#asahi-dev · 2023-07-28 08:19
eilnisp/ave/jpeg would go under drivers/media/platform/apple?#asahi-dev · 2023-07-28 08:19
eilnin a hypothetical scenario we all dream of#asahi-dev · 2023-07-25 12:30
eiln:))#asahi-dev · 2023-07-25 11:36
eilnmarcan: looks way better than my intel webcam :P#asahi-dev · 2023-07-25 11:31
eilnalso the frames are slightly blue? luma looks good, just the colors are a bit raw(?)#asahi-dev · 2023-07-25 11:31
eilnstill dropping about 1 in 2 frames bc hypervisor + python + i haven't quite figured out interrupts#asahi-dev · 2023-07-25 11:31
eilnhad to slow the framerate down by 1/5 bc m1n1 can't keep up#asahi-dev · 2023-07-25 11:31
eilnalriiight it streams https://streamable.com/c2jevl#asahi-dev · 2023-07-23 11:35
eilnive yet to decode it, but im guessing it's luma & chroma, like you said#asahi-dev · 2023-07-23 11:35
eiln(158): ISPASC: BUFFER: child of FrontRGBLocalRawlevel = 0, diff: 2 us, delta: 28655375 us#asahi-dev · 2023-07-23 11:35
eilni set the receiver pool as (1280*720, 1280*360) and i got my first frame :D#asahi-dev · 2023-07-23 11:35
eilnmarcan: the issue was actually the buffer protocol. still related to colors though#asahi-dev · 2023-07-23 02:29
eilnmarcan: sigh, currently reading the wikipedia entry on color spaces#asahi-dev · 2023-07-23 02:28
eiln(472): ISPASC: FLOW ERR: ./h10isp/filters/Flow/H13/CVideoFlowH13.cpp, 5351: flow=0, ch=0 can't get output yuv buf, dropped frame=0 (1) 0#asahi-dev · 2023-07-23 02:28
eiln(471): ISPASC: FLOW: CmdStartMSTF, 867: ch=0,numOfScales=5, bLumaOnly=0, bComp=1, bCTFCircularEn=1#asahi-dev · 2023-07-20 23:41
eilnif it's noticeably slower, i'll try to make up for it in userspace (if i get to that point..)#asahi-dev · 2023-07-20 23:40
eilncam output is the same with or without ane. i'm pretty sure it just speeds up face detection.#asahi-dev · 2023-07-20 23:28
eilnmarcan: also the ane-isp ipc can be disabled. i trace with the ane firmware range zeroed out.#asahi-dev · 2023-07-20 23:27
eilni think we're like a quarter there#asahi-dev · 2023-07-20 23:27
eilnMSG: UpdateDecimationConfig, 3697: ch=0,Config=0,standFrameRate=30.00,bNoChSynced=1,actChCnt=0,maxFRate=29.97,totFRate=0#asahi-dev · 2023-07-20 23:26
eilnUpdateEITBoundaries: cfg 0 sensorClk 38400000 vFrame W+H 1428, 896 max H 1792#asahi-dev · 2023-07-18 09:41
eilnmikelee: well, you're gonna need two for m1n1/hypervisor. if you know exactly what you're writing, you *can* develop drivers on asahi itself with modprobe#asahi-dev · 2023-07-18 09:36
eilnfirmware & i are now friends!#asahi-dev · 2023-07-18 09:36
eilnCH = 0x1, CMD = 0x0004 [CISP_CMD_PRINT_ENABLE], receive at 422127.906 ms, delta: 0 ms.#asahi-dev · 2023-07-18 09:35
eiln[ISP] CMD = 0x0004 [CISP_CMD_PRINT_ENABLE]#asahi-dev · 2023-07-18 09:35
eiln[ISP] CTRL READY#asahi-dev · 2023-07-16 20:10
eilnChaosPrincess: the sensors should be next. im really not familiar with the color/calibration magic though#asahi-dev · 2023-07-16 20:07
eilnsven: thank you, dcp/ipc.py was really helpful :)#asahi-dev · 2023-07-16 20:03
eilnand the code: https://pastebin.com/YpaH18Ne#asahi-dev · 2023-07-16 20:02
eilnlog: https://pastebin.com/ExWV7WNh#asahi-dev · 2023-07-16 20:02
eilnwtf pastebin is down#asahi-dev · 2023-07-16 20:01
eiln5am again, but i got the isp firmware to do its init dance#asahi-dev · 2023-07-16 14:29
eiln*powermetrics#asahi-dev · 2023-07-16 14:29
eilnthe +0x8000s aren't necessary, they're something to do with voltages in powertop#asahi-dev · 2023-07-16 14:27
eilnChaosPrincess: without the reset it doesn't run on the second run (and requires reboot). i need to look into it though#asahi-dev · 2023-07-16 14:22
eilnim working on the irq & command arg structs#asahi-dev · 2023-07-16 14:22
eilnthe next step is enabling fw print through the "ioctl"/cmd#asahi-dev · 2023-07-16 14:22
eilnthis is a much better script with the bootargs parsed#asahi-dev · 2023-07-16 14:21
eilnChaosPrincess: https://pastebin.com/RyVWR9Mt#asahi-dev · 2023-07-15 16:06
eilnaha, ty#asahi-dev · 2023-07-15 16:03
eilnChaosPrincess: i recall the ane dapf writes were slightly different so i played safe. this is much better, thank you#asahi-dev · 2023-07-15 15:53
eilnoh well, i've gotta figure out CISP_CMD_START first#asahi-dev · 2023-07-15 15:53
eilni'm worried calls like CISP_CMD_CH_AE_BIAS_EXPOSURE_SET aren't one-time, static params though (unlike, say, CISP_CMD_CH_COLOR_CAL_DATA_SET)#asahi-dev · 2023-07-15 15:52
eilnmarcan: makes sense. and if i can figure out one "ioctl", the 400 others should be relatively straightforward#asahi-dev · 2023-07-14 23:58
eilnim headed to work, ill reply tonight#asahi-dev · 2023-07-14 23:42
eilnyeah i should document this#asahi-dev · 2023-07-14 23:41
eilnand offset 0x0000 isn't cpu control & shouldn't be written to. the first +0x4000 are all fabrics under w_tun_a(). asc control is +0x1400000#asahi-dev · 2023-07-14 23:37
eilnalso pmu_reset is handled differently for the first three (so last three in the sequence) in power_off#asahi-dev · 2023-07-14 23:36
eilnin w_tun_d() the ttbr of the first dart domain gets copied to the 2nd & 3rd domains. this is unique to the isp & ane. dma doesn't work w/o it#asahi-dev · 2023-07-14 23:33
eilnbut it does boot & responds to irq