[kwin] [Bug 524279] Night Light temperature transitions cause transient scanout corruption

https://bugs.kde.org/show_bug.cgi?id=524279

Sean Brar <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REPORTED                    |RESOLVED
         Resolution|---                         |WORKSFORME

--- Comment #1 from Sean Brar <[email protected]> ---
I recently updated my system to the Fedora 44 branch and I'm happy to report
this issue is completely resolved.

**New Environment:** KWin/Plasma 6.7.4 (Wayland), Mesa 26.1.6, kernel 7.1.8,
Fedora 44. Same hardware and 3-monitor layout.

Running the same DBus reproducer script now produces a perfectly smooth
transition with zero visible corruption.

The CRC measurements back this up. On 6.6.4, a single temperature step was
generating 10–19 distinct CRCs (way more than the expected 6–7 frames for the
color fade), and it was spitting out anomalous frames even when the temperature
was supposedly stable. On 6.7.4, the same sweep produces exactly 5–9 distinct
CRCs, which perfectly matches the expected fade interpolation. The visual and
mathematical anomalies are completely gone.

Since KWin, Mesa, and the kernel all updated at once, I'm not sure exactly
whose commit fixed it. (I did see similar symptoms under wlroots on a recent
Mesa build a week ago, so I suspect the fix wasn't solely in Mesa).

If any of the devs know what underlying plumbing changed to fix this between
6.6.4 and 6.7.4, I'd love to hear the post-mortem! Otherwise, I'll go ahead and
close this out.

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