[frameworks-kio] [Bug 480797] "PrefersNonDefaultGPU" setting can nonsensically end up preferring the less powerful integrated GPU in cases where the more powerful GPU is the default one
https://bugs.kde.org/show_bug.cgi?id=480797 Berk Elyesa Yıldırım changed: What|Removed |Added CC||bery...@sayafe.org --- Comment #8 from Berk Elyesa Yıldırım --- Can you check whether the order is also broken with switcherooctl as well, since https://invent.kde.org/frameworks/kio/-/merge_requests/1332 plasma should just be probing the same thing AFAIS, so it might not be a kio bug at all. To add up, the upstream might push that to the next upstream (kernel) https://gitlab.freedesktop.org/hadess/switcheroo-control/-/issues/19 as "wrong boot_vga", i.e. the linked https://bugzilla.kernel.org/show_bug.cgi?id=211189 ... It's a little messy, but can you check your HDMI ports like the reporter: "I put the HDMI output on the motherboard and now the cards are well identified." as well? -- You are receiving this mail because: You are watching all bug changes.
[frameworks-kio] [Bug 481538] New: Discrete GPU dialog sets the legacy X-KDE-RunOnDiscreteGpu desktop key as overlooked backwards compatibility instead of upstream PrefersNonDefaultGPU
https://bugs.kde.org/show_bug.cgi?id=481538 Bug ID: 481538 Summary: Discrete GPU dialog sets the legacy X-KDE-RunOnDiscreteGpu desktop key as overlooked backwards compatibility instead of upstream PrefersNonDefaultGPU Classification: Frameworks and Libraries Product: frameworks-kio Version: git master Platform: Compiled Sources OS: Linux Status: REPORTED Severity: minor Priority: NOR Component: Properties dialog Assignee: kio-bugs-n...@kde.org Reporter: bery...@sayafe.org CC: kdelibs-b...@kde.org Target Milestone: --- SUMMARY If a desktop file doesn't already have PrefersNonDefaultGPU, the dialog sets the legacy X-KDE-RunOnDiscreteGpu key due to an overlooked conditional order in the name of backwards compatibility. STEPS TO REPRODUCE 0. Use 2 GPUs/Hybrid laptop 1. Find an app that doesn't already have PrefersNonDefaultGPU key (almost all regular apps) 2. Mark "Use discrete GPU" in advanced launch settings 3. It writes X-KDE-RunOnDiscreteGpu instead 4. Find an app that does have PrefersNonDefaultGPU key (i.e. switcheroo-control-test-app) 5. (Un)mark "Use discrete GPU" in advanced settings 6. Correctly (un)sets the key OBSERVED RESULT (Un)sets PrefersNonDefaultGPU if the key already exists, but writes X-KDE-RunOnDiscreteGpu on new entries. KIO checks for PrefersNonDefaultGPU and falls back to X-KDE-RunOnDiscreteGpu but in the wrong conditional order Kservice part is correct (but backwards compatibility part can still be dropped) EXPECTED RESULT (Un)set X-KDE-RunOnDiscreteGpu if the key already exists, but write PrefersNonDefaultGPU for new entries. OR Only take PrefersNonDefaultGPU into account since the legacy one should be deprecated since 2020 and Plasma 6 can break backwards compatibility. SOFTWARE/OS VERSIONS KDE Plasma Version: Broken since bundled, Plasma 5.23 KDE Frameworks Version: Broken since added, KF 5.86. ADDITIONAL INFORMATION Seems to be added with https://invent.kde.org/frameworks/kio/-/commit/85c5815e10a3b448583817de98486d14ebcba3fc The current file is in https://invent.kde.org/frameworks/kio/-/blob/master/src/widgets/kpropertiesdialogbuiltin_p.cpp?ref_type=heads#L2829 (moved since the mentioned commit) Should either drop X-KDE-RunOnDiscreteGpu entirely, or swap the two conditionals (and backport to 5.27 as well?) Or drop in 5.27 as well since apps can't set X-KDE-RunOnDiscreteGpu as they should follow the upstream spec anyway. It's tiny, so can I take it as a good first fix/contribution if you can guide the decision? Thanks xP -- You are receiving this mail because: You are watching all bug changes.
[frameworks-kio] [Bug 481538] Discrete GPU dialog sets the legacy X-KDE-RunOnDiscreteGpu desktop key as overlooked backwards compatibility instead of upstream PrefersNonDefaultGPU
https://bugs.kde.org/show_bug.cgi?id=481538 Berk Elyesa Yıldırım changed: What|Removed |Added CC||bery...@sayafe.org -- You are receiving this mail because: You are watching all bug changes.
