https://bugs.kde.org/show_bug.cgi?id=517625
--- Comment #2 from vindicator <[email protected]> --- https://discuss.kde.org/t/high-memory-usage-from-plasmashell-is-it-a-memory-leak/29105/ https://bugs.kde.org/show_bug.cgi?id=520190 Nate Graham 2026-08-12 19:39:16 UTC: ``` "There's a memory leak somewhere" unfortunately is not actionable for developers. It's a bit like saying, "My car doesn't drive as well as it used to". You'll have to pinpoint where the issue is so we can debug that thing. ``` IOW, mechanic to car owner: "You need to know your car inside and out and point to US, the mechanics, where the problem is, rather than just say it isn't getting as good mileage as similar cars." You know it's plasmashell (restarting frees up lots of shmem). You know it's related to graphics (probably GEM involving haswell (and related) hardware). You know it's related to the window previews (disabling compositor (no window previews) corrects the issue). *WE*'ve all pointed to what we believe is the impetus of the issue (window preview). Heaps/grinds have all been submitted at probably every link I've provided. I'll state here (stated elsewhere) that my system is an ancient Haswell, which has been known for having GEM issues. I had another thread open at freedesktop (and another site) regarding an app that downright destroys my session altogether. Like the rendering server (or driver?) never recovers, even when I manage to get into an alt tty and try as I might to restart any graphical component, and am forced to reboot. Intel even believed the issue is with their driver, but near as I can tell (and somewhat understandable), it's going to go nowhere since it is so ancient, and I'll be waiting for Nova Lake (or rather whenever the AI bubble pops and RAM comes back down). Until then, this issue will remain in-fact an issue. Whether or not anything gets done about it... -- You are receiving this mail because: You are watching all bug changes.
