| Message ID | 20230116220909.196926-1-marijn.suijten@somainline.org (mailing list archive) |
|---|---|
| Headers | show |
| Series | iio: adc: qcom-spmi-vadc: Propagate fw node label to userspace | expand |
On Mon, 16 Jan 2023 23:09:04 +0100 Marijn Suijten <marijn.suijten@somainline.org> wrote: > Implement read_label in qcom-spmi-vadc to see DT-specified label names > in userspace. At the same time clear up some documentation around > extend_name to promote read_label, and normalize similar code in > qcom-spmi-adc5. > I think this is a good route forwards, but it is making changes to ABI so I definitely want input on this from at least one of the qualcomm maintainers before I pick it up - particularly the changes in patch 3. Don't want to cause anyone nasty surprises. J > Changes since v1: > - qcom-spmi-vadc: Use read_label instead of extend_name; > > New since v1: > - core: Point users of extend_name field to read_label callback > - qcom-spmi-adc5: Use datasheet_name string literal for > iio_chan_spec::datasheet_name; > - qcom-spmi-adc5: Fall back to datasheet_name instead of > fwnode_get_name() for iio_chan_spec::extend_name (gets rid of @xx in > sysfs filenames and labels); > - qcom-spmi-adc5: Remove unnecessary datasheet_name NULL check. > > v1: https://lore.kernel.org/linux-arm-msm/20221106193018.270106-1-marijn.suijten@somainline.org/ > > Marijn Suijten (5): > iio: core: Point users of extend_name field to read_label callback > iio: adc: qcom-spmi-adc5: Use driver datasheet_name instead of DT > label > iio: adc: qcom-spmi-adc5: Fall back to datasheet_name instead of > fwnode name > iio: adc: qcom-spmi-adc5: Remove unnecessary datasheet_name NULL check > iio: adc: qcom-spmi-vadc: Propagate fw node label to userspace > > drivers/iio/adc/qcom-spmi-adc5.c | 15 +++++++-------- > drivers/iio/adc/qcom-spmi-vadc.c | 19 ++++++++++++++++++- > include/linux/iio/iio.h | 3 +++ > 3 files changed, 28 insertions(+), 9 deletions(-) > > -- > 2.39.0 >
On 2023-01-22 16:59:47, Jonathan Cameron wrote: > On Mon, 16 Jan 2023 23:09:04 +0100 > Marijn Suijten <marijn.suijten@somainline.org> wrote: > > > Implement read_label in qcom-spmi-vadc to see DT-specified label names > > in userspace. At the same time clear up some documentation around > > extend_name to promote read_label, and normalize similar code in > > qcom-spmi-adc5. > > > > I think this is a good route forwards, but it is making changes > to ABI so I definitely want input on this from at least one of > the qualcomm maintainers before I pick it up - particularly > the changes in patch 3. > > Don't want to cause anyone nasty surprises. Ack, patch 3 is the odd one of the bunch (as discussed many times prior). It is an ABI break and would effectively obsolete "iio: adc: qcom-spmi-adc5: Fix the channel name" [1] as it then only affects a few dev_err/dev_dbg. Let's wait and hear from Qcom maintainers. [1]: https://lore.kernel.org/linux-arm-msm/20230118100623.42255-1-andriy.shevchenko@linux.intel.com/ - Marijn
On Mon, 23 Jan 2023 00:41:37 +0100 Marijn Suijten <marijn.suijten@somainline.org> wrote: > On 2023-01-22 16:59:47, Jonathan Cameron wrote: > > On Mon, 16 Jan 2023 23:09:04 +0100 > > Marijn Suijten <marijn.suijten@somainline.org> wrote: > > > > > Implement read_label in qcom-spmi-vadc to see DT-specified label names > > > in userspace. At the same time clear up some documentation around > > > extend_name to promote read_label, and normalize similar code in > > > qcom-spmi-adc5. > > > > > > > I think this is a good route forwards, but it is making changes > > to ABI so I definitely want input on this from at least one of > > the qualcomm maintainers before I pick it up - particularly > > the changes in patch 3. > > > > Don't want to cause anyone nasty surprises. > > Ack, patch 3 is the odd one of the bunch (as discussed