL
configurations. (I'll need help).
Best regards,
Pavel
--
DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.cip-project.org/pipermail/cip-dev/attachments/20190715/9643f7e5/attachment.sig>
next parent reply other threads:[~2019-07-15 8:41 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <TYAPR01MB2285F72DD7C51D8D1E2C38B3B7F10@TYAPR01MB2285.jpnprd01.prod.outlook.com>
2019-07-15 8:41 ` Pavel Machek [this message]
2020-01-14 17:01 [cip-dev] RT Testing Chris Paterson
2020-01-16 6:54 ` kazuhiro3.hayashi at toshiba.co.jp
2020-01-16 9:24 ` Daniel Wagner
2020-01-16 10:42 ` kazuhiro3.hayashi at toshiba.co.jp
2020-01-16 9:57 ` Chris Paterson
2020-01-17 1:01 ` Punit Agrawal
2020-01-17 11:34 ` Chris Paterson
2020-01-20 1:54 ` Punit Agrawal
2020-01-17 1:41 ` kazuhiro3.hayashi at toshiba.co.jp
2020-01-17 11:39 ` Chris Paterson
2020-01-16 9:24 ` Daniel Wagner
2020-01-16 11:36 ` Chris Paterson
2020-01-16 13:25 ` Pavel Machek
2020-01-16 17:13 ` Daniel Wagner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190715084130.GA27894@amd \
--to=pavel@denx.de \
--cc=cip-dev@lists.cip-project.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox