#asahi-gpu · 2022-11-27 17:53
jannaureproducible but much more than a few seconds. more like 2-3 minutes. it complains after the poweron about idac value out of range, which has to be the 0x0 fromm the poweroff#asahi-gpu · 2022-11-27 17:48
jannauwtf, it fixed itself after a quite several seconds without setting the brightness#asahi-gpu · 2022-11-27 17:25
jannaulina: still broken on the imac. I suspect we need to wait after the iomfb has finished it's callbacks after the setPowerState#asahi-gpu · 2022-11-27 16:35
jannauI can't remember seeing that on devices with external displays or on the j314#asahi-gpu · 2022-11-27 16:35
jannauI've also noticed that on the imac macos 12.4 shuts dcp downs completely (including a MMIO CPU_RUN=0) when it turns off the display#asahi-gpu · 2022-11-27 16:19
jannaulina: if you chain the the requests like that there no need to use out of band for dcp_set_power_state(). the true as second parameter#asahi-gpu · 2022-11-27 16:13
jannaulina: yes, I've come to the same conclusion, seems to be bug on t8103 and t8112. I've never seen seen that on t6001#asahi-gpu · 2022-11-27 15:10
jannauthere's /topic which says "GPU / 3D graphics stack black-box RE and development (NO binary reversing) | Keep things on topic"#asahi-gpu · 2022-11-27 14:14
jannauatomic_check gets a all zero mode on disable#asahi-gpu · 2022-11-27 14:14
jannauthe check for the fake mode has to happen in mode_fixup and not in atomic_check. at least when checking the mode against the modes known by dcp#asahi-gpu · 2022-11-27 14:12
jannaufixed by https://github.com/jannau/linux/commit/021c3a56e99cad7b59f12c68eb310897c05254e3#asahi-gpu · 2022-11-27 14:11
jannaulina: that's fixed inside there#asahi-gpu · 2022-11-27 14:11
jannausee here: https://github.com/jannau/linux/tree/asahi-6.1-rc6-5_dcp-fixes#asahi-gpu · 2022-11-27 14:10
jannaulina: it could be that's broken as it is currently merged#asahi-gpu · 2022-11-27 13:53
jannaujoske: there are two fixes needed. for x11 with gpu you need to place https://www.jannau.net/asahi/30-modeset.conf into /etc/X11/xorg.conf.d/#asahi-gpu · 2022-11-27 11:35
jannauah, yes, the drm code for DRM_IOCTL_CRTC_{GET,QUEUE}_SEQUENCE made me believe that there's no support for hw without vblank irq in drm#asahi-gpu · 2022-11-27 11:31
jannauwhat means vsync is broken? Does it run significantly faster/slower than the display refresh rate?#asahi-gpu · 2022-11-27 11:26
jannauthanks, I'll see if it's possible to remove more vblank hacks or switch back to the default drm_mode_config_helper_funcs.atomic_commit_tail#asahi-gpu · 2022-11-27 10:07
jannaudon't you have a blog post to write? I can look after it later today#asahi-gpu · 2022-11-27 09:59
jannauI think not calling drm_vblank_init() and setting drm_crtc_state.no_vblank to true during modeset and swap might do what we need#asahi-gpu · 2022-11-27 09:56
jannaudrivers/gpu/drm/xen seems to do something which could match what dcp needs#asahi-gpu · 2022-11-27 09:45
jannauadvise I got in #dri-devel last weekend was to synthesize vblank events but I'm not sure if the weekend is the best time to ask this#asahi-gpu · 2022-11-27 09:42
jannaunot sure there is other complex hw supporting tear free swaps without vblank interrupts#asahi-gpu · 2022-11-27 09:40
jannaualso I've never seen overlapping swap sequences in macos traces, it's swap_start, swap, swap_complete#asahi-gpu · 2022-11-27 09:38
jannauwe might need to to use timestamps for queued swap. without them they are "blocking" and dcp doesn't like it to start the next swap before the previous one completed#asahi-gpu · 2022-11-27 09:35
jannauI'd rather have the the workaround affect users of the legacy kms api than the atomic kms api#asahi-gpu · 2022-11-27 09:33
jannauif we don't use vblanks after page swaps we need something else to prevent two concurrent swaps#asahi-gpu · 2022-11-27 09:31
jannauto use generated vblank events we would need at least control them to disable/enabled during modesets#asahi-gpu · 2022-11-27 09:30
jannauthe default helper waits for vblank with some large timeout (iirc 300ms)#asahi-gpu · 2022-11-27 09:26
jannaualso I think we have to do the vblank events after a swap to get fb swaps to work properly#asahi-gpu · 2022-11-27 09:25
jannauI haven't found that code yet or how I tell drm that the driver doesn't support vblank irq#asahi-gpu · 2022-11-27 09:22
jannauyes, it's the vblank issue. it "works" when switching the brightness quickly between 2 values#asahi-gpu · 2022-11-26 11:59
jannauI've seen xonotic crashes before though on the same machine#asahi-gpu · 2022-11-26 11:58
jannauyuka: xonotic works for me on m1 ultra. lina's kernel from yesterday and mesa asahi/main#asahi-gpu · 2022-11-26 11:57
jannauoh, that looks indeed like broken state#asahi-gpu · 2022-11-26 11:47
jannauyuka: after the 2nd frame? probably missing debug information and frame pointers. it would require code analysis / memory access to unwind the stack under these conditions#asahi-gpu · 2022-11-25 23:20
jannaudocumentation suggest that drm_aperture_remove_framebuffers should be called first but that resulted sometimes in no /dev/dri/cardX at all#asahi-gpu · 2022-11-25 23:19
jannauor I need to investigate why drm_aperture_remove_framebuffers() and devm_drm_dev_alloc() race each other#asahi-gpu · 2022-11-25 23:12
jannauworks with https://www.jannau.net/asahi/30-modeset.conf#asahi-gpu · 2022-11-25 23:12
jannaux11 works, the modesetting driver just can not deal with card2 (or not card0) as kms driver#asahi-gpu · 2022-11-25 22:51
jannauerr, forget that. there is no 8 core gpu model of the m2 mbp 13". so at least too much apple hw to keep track of the configurations#asahi-gpu · 2022-11-25 22:22
jannau8 GPU core M2 works as well#asahi-gpu · 2022-11-24 06:59
jannaualyssa: works and is faster, thanks#asahi-gpu · 2022-11-20 09:19
jannauthis seems to be unrelated though#asahi-gpu · 2022-11-20 09:19
jannauI would in genral recommend to use wayland with dcp/appledrm since it's 1) much closer to how macOS uses it 2) better tested#asahi-gpu · 2022-11-19 22:46
jannau"req->swap.src_rect[l].x = req->swap.src_rect[l].y = 0;" works for me#asahi-gpu · 2022-11-19 22:37
jannaudcp uses drm_fb_dma_get_gem_addr() to set the surf address for dcp and we set the src rectangle also based on the offset#asahi-gpu · 2022-11-19 22:35
jannauit allocates a single fb and one display is offsetted#asahi-gpu · 2022-11-19 22:35
jannauX11 multi-display support#asahi-gpu · 2022-11-19 22:34
jannaus/can't/can/