On 11/06/2019 20:05, Bjorn Andersson wrote: > Also, what's up with tsif0 vs tsif1? Some people count from 0; other people count from 1. The HW programming guide mentions TSIF_0 and TSIF_1. The pinctrl driver defines tsif1 and tsif2. I propose we use 0 and 1 consistently everywhere. > PS. I would suggest that you send a patch to the MSM8998 pinctrl driver > (and binding) where you squash tsifN_* to tsifN. It would break > backwards compatibility, but I think we can take that risk now before > someone starts to use it... And after that you can go with your proposed > squashed node. Here's what I have right now: $ git ls --stat 85c02fb4dfd1..HEAD 5ed38c44a92a (HEAD -> tsif-fixup) Fixup qcom,msm8998-pinctrl.txt example Documentation/devicetree/bindings/pinctrl/qcom,msm8998-pinctrl.txt | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) 823402af81a6 Fixup qcom,msm8998-pinctrl.txt binding Documentation/devicetree/bindings/pinctrl/qcom,msm8998-pinctrl.txt | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) afb686b8b3e7 Squash all 5 tsif1 pins together drivers/pinctrl/qcom/pinctrl-msm8998.c | 38 +++++++++----------------------------- 1 file changed, 9 insertions(+), 29 deletions(-) 8e4d31c8d455 Squash all 5 tsif0 pins together drivers/pinctrl/qcom/pinctrl-msm8998.c | 38 +++++++++----------------------------- 1 file changed, 9 insertions(+), 29 deletions(-) ee850fa510a6 sed -i 's/tsif2/tsif1/g' drivers/pinctrl/qcom/pinctrl-msm8998.c drivers/pinctrl/qcom/pinctrl-msm8998.c | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-) c7ffe4075623 sed -i 's/tsif1/tsif0/g' drivers/pinctrl/qcom/pinctrl-msm8998.c drivers/pinctrl/qcom/pinctrl-msm8998.c | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-) I'm wondering if the series needs to be split up (as it is) or squashed into a single patch (might be harder to review for mistakes). What do you think? Regards.
