[plasmashell] [Bug 395306] AMD GPUs changed screen wakeup behavior, kde assumes old behavior

2018-08-09 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=395306

Mariusz Mazur  changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 Resolution|--- |INVALID

--- Comment #2 from Mariusz Mazur  ---
There's a bug in the amdgpu displayport link training code that causes monitors
to disconnect when trying to wake them up. It's not an intentional change on
their part.

-- 
You are receiving this mail because:
You are watching all bug changes.

[plasmashell] [Bug 395306] New: AMD GPUs changed screen wakeup behavior, kde assumes old behavior

2018-06-13 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=395306

Bug ID: 395306
   Summary: AMD GPUs changed screen wakeup behavior, kde assumes
old behavior
   Product: plasmashell
   Version: 5.12.4
  Platform: Other
OS: Linux
Status: UNCONFIRMED
  Severity: normal
  Priority: NOR
 Component: Multi-screen support
  Assignee: aleix...@kde.org
  Reporter: mariusz.g.ma...@gmail.com
CC: plasma-b...@kde.org
  Target Milestone: 1.0

Full explanation of the issue here:
https://bugs.freedesktop.org/show_bug.cgi?id=106529
Short version: when using recent amd gpus with recent kernels, the behavior
when waking up monitors is to (I suppose) mark them as disconnected until they
fully wake up. This is a serious problem for multi-monitor setups, since if the
primary monitor wakes up more slowly then any of the secondaries, then plasma
will for a few hundred ms switch the primary desktop to one of the secondary
monitors, then switch back when the actual primary monitor wakes up.
This looks like crap and tends to misbehave regularly (windows change their
position, plasmoids occasionally get misplaced).

Since the kernel bug report got ignored for the past month, I must assume
that's just how amdgpu driver will behave going forward. Which means the only
place that can mitigate the issue is the DE.

Can KDE mitigate the issue (by, dunno, adding 500ms of delay before switching
primary displays, or something)?

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-05-09 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #15 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
My last comment contains a proposal to work around the issue pointed out by the
reviewer. And I end with: "(…)  I'm waiting to know how you'd like to proceed
on this patch and the other one.".

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-05-09 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #13 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
Upstream reviewer went silent two weeks ago without making a decision
(https://phabricator.kde.org/D12536), so anybody interested in having this
applet working can patch it themselves (no rebuild needed), though out of the
box it'll remain broken for the foreseeable future.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-04-26 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #9 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
Ok, found a spare hour to figure this out (courtesy of car mechanic having
scheduling problems). All patches submitted, including the one that fixes this
issue (D12536).

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-04-25 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #7 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
Fixes for this bug (and probably 391634) can be found here:
https://github.com/mmazur/kdeplasma-addons/tree/timerfixes

Anybody willing to get them merged has my blessing. I'll probably try to do it
eventually, but likely not anytime soon, since that phabricator thing does not
look overly inviting to someone trying to quickly drop a few lines of patches.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-04-24 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #6 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
Ok, so I've rewritten the timer logic and now the plasmoid is (much more)
precise independently of issues with Qt. And fixed a few other bugs while I was
at it.

So first order of business is getting that code merged. Dealing with the Qt
bug, well, isn't really my priority, considering that (a) I don't want to spend
the time on learning how to build qt things and (b) if it's not just the qt
timer itself, but it interacting with something else in the plasmoid, then it
could take a while to pinpoint a test case. Especially for someone who has no
idea how to write plasmoids.

If anybody in here has a qt dev env and the know-how to use it to do a quick
test, then the offending code I'd start with would look something like this:

Timer {
  id: t;
  interval: 1000;
  onTriggered: {
var ts = new Date().getTime();
console.log(ts);
  }
  repeat: true;
  running: true;
}

What should happen is a log output roughly every second. What instead happened
in my debugging of the plasmoid was two log outputs every second, a few
milliseconds apart.

Which nicely explains why the plasmoid on my ubuntu 17.10 worked fine, and then
got the double boost when I upgraded to 18.04, all the while the code of the
plasmoid itself has not changed one bit. Only possible conclusion: the problem
must be in the underlying libraries.

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-04-24 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

--- Comment #4 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
QtQuick's Timer became unreliable and is notifying the plasmoid that a second
has passed twice each time it passes. I'll rewrite the handling to make sure
that's no longer the issue.

(And while I'm at it, I'll fix the precision, cause the timer plasmoid really
isn't precise, which I wasn't ware of before.)

-- 
You are receiving this mail because:
You are watching all bug changes.

[kdeplasma-addons] [Bug 381173] Wrong countdown speed in timer plasmoid's window.

2018-04-24 Thread Mariusz Mazur
https://bugs.kde.org/show_bug.cgi?id=381173

Mariusz Mazur <mariusz.g.ma...@gmail.com> changed:

   What|Removed |Added

 CC||mariusz.g.ma...@gmail.com

--- Comment #3 from Mariusz Mazur <mariusz.g.ma...@gmail.com> ---
Two days left to official ubuntu 18.04 release and this bug is present in kde.

-- 
You are receiving this mail because:
You are watching all bug changes.