many times > prior). It is an ABI break and would effectively obsolete "iio: adc: > qcom-spmi-adc5: Fix the channel name" [1] as it then only affects a few > dev_err/dev_dbg. Let's wait and hear from Qcom maintainers. > > [1]: https://lore.kernel.org/linux-arm-msm/20230118100623.42255-1-andriy.shevchenko@linux.intel.com/ Perhaps time for a resend of this series (drop the RFC marking as that tends to mean people don't read things if they are low on time). I'm going to mark this in patchwork as changes requested on basis a v3 will show up shortly. Thanks, Jonathan > > - Marijn
On 2023-05-01 17:21:51, Jonathan Cameron wrote: > On Mon, 23 Jan 2023 00:41:37 +0100 > Marijn Suijten <marijn.suijten@somainline.org> wrote: > > > On 2023-01-22 16:59:47, Jonathan Cameron wrote: > > > On Mon, 16 Jan 2023 23:09:04 +0100 > > > Marijn Suijten <marijn.suijten@somainline.org> wrote: > > > > > > > Implement read_label in qcom-spmi-vadc to see DT-specified label names > > > > in userspace. At the same time clear up some documentation around > > > > extend_name to promote read_label, and normalize similar code in > > > > qcom-spmi-adc5. > > > > > > > > > > I think this is a good route forwards, but it is making changes > > > to ABI so I definitely want input on this from at least one of > > > the qualcomm maintainers before I pick it up - particularly > > > the changes in patch 3. > > > > > > Don't want to cause anyone nasty surprises. > > > > Ack, patch 3 is the odd one of the bunch (as discussed many times > > prior). It is an ABI break and would effectively obsolete "iio: adc: > > qcom-spmi-adc5: Fix the channel name" [1] as it then only affects a few > > dev_err/dev_dbg. Let's wait and hear from Qcom maintainers. > > > > [1]: https://lore.kernel.org/linux-arm-msm/20230118100623.42255-1-andriy.shevchenko@linux.intel.com/ > > Perhaps time for a resend of this series (drop the RFC marking as that > tends to mean people don't read things if they are low on time). > > I'm going to mark this in patchwork as changes requested on basis a > v3 will show up shortly. v3 is out now after once again messing with b4 to import an older series, with a RESEND to include all the email addresses which b4 only addes after using prep --auto-to-cc, and there not being any obvious way to retroactively add the bunch of To:/Cc: commands to a previously sent version before running b4 send --resend v3. Fingers crossed it all went right and reaches the right people. - Marijn
Implement read_label in qcom-spmi-vadc to see DT-specified label names in userspace. At the same time clear up some documentation around extend_name to promote read_label, and normalize similar code in qcom-spmi-adc5. Changes since v1: - qcom-spmi-vadc: Use read_label instead of extend_name; New since v1: - core: Point users of extend_name field to read_label callback - qcom-spmi-adc5: Use datasheet_name string literal for iio_chan_spec::datasheet_name; - qcom-spmi-adc5: Fall back to datasheet_name instead of fwnode_get_name() for iio_chan_spec::extend_name (gets rid of @xx in sysfs filenames and labels); - qcom-spmi-adc5: Remove unnecessary datasheet_name NULL check. v1: https://lore.kernel.org/linux-arm-msm/20221106193018.270106-1-marijn.suijten@somainline.org/ Marijn Suijten (5): iio: core: Point users of extend_name field to read_label callback iio: adc: qcom-spmi-adc5: Use driver datasheet_name instead of DT label iio: adc: qcom-spmi-adc5: Fall back to datasheet_name instead of fwnode name iio: adc: qcom-spmi-adc5: Remove unnecessary datasheet_name NULL check iio: adc: qcom-spmi-vadc: Propagate fw node label to userspace drivers/iio/adc/qcom-spmi-adc5.c | 15 +++++++-------- drivers/iio/adc/qcom-spmi-vadc.c | 19 ++++++++++++++++++- include/linux/iio/iio.h | 3 +++ 3 files changed, 28 insertions(+), 9 deletions(-) -- 2.39.0