On Mon, Jun 24, 2024 at 6:13 AM Bryan O'Donoghue <bryan.odonoghue@xxxxxxxxxx> wrote: > > On 23/06/2024 22:37, george chan wrote: > > User-space tool can't tell so I made some guesses. Sorry for misleading, actually i mean user-space too can't tell the difference. As all 3 kinds of init sequences are working, I can't get a strong conclusion of "correct" init sequence between atoll's and trodger's. > So how are you testing ? > > Libcamera on your target rootfs ? Yes, a similar test was carried out early days with the "wrong" v1.2.1 init sequence, on pmOS qcam installed into xiaomi redmi note 9 pro (sm7125). It showed nice output. And I was excited so I took a video recording too: https://www.youtube.com/watch?v=U_do11pSf1s After your indication, I replaced the v1.2.1 init sequence with atoll's and trodger's and carried some simple test with below cmd and both are outputting files. media-ctl --reset media-ctl -V '"msm_csid0":0[fmt:SRGGB10/2592x1944 field:none]' media-ctl -V '"msm_vfe0_rdi0":0[fmt:SRGGB10/2592x1944 field:none]' media-ctl -l '"msm_csid0":1->"msm_vfe0_rdi0":0[1]' v4l2-ctl -d /dev/v4l-subdev4 -c test_pattern=0 v4l2-ctl -d /dev/v4l-subdev5 -c test_pattern=0 v4l2-ctl -d /dev/v4l-subdev6 -c test_pattern=0 v4l2-ctl -d /dev/v4l-subdev19 -c test_pattern=$1 media-ctl -V '"s5k5e9 13-002d":0[fmt:SRGGB10/2592x1944 field:none]' media-ctl -V '"msm_csiphy2":0[fmt:SRGGB10/2592x1944 field:none]' media-ctl -l '"msm_csiphy2":1->"msm_csid0":0[1]' yavta -B capture-mplane --capture=3 -n 3 -f SRGGB10P -s 2592x1944 /dev/video0 -F As you can see the cmos named s5k5e9. and this time simply do yavta dump, no pmOS qcam test. Since this test is carried out in sm7125 SOC, in theory, it is better to test with sc7180 (less likely form-factor available in the market) so I will send out v2 with trogdor init sequence for other dev have sc7180 board to have a test. Stay tuned.