| Message ID | 20240619230400.8459-1-ansuelsmth@gmail.com |
|---|---|
| Headers | show |
| Series | misc: introduce STATUS LED activity function | expand |
On Thu, Jun 20, 2024 at 08:34:03AM +0200, Heinrich Schuchardt wrote: > On 6/20/24 01:03, Christian Marangi wrote: > > This series expand the STATUS LED framework with a new color > > and a big new feature. One thing that many device need is a way > > to communicate to the user that the device is actually doing > > something. > > > > This is especially useful for recovery steps where an > > user (for example) insert an USB drive, keep a button pressed > > and the device autorecover. > > > > There is currently no way to signal the user externally that > > the bootloader is processing/recoverying aside from setting > > a LED on. > > > > A solid LED on is not enough and won't actually signal any > > kind of progress. > > Solution is the good old blinking LED but uboot doesn't > > suggest (and support) interrupts and almost all the LED > > are usually GPIO LED that doesn't support HW blink. > > > > To fix this and handle the problem of device not supporting > > HW blink, expand the STATUS LED framework with new API. > > > > We introduce a new config LED_STATUS_ACTIVITY, that similar > > to the RED, GREEN and others, takes the LED ID set in > > the LED_STATUS config and is used as the global LED for activity > > operations. > > > > We add status_led_activity_start/stop() used to turn the > > activity LED on and register a cyclic to make it blink. > > LED will continue to blink until _stop() is called. > > > > Function that will start some kind of action will first call > > _start() and then at the end will call stop(). > > If (on the broken implementation) a _start() is called before > > calling a _stop(), the Cyclic is first unregister and is > > registered again. > > > > Support for this is initially added to the tftp, mtd and ubi > > command. > > > > This also contains a big fixup for the gpio_led driver that > > currently deviates from the Documentation and make the > > coloured status led feature unusable. > > > > (world tested with the azure pipeline) > > Hello Christian, > > What is the relationship between CONFIG_LED and CONFIG_LED_STATUS? > Can they be used at the same time or should they be exclusive? > Should CONFIG_LED_STATUS depend on CONFIG_LED? Well this is something that should have been fixed long time ago. Given the 2 cmd and some function clash with each other they are totally incompatible with each other and one should exclude the other. > > This should be clarified both in Kconfig and in the documentation. > > At least cmd/led.c and cmd/legacy_led.c are conflicting as both provide > a command 'led'. > > Can we unify the too? Unify them is complicated (I tried) as the Status LED goes very deep and can be supported by the board with specific implementation. LED driver is all based on DT and most of the time it's just GPIO. (aka require OF support) > > Best regards > > Heinrich > > > > > Changes v4: > > - Rebase on top of next > > - Rework for cyclic changes on next > > - Fix compilation error > > - Fix compilation warning due to missing header > > Changes v3: > > - Add log on Status LED error conditions > > Changes v2: > > - Add Documentation conversion > > - Rework to Cyclic and simplify implementation > > > > Christian Marangi (9): > > misc: gpio_led: fix broken coloured LED status functions > > led: status_led: add support for white LED colour > > led: status_led: add function to toggle a status LED > > led: status_led: add warning when an invalid Status LED ID is used > > led: status_led: add new activity LED config and functions > > tftp: implement support for LED status activity > > mtd: implement support for LED status activity > > ubi: implement support for LED status activity > > doc: convert README.LED to .rst documentation > > > > cmd/legacy_led.c | 6 + > > cmd/mtd.c | 19 ++++ > > cmd/ubi.c | 15 ++- > > common/board_f.c | 2 + > > doc/README.LED | 77 ------------- > > doc/api/index.rst | 1 + > > doc/api/status_led.rst | 35 ++++++ > > drivers/led/Kconfig | 30 +++++ > > drivers/misc/gpio_led.c | 41 +++++-- > > drivers/misc/status_led.c | 75 ++++++++++++- > > include/status_led.h | 231 +++++++++++++++++++++++++++++++++++++- > > net/tftp.c | 7 ++ > > 12 files changed, 440 insertions(+), 99 deletions(-) > > delete mode 100644 doc/README.LED > > create mode 100644 doc/api/status_led.rst > > >
On 6/20/24 01:03, Christian Marangi wrote: > This series expand the STATUS LED framework with a new color > and a big new feature. One thing that many device need is a way > to communicate to the user that the device is actually doing > something. > > This is especially useful for recovery steps where an > user (for example) insert an USB drive, keep a button pressed > and the device autorecover. > > There is currently no way to signal the user externally that > the bootloader is processing/recoverying aside from setting > a LED on. > > A solid LED on is not enough and won't actually signal any > kind of progress. > Solution is the good old blinking LED but uboot doesn't > suggest (and support) interrupts and almost all the LED > are usually GPIO LED that doesn't support HW blink. > > To fix this and handle the problem of device not supporting > HW blink, expand the STATUS LED framework with new API. > > We introduce a new config LED_STATUS_ACTIVITY, that similar > to the RED, GREEN and others, takes the LED ID set in > the LED_STATUS config and is used as the