Hi Jussi,

Thanks for the help.

Here are the relevant steps of the problem:


1.      weston.service sets XDG_RUNTIME_DIR, then starts Weston

2.      Weston starts up and registers its Wayland server socket in 
XDG_RUNTIME_DIR

3.      pam starts and sets XDG_RUNTIME_DIR to another location

4.      Wayland client fails to find the Wayland server socket.

At this point, I realize I’m not exactly sure what will happen. Looking at the 
Weston man page, it appears possible that the Wayland client will still be able 
to connect through the default Wayland server socket. I can’t immediately test 
this to know for sure, but regardless, there is still an issue except for the 
default case.

Tom

From: Jussi Kukkonen [mailto:jussi.kukko...@intel.com]
Sent: Friday, September 09, 2016 1:41 AM
To: Tom Hochstein <tom.hochst...@nxp.com>
Cc: Patches and discussions about the oe-core layer 
(openembedded-core@lists.openembedded.org) 
<openembedded-core@lists.openembedded.org>
Subject: Re: [OE-core] weston startup problem

On 24 August 2016 at 20:08, Tom Hochstein 
<tom.hochst...@nxp.com<mailto:tom.hochst...@nxp.com>> wrote:
Hi All,

Hi, sorry for missing this in the originally,

The weston systemd startup implementation is not working correctly and I need 
some help fixing it.

With systemd and pam on the image, the expected behavior is that 
XDG_RUNTIME_DIR would already be set when weston is launched from 
weston.service. Turns out XDG_RUNTIME_DIR is not set, and weston.service sets 
it itself. This is already a problem, and the problem would be worse except 
that weston happens to use a different folder. If weston.service is modified to 
use the same folder name, the folder ends up getting replaced later by pam.


I can see our XDG_RUNTIME_DIR handling indeed should be cleaned up but just so 
I understand the whole issue: can you explain the problem you have in detail? 
It seems one issue is that the variable isn't set outside of  Xsession (it 
really should) but why is it a problem if weston.service decides a runtime 
directory for itself? How does pam interfere with that?


Thanks,
 Jussi


From looking at the weston-launch code and its pam handling, I thought to try 
the weston-launch option '--user=root' to force the creation of the pam 
session. This does seem to help, and I was able to get it mostly working by 
modifying the openvt args in weston-start, removing -e and adding -w:

openvt -v -s -w -- weston-launch --user=root -- --modules=xwayland.so 
--log=/var/log/weston.log

The problem I have now is that keyboard input goes to both weston and the 
desktop. I tried with and without the -s option, and there is no difference.

I'm stuck at this point and would appreciate some help.

Tom

-- 
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Reply via email to