#asahi-dev · 2026-07-25 16:22
jannausofus: which decoder are you using on linux? JVT-AVC_V1 passes only 28/135 with GStreamer-H.264-V4L2SL on t6020 here#asahi-dev · 2026-07-25 16:19
jannauI suspect 'iommu-map*' in the pcie devicetree node#asahi-dev · 2026-07-25 16:18
jannauuse_dma_iommu() should return true. do you have CONFIG_IOMMU_DMA=y?#asahi-dev · 2026-07-25 16:17
jannauI don't think it should call dma_direct_supported()#asahi-dev · 2026-07-25 16:02
jannau16k page size, iommu / dart support?#asahi-dev · 2026-07-25 16:01
jannauthe wider address space shouldn't be an issue since the mac mini with m2 pro works with tg3 as well#asahi-dev · 2026-07-25 16:00
jannaunot sure why dma_set_coherent_mask() would fail. it should not with iommu. maybe print the error and follow the failure path up#asahi-dev · 2026-07-25 15:54
jannauI don't think it'll help, different issue than I assumed\#asahi-dev · 2026-07-25 15:51
jannausearch for dma_mask in tg3_init_one(), maybe try to 'apple,dma-range = <0x0 0x0 0x0 0xfc000000>;' to the ethernet pcie dart node#asahi-dev · 2026-07-25 12:44
jannausee drivers/hid/hid-haptic.c#asahi-dev · 2026-07-25 12:42
jannaulinux itself has now support for haptic trackpads#asahi-dev · 2026-07-25 11:38
jannau"Vp: 3 (00)" and "H3 00 error"#asahi-dev · 2026-07-25 11:36
jannauframe processing timeout's in kernel log#asahi-dev · 2026-07-25 11:26
jannauthere is a single test with ERROR: vp90-2-02-size-lf-1920x1080.webm maybe it's that one#asahi-dev · 2026-07-25 11:16
jannauJVT-AVC_V1 passes 28/135 on t8103#asahi-dev · 2026-07-25 11:14
jannausofus: are some fails / errors expected for vp9? t8103 passes 215/315 and none of the high bit depth ones#asahi-dev · 2026-07-25 11:07
jannauI think u-boot re-arms it as well#asahi-dev · 2026-07-25 11:07
jannauhuh, it looks like I'm mistaken and it was just what I was expecting (to trigger reboots if the next payload doesn't start)#asahi-dev · 2026-07-25 09:07
jannau`lsusb -t` displays the link speed#asahi-dev · 2026-07-24 18:18
jannauv3 M3 Pro/Max/Ultra intial devicetrees sent#asahi-dev · 2026-07-24 17:42
jannauI was planning to bring this up with M3* nvme support as the M3 Ultra requires macOS 15.x fw support#asahi-dev · 2026-07-24 17:41
jannauwhether it's a good idea to encode the fw version in the compatible itself or in a property is more of a style question#asahi-dev · 2026-07-24 17:37
jannauwe need a new compatible anyway because the the macOS 15 fw is very not compatible. it even breaks boot if it is not used#asahi-dev · 2026-07-24 08:07
jannaubut the difference is that the pmgr node doesn't list power-states for non-existing HW and misses one (or two) pmgr region completely#asahi-dev · 2026-07-24 08:02
jannauI suspect the pmgr* nodes for t60x0 weren't derived directly from the t60x0 ADT. The t60x1 ADT was used and we deleted the absent nodes manually#asahi-dev · 2026-07-23 12:16
jannauchaos_princess: the property exists for both but is None for M3. M3Pro has values#asahi-dev · 2026-07-20 14:10
jannaugentoo might have an use flag but I think firefox vendors ffmpeg#asahi-dev · 2026-07-20 14:08
jannauI don't think firefox has patched ffmpeg. so https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20847 still has to be released and merged in firefox ffmpeg copy#asahi-dev · 2026-07-17 08:43
jannauI was planning to do that anyway#asahi-dev · 2026-07-17 08:42
jannauchadmed: I can look after 14.8.3 dcp from middle of next week when I'm back from vacation#asahi-dev · 2026-07-15 20:54
jannauPPS_A_qualcomm_7 doesn't decode anything here. it produces "avd 269080000.avd: H0 00 error" and "269080000.avd: Frame processing timed out! Vp: 0 (00)" in dmesg#asahi-dev · 2026-07-15 20:50
jannauthose 3 samples are no correct with v4l2. remaining diff to your results: PICSIZE_A_Bossen_1, PICSIZE_B_Bossen_1, PPS_A_qualcomm_7, SAO_C_Samsung_5, SAO_D_Samsung_5, SDH_A_Orange_4, TUSIZE_A_Samsung_1#asahi-dev · 2026-07-15 20:36
jannauiirc it's less annoying in HEVC. a conformance sample for it looks like a soft reason to support it but probably not relevant in practice#asahi-dev · 2026-07-15 20:31
jannaualso missing in linux, so those 3 samples might ok there as well#asahi-dev · 2026-07-15 20:28
jannauCAINIT_H_SHARP_3, SLICES_A_Rovi_3 and SLIST_A_Sony_5 are my fault, samples were missing since I downloaded them manually. TSUNEQBD_A_MAIN10_Technicolor_2 appears to be ffmpeg issue due luma and chroma bit depth mismatch#asahi-dev · 2026-07-15 20:18
jannausofus: common errors between macos/linux on t8103 are CAINIT_H_SHARP_3, SLICES_A_Rovi_3, SLIST_A_Sony_5, TSUNEQBD_A_MAIN10_Technicolor_2#asahi-dev · 2026-07-15 17:29
jannausofus: macos 15.4 has 4 errors and 4 fails on t8103 for JCT-VC-HEVC_V1. I'll take a look which samples this affects later#asahi-dev · 2026-07-15 16:51
jannauI guess Fluendo-H.264-VT-Gst1.0 / Fluendo-H.265-VT-Gst1.0 and need install gstreamer#asahi-dev · 2026-07-15 16:46
jannausofus: how are testing vtb on macos?#asahi-dev · 2026-07-15 13:58
jannauI'll test later#asahi-dev · 2026-07-15 13:29
jannauhttps://paste.debian.net/hidden/8a5cab9b#asahi-dev · 2026-07-15 13:29
jannaudmesg once "avd 269080000.avd: H0 00 error" and 8 times "Frame processing timed out! Vp: 0 (00)"#asahi-dev · 2026-07-15 13:17
jannausofus: https://paste.centos.org/view/1b134151#asahi-dev · 2026-07-15 13:13
jannauRan 136/147 tests successfully, in 252.689 secs#asahi-dev · 2026-07-15 13:11
jannauI've ssen. thanks seems to work now. CAINIT_H_SHARP_3 errors but other samples so far work#asahi-dev · 2026-07-15 13:00
jannaudid you push a fix for the issue?#asahi-dev · 2026-07-15 12:51
jannauwould have been easier for me but nothing I can't deal with#asahi-dev · 2026-07-15 12:51
jannauno ptoblem#asahi-dev · 2026-07-15 12:50
jannauplease use a fixup commit as I merged the PR already locally#asahi-dev · 2026-07-15 12:44
jannauI'm not testing your latest update