All of lore.kernel.org
 help / color / mirror / Atom feed
From: Per Oberg <pero@wolfram.com>
To: Jan Kiszka <jan.kiszka@siemens.com>
Cc: xenomai <xenomai@xenomai.org>
Subject: Re: RTUdp incoming msg are ok but unable to send UDP
Date: Tue, 28 May 2019 06:01:53 -0500 (CDT)	[thread overview]
Message-ID: <1368542045.799892.1559041313687.JavaMail.zimbra@wolfram.com> (raw)
In-Reply-To: <81b1d78e-2835-7e53-3bd9-67e3e3f1daa8@siemens.com>


----- Den 28 maj 2019, på kl 8:45, Jan Kiszka jan.kiszka@siemens.com skrev:

> On 27.05.19 18:13, Per Oberg wrote:

> > ----- Den 27 maj 2019, på kl 10:22, Jan Kiszka jan.kiszka@siemens.com skrev:

> >> On 27.05.19 09:27, Per Oberg via Xenomai wrote:
> >>> Hi folks

> >>> I am having some troubles with UDP and rtnet. I think it used to work, but I'm
> >>> not sure in what configuration. I had to update to meet some driver demands and
> >>> now i am using the following versions:

> >>> Xenomai 3.0.7
> >>> ipipe-core-4.9.90-x86-6
> >>> Kernel 4.9.90


> >>> When trying to send a UDP message I got a kernel paging request error so I tried
> >>> adding nosmap to the boot parameters. This solved the kernel paging bug but no
> >>> messages are coming out.


> >>> Does anyone know if there went in some patches to rtudp lately that would fix
> >>> this?

> >>> Any other tips out there ?


> >> Can you try if the issue is also present in stable/v3.0.x or master?

>> At a first glance it looks like the bug is fixed. However, I couldn't make it
>> work in 3.0.8 because of compilation issues with the kernel. In current master
> > it works however.


> Can you share the compilation issues? Either they have been fixed in git, or
> they better should be before the next release.


I'm compiling using my own Yocto setup based on meta-xenomai from [1] below. For some reason I couldn't compile the kernel modules because of a missing  rtdm/net.h. I noticed that there was a split in the git tree while back so I guess that's the original culprit. 

However, I was never able to figure out if I had some old Xenomai-stuff laying around in the wrong places or if I forgot to update my sdk. Further, I was not able to get the actual compilation environment or see the real "cc commands". Instead of digging around further I decided to check master.

So, either Yocto kept the old kernel source or there is something else broken in 3.0.8 compilation. I haven't been able to tell which is more likely. When I created my own 3.0.8a release files based on master it started working again without cleaning anything else out. 

This is one example of the error:
------------------------------------------------------------------------------------------------------------------------
In file included from .../kernel-source/drivers/xenomai/net/stack/include/rtdev.h:37:0,
                 from .../kernel-source/drivers/xenomai/net/stack/include/rtnet_port.h:32,
                 from .../kernel-source/drivers/xenomai/net/drivers/8139too.c:54:
.../kernel-source/drivers/xenomai/net/stack/include/rtskb.h:32:10: fatal error: rtdm/net.h: No such file or directory
------------------------------------------------------------------------------------------------------------------------

I just realized that I probably hit this bug: [2/3]. I knew I saw it somewhere, but still couldn't fint it when I was looking yesterday. 

Does anyone have a good idea on how to effectively search the mailing list ? 


> >> Jan

> >> PS: 4.9 is EOL by now.

>> Yes, I've been meaning to upgrade to 4.14, but I got the feeling there was
> > something iffy going on for quite a while?


> What exactly do you mean?

Sorry for being unclear. With iffy I meant that I've been seeing some reports of issues people have had, and I was probably more wrong than right. It's kind of hard to figure out afterwards if issues you only glanced on when they passed the mailing list was ever fixed, only a glitch, or something serious. The bug-reporting system seems a bit under-used =)

Anyway, I was referring to these:
- [4] In my mind it was only yesterday it was released, but time flies. 
- [5] with a follow up that it was also for real hardware
- [6] perhaps only amd issue ?
- [7] kvm only issue ?



> Jan

[1] https://bitbucket.org/pficheux/meta-xenomai/src
[2] https://www.xenomai.org/pipermail/xenomai/2019-February/040364.html
[3] https://www.xenomai.org/pipermail/xenomai/2019-May/040866.html
[4] https://www.xenomai.org/pipermail/xenomai/2018-October/039653.html
[5] https://www.xenomai.org/pipermail/xenomai/2019-March/040460.html
[6] https://www.xenomai.org/pipermail/xenomai/2019-March/040619.html
[7] https://www.xenomai.org/pipermail/xenomai/2019-April/040722.html

Per Öberg 


      reply	other threads:[~2019-05-28 11:01 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-27  7:27 RTUdp incoming msg are ok but unable to send UDP Per Oberg
2019-05-27  8:22 ` Jan Kiszka
2019-05-27 16:13   ` Per Oberg
2019-05-28  6:45     ` Jan Kiszka
2019-05-28 11:01       ` Per Oberg [this message]

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=1368542045.799892.1559041313687.JavaMail.zimbra@wolfram.com \
    --to=pero@wolfram.com \
    --cc=jan.kiszka@siemens.com \
    --cc=xenomai@xenomai.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.