global LED for activity > operations. > > We add status_led_activity_start/stop() used to turn the > activity LED on and register a cyclic to make it blink. > LED will continue to blink until _stop() is called. > > Function that will start some kind of action will first call > _start() and then at the end will call stop(). > If (on the broken implementation) a _start() is called before > calling a _stop(), the Cyclic is first unregister and is > registered again. > > Support for this is initially added to the tftp, mtd and ubi > command. > > This also contains a big fixup for the gpio_led driver that > currently deviates from the Documentation and make the > coloured status led feature unusable. > > (world tested with the azure pipeline) Hello Christian, What is the relationship between CONFIG_LED and CONFIG_LED_STATUS? Can they be used at the same time or should they be exclusive? Should CONFIG_LED_STATUS depend on CONFIG_LED? This should be clarified both in Kconfig and in the documentation. At least cmd/led.c and cmd/legacy_led.c are conflicting as both provide a command 'led'. Can we unify the too? Best regards Heinrich > > Changes v4: > - Rebase on top of next > - Rework for cyclic changes on next > - Fix compilation error > - Fix compilation warning due to missing header > Changes v3: > - Add log on Status LED error conditions > Changes v2: > - Add Documentation conversion > - Rework to Cyclic and simplify implementation > > Christian Marangi (9): > misc: gpio_led: fix broken coloured LED status functions > led: status_led: add support for white LED colour > led: status_led: add function to toggle a status LED > led: status_led: add warning when an invalid Status LED ID is used > led: status_led: add new activity LED config and functions > tftp: implement support for LED status activity > mtd: implement support for LED status activity > ubi: implement support for LED status activity > doc: convert README.LED to .rst documentation > > cmd/legacy_led.c | 6 + > cmd/mtd.c | 19 ++++ > cmd/ubi.c | 15 ++- > common/board_f.c | 2 + > doc/README.LED | 77 ------------- > doc/api/index.rst | 1 + > doc/api/status_led.rst | 35 ++++++ > drivers/led/Kconfig | 30 +++++ > drivers/misc/gpio_led.c | 41 +++++-- > drivers/misc/status_led.c | 75 ++++++++++++- > include/status_led.h | 231 +++++++++++++++++++++++++++++++++++++- > net/tftp.c | 7 ++ > 12 files changed, 440 insertions(+), 99 deletions(-) > delete mode 100644 doc/README.LED > create mode 100644 doc/api/status_led.rst >
On Thu, Jun 20, 2024 at 06:29:23AM +0200, Christian Marangi wrote: > On Thu, Jun 20, 2024 at 08:34:03AM +0200, Heinrich Schuchardt wrote: > > On 6/20/24 01:03, Christian Marangi wrote: > > > This series expand the STATUS LED framework with a new color > > > and a big new feature. One thing that many device need is a way > > > to communicate to the user that the device is actually doing > > > something. > > > > > > This is especially useful for recovery steps where an > > > user (for example) insert an USB drive, keep a button pressed > > > and the device autorecover. > > > > > > There is currently no way to signal the user externally that > > > the bootloader is processing/recoverying aside from setting > > > a LED on. > > > > > > A solid LED on is not enough and won't actually signal any > > > kind of progress. > > > Solution is the good old blinking LED but uboot doesn't > > > suggest (and support) interrupts and almost all the LED > > > are usually GPIO LED that doesn't support HW blink. > > > > > > To fix this and handle the problem of device not supporting > > > HW blink, expand the STATUS LED framework with new API. > > > > > > We introduce a new config LED_STATUS_ACTIVITY, that similar > > > to the RED, GREEN and others, takes the LED ID set in > > > the LED_STATUS config and is used as the global LED for activity > > > operations. > > > > > > We add status_led_activity_start/stop() used to turn the > > > activity LED on and register a cyclic to make it blink. > > > LED will continue to blink until _stop() is called. > > > > > > Function that will start some kind of action will first call > > > _start() and then at the end will call stop(). > > > If (on the broken implementation) a _start() is called before > > > calling a _stop(), the Cyclic is first unregister and is > > > registered again. > > > > > > Support for this is initially added to the tftp, mtd and ubi > > > command. > > > > > > This also contains a big fixup for the gpio_led driver that > > > currently deviates from the Documentation and make the > > > coloured status led feature unusable. > > > > > > (world tested with the azure pipeline) > > > > Hello Christian, > > > > What is the relationship between CONFIG_LED and CONFIG_LED_STATUS? > > Can they be used at the same time or should they be exclusive? > > Should CONFIG_LED_STATUS depend on CONFIG_LED? > > Well this is something that should have been fixed long time ago. > Given the 2 cmd and some function clash with each other they are totally > incompatible with each other and one should exclude the other. Yes, sorry, I missed this earlier on, my fault. The CONFIG_LED_STATUS code is legacy and indeed shouldn't be enabled on anything new (and ideally, someone would care enough to transition the few platforms using the old framework over). So this functionality should be added to the drivers/led/ framework and then yes, your platforms defining the LED via device tree, with the same compatibles the Linux needs anyhow to be able to control it too.