#asahi-gpu · 2024-02-16 04:02
chergertas soon as i asan most my things, webkit stops working #asahi-gpu · 2024-02-16 04:01
chergertsighs #asahi-gpu · 2024-02-16 04:01
chergertalways work to do #asahi-gpu · 2024-02-16 04:01
chergerti probably need to do some surgery for the client side for GTK soon, because when I profile we waste so much time going through generic FFI marshalling too#asahi-gpu · 2024-02-16 03:57
chergerti think the valuable thing about something like io_uring in the context of a compositor is that you could make your normal "for every client write this little message" thing do a single context switch #asahi-gpu · 2024-02-16 03:53
chergerthonestly, unless you're over 1000 fds poll() is probably faster anyway #asahi-gpu · 2024-02-16 03:51
chergertneat stuff you could never make work 10 years ago w/ epoll #asahi-gpu · 2024-02-16 03:51
chergertbut i did learn something super cool in the process, you can use io_uring now + eventfd semaphores to do global work queues and only wake up a single worker#asahi-gpu · 2024-02-16 03:51
chergertbetter than having to maintain an entire futures/fiber library like i did#asahi-gpu · 2024-02-16 03:45
chergertno luck from compiling mesa w/ asan at least, time to add the rest #asahi-gpu · 2024-02-16 03:30
chergertthe nice thing about async/await in C is you can finally gdb across segmented stacks #asahi-gpu · 2024-02-16 03:11
chergertthe thing i love about doing GPU for toolkits is that even when its slow to valgrind, it's 100x faster than it was doing pixman in a trenchcoat for the past 20 years #asahi-gpu · 2024-02-16 03:11
chergertalready on it, just waiting heh #asahi-gpu · 2024-02-16 03:08
chergerti get a bunch of 8 >= 8, then a single 11 >= 11, then the crashing, 0 >= 11#asahi-gpu · 2024-02-16 03:05
chergertsure, one sec #asahi-gpu · 2024-02-16 02:35
chergertadded a gdb backtrace at least #asahi-gpu · 2024-02-16 02:33
chergertmaybe it's getting dropped from buffering + crash #asahi-gpu · 2024-02-16 02:32
chergertalyssa: is there a way to force apitrace to write synchronously?#asahi-gpu · 2024-02-16 01:32
chergertuploaded to #38 #asahi-gpu · 2024-02-16 01:28
chergertalyssa: sorry for the delay, will do so quick #asahi-gpu · 2024-02-16 00:18
chergertalyssa: https://gitlab.freedesktop.org/asahi/mesa/-/issues/38 hope that's the right place to report!#asahi-gpu · 2024-02-16 00:11
chergert(it prefers GLES iirc to ensure support for DMA-BUF importing)#asahi-gpu · 2024-02-16 00:10
chergertthankfully same error with GL too#asahi-gpu · 2024-02-16 00:10
chergertoh, i think i need to force GL instead of GLES #asahi-gpu · 2024-02-16 00:07
chergertsure, ill record one #asahi-gpu · 2024-02-16 00:07
chergertwhat i do notice is that it's when using webkit here, which uses a gl texture to pass data to GTK #asahi-gpu · 2024-02-16 00:06
chergertbut if you want an "easy" way to test, build GTK from source, and run epiphany with that. "show overview" (ctrl+shift+o) should expose it #asahi-gpu · 2024-02-16 00:05
chergertthe older renderer which i worked on in GTK works fine, it's just our new one (which admitadly pushes things much harder to do more dmabuf stuff)#asahi-gpu · 2024-02-16 00:04
chergertalyssa: it works on intel Arc and iGPU it looks like here #asahi-gpu · 2024-02-15 23:29
chergertis building mesa here so can definitely try different things to test stuff #asahi-gpu · 2024-02-15 23:29
chergertdoes that sound like a driver bug or an app/library bug?#asahi-gpu · 2024-02-15 23:29
chergert ../src/mesa/state_tracker/st_gen_mipmap.c:114: st_generate_mipmap: Assertion `pt->last_level >= lastLevel' failed.#asahi-gpu · 2024-02-15 23:29
chergerti notice this assertion when testing the new GTK renderer for GNOME 46 and the tab overview used in various apps