[frameworks-kio] [Bug 480797] "PrefersNonDefaultGPU" setting can nonsensically end up preferring the less powerful integrated GPU in cases where the more powerful GPU is the default one
https://bugs.kde.org/show_bug.cgi?id=480797 --- Comment #10 from Berk Elyesa Yıldırım --- So it's technically correct, and not caused by plasma as well To set intel as the default gpu, you should connect to the motherboard display out, but then your purpose of using your amd card for desktop rendering doesn't apply. Maybe there can be an "ignore non-default gpu" global toggle in plasma to support such use cases, though; which would stick with the default gpu, i.e. your amd one here. Adding that as && pre conditional for KIO::hasDiscreteGpu() bool might be the easiest solution? -- You are receiving this mail because: You are watching all bug changes.
[frameworks-kio] [Bug 480797] "PrefersNonDefaultGPU" setting can nonsensically end up preferring the less powerful integrated GPU in cases where the more powerful GPU is the default one
https://bugs.kde.org/show_bug.cgi?id=480797 --- Comment #13 from Berk Elyesa Yıldırım --- Don't forget hybrid laptops, which as far as I'm aware has more dev attention and those are intentionally using iGPU as default and dGPU is offloaded to save battery life, but the Lutris issue is right on the PC main reversal you're currently experiencing. -- You are receiving this mail because: You are watching all bug changes.
[frameworks-kio] [Bug 481538] Discrete GPU dialog sets the legacy X-KDE-RunOnDiscreteGpu desktop key as overlooked backwards compatibility instead of upstream PrefersNonDefaultGPU
https://bugs.kde.org/show_bug.cgi?id=481538 --- Comment #3 from Berk Elyesa Yıldırım --- Thanks, I checked 1556 and this is way above that to interfere anything, just desktop key checking hehe Just the write action condition can simply be reversed (instead of being dropped) for 5.27, but either way that needs approval and I don't know how to do stuff for LTS yet.. -- You are receiving this mail because: You are watching all bug changes.
[kwin] [Bug 459860] [Wayland] Random, seemingly unreproducible crash at KWin::Workspace::takeActivity
https://bugs.kde.org/show_bug.cgi?id=459860 --- Comment #13 from Berk Elyesa Yıldırım --- For the one marked as "duplicate" (still doesn't seem related to me, I couldn't reproduce OPs bug anyway), I can't reproduce it anymore :) 11 Ekim 2024 05:49:14 GMT+02:00, Bug Janitor Service yazdı: >https://bugs.kde.org/show_bug.cgi?id=459860 > >--- Comment #12 from Bug Janitor Service --- >🐛🧹 ⚠️ This bug has been in NEEDSINFO status with no change for at least 15 >days. >Please provide the requested information, then set the bug status to REPORTED. >If there is no change for at least 30 days, it will be automatically closed as >RESOLVED WORKSFORME. > >For more information about our bug triaging procedures, please read >https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging. > >Thank you for helping us make KDE software even better for everyone! > >-- >You are receiving this mail because: >You are on the CC list for the bug. -- You are receiving this mail because: You are watching all bug changes.
[plasmashell] [Bug 493679] New: Plasmashell crash on gpu reset
https://bugs.kde.org/show_bug.cgi?id=493679 Bug ID: 493679 Summary: Plasmashell crash on gpu reset Classification: Plasma Product: plasmashell Version: 6.1.5 Platform: Fedora RPMs OS: Linux Status: REPORTED Keywords: drkonqi Severity: crash Priority: NOR Component: generic-crash Assignee: plasma-b...@kde.org Reporter: bery...