#asahi-dev · 2026-06-21 08:57
sofusyeah, im thinking its timing out for some reason, ffmpeg frees the buffers and then on the next frame it tries to decode those deallocated buffers#asahi-dev · 2026-06-21 08:52
sofusone irq per frame? or does it keep firing?#asahi-dev · 2026-06-21 08:49
sofushow does it explode?#asahi-dev · 2026-06-21 08:45
sofusI thought address handling was version dependent#asahi-dev · 2026-06-21 08:43
sofusthis is very helpfull thanks a lot!!#asahi-dev · 2026-06-21 08:43
sofusin hw.c t8112_configure_stream AVD_V4_VP_INSN_FIFO_IOVA_LO change addr & 0xffffffff into addr >> 8#asahi-dev · 2026-06-21 08:41
sofusnow its the instruction stream address thats wrong#asahi-dev · 2026-06-21 08:41
sofusNice! :)#asahi-dev · 2026-06-21 08:34
sofusand remove the dart fix i think its the wrong solution#asahi-dev · 2026-06-21 08:33
sofustry setting DMA_BIT_MASK to 38 in avd-drv.c#asahi-dev · 2026-06-21 08:32
sofusamazing#asahi-dev · 2026-06-21 08:24
sofusJamesCalligeros[m]: in avd-inst.h force it to take the else path in push_rvra and push_address#asahi-dev · 2026-06-21 08:03
sofusand i could not find a way to force dart to turn off#asahi-dev · 2026-06-21 08:02
sofusthe reset handler is not actually needed, i just found that the device does not power down unless dart also does#asahi-dev · 2026-06-21 08:01
sofusmaybe, the dart errors are very weird#asahi-dev · 2026-06-21 07:57
sofusthe version stuff is only based on the tunables#asahi-dev · 2026-06-21 07:54
sofusthanks, out of curiosity, what version does the driver log?#asahi-dev · 2026-06-21 07:51
sofuscan you send the logs?#asahi-dev · 2026-06-21 07:50
sofusbut ffmpeg should be fine for that video#asahi-dev · 2026-06-21 07:50
sofusi did not even know there was a ebuild#asahi-dev · 2026-06-21 07:42
sofusjust to rule out ffmpeg doing something weird#asahi-dev · 2026-06-21 07:41
sofuswould you mind testing https://www.itu.int/wftp3/av-arch/jvt-site/draft_conformance/AVCv1/AUD_MW_E.zip#asahi-dev · 2026-06-21 07:38
sofusthe t8112 should also be the exact same as mine, so im very confused#asahi-dev · 2026-06-21 07:36
sofusor not one of, vp number 4#asahi-dev · 2026-06-21 07:35
sofusthe error means one of the vp's had an error#asahi-dev · 2026-06-21 07:35
sofusi had that in the start but its not really usefull anymore. This is just a documentation problem on my part sorry.#asahi-dev · 2026-06-21 07:32
sofusi really need to add some debug prints...#asahi-dev · 2026-06-21 07:20
sofusthats very weird! What does PMD stand for?#asahi-dev · 2026-06-21 07:11
sofusi recommend gstreamer, ffmpeg has some issues with timestamps, some output will be wrong#asahi-dev · 2026-06-21 07:08
sofusits not really a v4l2m2m device, ffmpeg only supports it with patches. I was pretty supprised to see gentoo had those patches applied. Run with -hwaccel v4l2request#asahi-dev · 2026-06-21 07:06
sofuschadmed: what kinda iova faults where you getting? Would you mind sending logs?#asahi-dev · 2026-06-20 07:14
sofuschadmed: thanks, that was needed... I will test and merge later#asahi-dev · 2026-06-19 11:15
sofusi dont mind, thanks#asahi-dev · 2026-06-19 10:37
sofusany thoughts on my tunables approach? https://github.com/sofus13/avd-fw/commit/4f5e26ce9884d19bfd081e97db46b48659634406#asahi-dev · 2026-06-18 08:57
sofusthat would kinda make sense#asahi-dev · 2026-06-18 08:56
sofuswhat about BFR? There is always one DMA_BFR_CONFIG with the same name after an DMA_CONFIG#asahi-dev · 2026-06-18 08:51
sofusits one of the things they are inconsistent with#asahi-dev · 2026-06-18 08:50
sofusits actually all dma config, any clue what RD and WR could mean?#asahi-dev · 2026-06-18 08:44
sofusthanks, the names are super inconsistent, some tunatables equal, but one or two names are different#asahi-dev · 2026-06-18 08:35
sofusnice#asahi-dev · 2026-06-18 08:34
sofusfor the tunatable registers, R basicly copied names from the kc, not sure if i can do the same?#asahi-dev · 2026-06-18 08:31
sofusJamesCalligeros[m]: yes ofc#asahi-dev · 2026-06-18 08:31
sofusi have gotten the tunatables names for all devices now, some of the newer ones have 346 different registers#asahi-dev · 2026-06-18 08:28
sofusokay great, thanks#asahi-dev · 2026-06-18 08:26
sofuswould it be silly to let the cm3 handle the tunatables?#asahi-dev · 2026-06-17 09:13
sofusbased on tunatables i think im right#asahi-dev · 2026-06-17 09:12
sofusany clue what codenames corespond to what devices?#asahi-dev · 2026-06-17 08:41
sofuscan i do the same with tunatables?#asahi-dev · 2026-06-17 08:41
sofusok sure#asahi-dev · 2026-06-17 08:40
sofusif anyone with an m3 (or higher) would send me fw for the cm3 i could verify it