[Kde-hardware-devel] Review Request 108407: systemd-login1 support for PowerDevil

[Dario Freddi]

> On Jan. 14, 2013, 12:40 p.m., Lamarque Vieira Souza wrote:
> > powerdevil/daemon/backends/upower/powerdevilupowerbackend.cpp, line 53
> > <http://git.reviewboard.kde.org/r/108407/diff/1/?file=107187#file107187line53>
> >
> >     Why not use QDBusPendingCallWatcher to track when the call above finishes and avoid this waitForFinished() call?
> 
> Lukáš Tinkl wrote:
>     What would be the advantage?
> 
> Lamarque Vieira Souza wrote:
>     Do not block the backend initialization while trying to launch logind. That can take time (several seconds) meanwhile anybody trying to access the backend will be blocked as well. I have not checked but I fear that can block the battery plasmoid and plasma-desktop as consequence.

I'm sorry I have to change my mind a second time here, I didn't take into account the fact that this happens while initializing, hence indeed this represents a potential problem even if the call is relatively fast. OTOH is true that if it was into a standard situation even a blocking solution would have been acceptable, but in this case we're potentially blocking quite a critical path.


- Dario


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/108407/#review25447
-----------------------------------------------------------


On Jan. 14, 2013, 2:24 p.m., Lukáš Tinkl wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/108407/
> -----------------------------------------------------------
> 
> (Updated Jan. 14, 2013, 2:24 p.m.)
> 
> 
> Review request for Solid and Dario Freddi.
> 
> 
> Description
> -------
> 
> This patch adds support for systemd-login1 service to Powerdevil's upower backend. The main purpose is that UPower will be soon dropping support[1] for suspend/resume features so we have to rely on systemd. With this login1, we are also gaining support for HybridSleep, where implemented by the system.
> 
> One caveat: the current login1 implementation doesn't support[2] emitting the "resume from suspend" signal
> 
> [1] http://lists.freedesktop.org/archives/devkit-devel/2013-January/001339.html
> [2] http://www.freedesktop.org/wiki/Software/systemd/inhibit
> 
> 
> This addresses bug https://bugzilla.redhat.com/show_bug.cgi?id=859227.
>     http://bugs.kde.org/show_bug.cgi?id=https://bugzilla.redhat.com/show_bug.cgi?id=859227
> 
> 
> Diffs
> -----
> 
>   powerdevil/daemon/BackendConfig.cmake 5dbe6f6 
>   powerdevil/daemon/backends/upower/login1suspendjob.h PRE-CREATION 
>   powerdevil/daemon/backends/upower/login1suspendjob.cpp PRE-CREATION 
>   powerdevil/daemon/backends/upower/powerdevilupowerbackend.h ba942bd 
>   powerdevil/daemon/backends/upower/powerdevilupowerbackend.cpp 97a409b 
>   powerdevil/daemon/backends/upower/upowersuspendjob.h bbe2f45 
>   powerdevil/daemon/backends/upower/upowersuspendjob.cpp fa64ab0 
> 
> Diff: http://git.reviewboard.kde.org/r/108407/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Lukáš Tinkl
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-hardware-devel/attachments/20130114/5697d76f/attachment.html>