@sayafe.org Target Milestone: 1.0 Application: plasmashell (6.1.5) Qt Version: 6.7.2 Frameworks Version: 6.6.0 Operating System: Linux 6.11.0-63.fc41.x86_64 x86_64 Windowing System: Wayland Distribution: "Fedora Linux 41 (KDE Plasma Prerelease)" DrKonqi: 6.1.5 [CoredumpBackend] -- Information about the crash: Plasma shell crashes when triggering amdgpu_gpu_recover sometimes Though, tere's some issues on the gpu side itself so linking for reference https://gitlab.freedesktop.org/drm/amd/-/issues/3647 Resetting fixes the issue temporarily, catched a crash while triggering. Hope it's helpful The crash can be reproduced sometimes. -- Backtrace: Application: plasmashell (plasmashell), signal: Aborted [New LWP 147683] [New LWP 147688] [New LWP 147686] [New LWP 147687] [New LWP 147668] [New LWP 147669] [New LWP 147648] [New LWP 147699] [New LWP 147635] [New LWP 147746] [New LWP 147942] [New LWP 147700] [New LWP 147684] [New LWP 147703] [New LWP 147946] [New LWP 147953] [New LWP 147702] [New LWP 147752] [New LWP 147745] [New LWP 147928] [New LWP 147751] [New LWP 147956] [New LWP 147954] [New LWP 147698] [New LWP 147748] [New LWP 147940] [New LWP 147754] [New LWP 147958] [New LWP 147929] [New LWP 147955] [New LWP 148481] [New LWP 147701] [New LWP 147952] [New LWP 147689] [New LWP 148477] [New LWP 148009] [New LWP 148468] [New LWP 147966] [New LWP 148499] [New LWP 147957] [New LWP 148482] [New LWP 147941] [New LWP 148497] [New LWP 148502] [New LWP 148464] [New LWP 148463] [New LWP 148480] [New LWP 148030] [New LWP 148500] [New LWP 148462] [New LWP 148483] [New LWP 148478] [New LWP 148465] [New LWP 148501] [New LWP 147943] [New LWP 147747] [New LWP 148466] [New LWP 147944] [New LWP 148467] [New LWP 147945] [New LWP 148498] [New LWP 148479] [New LWP 148495] [New LWP 149216] [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib64/libthread_db.so.1". Core was generated by `/usr/bin/plasmashell --no-respawn'. Program terminated with signal SIGABRT, Aborted. #0 0x7fcdc8680724 in __pthread_kill_implementation () from /lib64/libc.so.6 [Current thread is 1 (Thread 0x7fcda54006c0 (LWP 147683))] Cannot QML trace cores :( [Current thread is 1 (Thread 0x7fcda54006c0 (LWP 147683))] Thread 64 (Thread 0x7fcd52a006c0 (LWP 149216)): #0 0x7fcdc867ae69 in __futex_abstimed_wait_common () at /lib64/libc.so.6 #1 0x7fcdc867dc22 in pthread_cond_timedwait@@GLIBC_2.3.2 () at /lib64/libc.so.6 #2 0x7fcdc8eb0e7d in QWaitCondition::wait(QMutex*, QDeadlineTimer) () at /lib64/libQt6Core.so.6 #3 0x7fcdc8eadc45 in QThreadPoolThread::run() () at /lib64/libQt6Core.so.6 #4 0x7fcdc8ea4626 in QThreadPrivate::start(void*) () at /lib64/libQt6Core.so.6 #5 0x7fcdc867e797 in start_thread () at /lib64/libc.so.6 #6 0x7fcdc870278c in __clone3 () at /lib64/libc.so.6 Thread 63 (Thread 0x7fccfd6006c0 (LWP 148495)): #0 0x7fcdc870055d in syscall () at /lib64/libc.so.6 #1 0x7fcd9d490bea in futex_wait () at /lib64/libgallium-24.2.3.so #2 0x7fcd9d49a5fe in do_futex_fence_wait () at /lib64/libgallium-24.2.3.so #3 0x7fcd9dd74308 in si_flush_from_st () at /lib64/libgallium-24.2.3.so #4 0x7fcd9dab2b71 in tc_flush () at /lib64/libgallium-24.2.3.so #5 0x7fcd9d52f4d5 in st_context_flush () at /lib64/libgallium-24.2.3.so #6 0x7fcd9d44eb79 in dri_flush () at /lib64/libgallium-24.2.3.so #7 0x7fcdb3ab25b8 in dri2_wl_swap_buffers_with_damage () at /lib64/libEGL_mesa.so.0 #8 0x7fcdb3aa685a in dri2_swap_buffers () at /lib64/libEGL_mesa.so.0 #9 0x7fcdb3a99de8 in eglSwapBuffers () at /lib64/libEGL_mesa.so.0 #10 0x7fcdb48372d3 in QtWaylandClient::QWaylandGLContext::swapBuffers(QPlatformSurface*) () at /lib64/libQt6WaylandEglClientHwIntegration.so.6 #11 0x7fcdc98b80ed in QRhiGles2::endFrame(QRhiSwapChain*, QFlags) () at /lib64/libQt6Gui.so.6 #12 0x7fcdc972c42e in QRhi::endFrame(QRhiSwapChain*, QFlags) () at /lib64/libQt6Gui.so.6 #13 0x7fcdcaaced89 in QSGRenderThread::syncAndRender() () at /lib64/libQt6Quick.so.6 #14 0x7fcdcaad0263 in QSGRenderThread::run() () at /lib64/libQt6Quick.so.6 #15 0x7fcdc8ea4626 in QThreadPrivate::start(void*) () at /lib64/libQt6Core.so.6 #16 0x7fcdc867e797 in start_thread () at /lib64/libc.so.6 #17 0x7fcdc870278c in __clone3 () at /lib64/libc.so.6 Thread 62 (Thread 0x7fcd08c006c0 (LWP 148479)): #0 0x7fcdc867ae69 in __futex_abstimed_wait_common () at /lib64/libc.so.6 #1 0x7fcdc867d8b9 in pthread_cond_wait@@GLIBC_2.3.2 () at /lib
