On 12/22/21 2:37 AM, Bryan O'Donoghue wrote:
V3: Publishing this V3 to make it a bit easier to merge - Adds Rob Herring's RB for the first DTS change - Adds Robert Foss' RB for the final three patches - Validated on latest 5.16-rc6 on rb3 and rb5 - No functional changes to v2 Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/commit/?h=v5.16-rc6-sm8250-camss-imx577-only-v3 Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/commit/?h=v5.16-rc3-sm8250-camss-imx577-only-v2 RESEND V2: + Cc: devicetree@xxxxxxxxxxxxxxx on first patch V2: - Adds Tested-by Julian Grahsl as indicated. - Fixes low-level noise checkpatch --strict I ended up having to push defines into static inlines for this. - Adds Reviewed-by Robert Foss. - Adds VFE/CSID fix for sdm845 as explicit additional sdm845 specific patch per Robert's suggestion. I took the time then to test out on sdm845 to ensure the new patch breaks nothing "./cam -c 1 --capture=1 --file=capture.raw" continues to produce expected data. Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/commit/?h=v5.16-rc1-sm8250-camss-imx577-only Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/commit/?h=v5.16-rc3-sm8250-camss-imx577-only-v2 git diff v5.16-rc1-sm8250-camss-imx577-only..v5.16-rc3-sm8250-camss-imx577-only-v2 -- drivers/media/platform/qcom/camss V1/RESEND: RESEND: git send-email --to=inux-arm-msm@xxxxxxxxxxxxxxx -> --to=linux-arm-msm@xxxxxxxxxxxxxxx This series is imported from excellent enabling work contributed by Jonathan Marek which we have picked up in Linaro for the RB5 project. Link: https://patchwork.kernel.org/project/linux-arm-msm/patch/20210511180728.23781-2-jonathan@xxxxxxxx/ Link: https://www.qualcomm.com/products/robotics-rb5-platform Link: https://www.96boards.org/product/qualcomm-robotics-rb5 Link: https://www.thundercomm.com/app_en/product/1590131656070623?index=1&categoryId=category0&tabIndex=1 The RB5 has a slightly different sku I think to Jonathan's platform - QRB5165 in our case with a different version of LK the relevant bootloader for the board. As a result we needed to make some additional updates to the provided patcheset, included here. The most notable changes are: - Addition of clocks. - Addition of bandwidth settings. - Representing the dependency on a VFE for a given CSID so that V4L subdevs will come out of reset right when poked. A full tree which is bootable and capable of producing pixel data from the imx577 main sensor is available here, with this series being a sub-set of the necessary patches on the SoC side only. Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=v5.16-rc1-sm8250-camss-imx577-only media-ctl --reset media-ctl -v -d /dev/media0 -V '"imx412 '20-001a'":0[fmt:SRGGB10/4056x3040 field:none]' media-ctl -V '"msm_csiphy2":0[fmt:SRGGB10/4056x3040]' media-ctl -V '"msm_csid0":0[fmt:SRGGB10/4056x3040]' media-ctl -V '"msm_vfe0_rdi0":0[fmt:SRGGB10/4056x3040]' media-ctl -l '"msm_csiphy2":1->"msm_csid0":0[1]' media-ctl -l '"msm_csid0":1->"msm_vfe0_rdi0":0[1]' yavta -B capture-mplane -c -I -n 5 -f SRGGB10P -s 4056x3040 -F /dev/video0
Tested-by: Vladimir Zapolskiy <vladimir.zapolskiy@xxxxxxxxxx> I've tested the series on top of v5.16-rc6 on RB5 board, image capturing from the imx577 sensor works fine, if I use the supplied rb5_defconfig. However I have to report that the camss driver is not probed successfully, if it is built-in, then there is an issue related to device link suppliers: platform ac6a000.camss: error -EPROBE_DEFER: wait for supplier camera@1a platform ac6a000.camss: Added to deferred list This could be analyzed separately. -- Best wishes, Vladimir
