You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mynewt.apache.org by "David G. Simmons" <sa...@mac.com> on 2016/12/21 19:24:40 UTC
Using the HAL/BSP
As you may or may not know, I've written a demo using an external ADC Sensor on the NRF52dk board. Since the Nordic stuff had to be moved out to an external repository, I nee do move the code that I had originally added to the -core repository out as well to comply with licensing, etc. It's also been an exercise in learning how to use the HAL/BSP stuff so that I can write a tutorial on that too.
I've been conversing with another mynewt developer off-list, but I think it's time to bring it out in public and see if I can get going again.
If anyone can shed light on any of this I'd be very grateful!
dg
>
>> On Dec 16, 2016, at 8:01 AM, David G. Simmons <santafen@mac.com <ma...@mac.com>> wrote:
>>
>> Ok, I'm finally back to this ...
>>
>> I think the majority of my confusion here is that there are HAL and BSP directories in various places, and I can't quite figure out where to jump in.
>>
> Yes, it can get confusing for sure :-)
>
>> Can you comment on whether or not I'm moving in the right direction here?
>>
>> there's hw/bsp/nrf52dk/include/bsp -- I'm guessing I'd have to define the ADC Pin here in bsp.h?
>>
>> I've already added
>> #if MYNEWT_VAL(ADC_0)
>> rc = os_dev_create((struct os_dev *) &os_bsp_adc0, "adc0",
>> OS_DEV_INIT_KERNEL, OS_DEV_INIT_PRIO_DEFAULT,
>> nrf52_adc_dev_init, &os_bsp_adc0_config);
>> assert(rc == 0);
>> #endif
>> }
>> to hal_bsp.c and I believe that stays there.
>>
> I think the way things are moving with this is that pin definitions are configured in structures in hal_bsp.c. There was a time that pins were defined in syscfg.yml but it seems the community has decided that pins should be set in structures defined in hal_bsp.c. For example, look at hal_bsp.c for the nrf51dk and how SPI pins are defined.
>
>> I believe that the hw/bsp/nrf52dk/syscfg.yml addition of
>>
>> ADC_0:
>> description: 'TBD'
>> value: 1
>> ADC_0_RESOLUTION:
>> description: 'TBD'
>> value: 'SAADC_CONFIG_RESOLUTION'
>> ADC_0_OVERSAMPLE:
>> description: 'TBD'
>> value: 'SAADC_CONFIG_OVERSAMPLE'
>> ADC_0_INTERRUPT_PRIORITY:
>> description: 'TBD'
>> value: 'SAADC_CONFIG_IRQ_PRIORITY'
>>
>> is also correct. But I might have to move this to my new LLS Driver syscfg.yml since it references SAADC_CONFIG_IRQ_PRIORITY which is in the Nordic licensed code.
> The BSP code that you are modifying with this would exist in the nordic directory right? This is one of the items that needs to be addressed in the code. There is a Jira ticket that I think is related if not directly about this subject; I cannot quite recall even though I added the Jira ticket! :-) The only piece of this that should be defined in syscfg.yml in the BSP is ADC_0. The rest of these items would be related to pin definitions and are either set in a structure in the bsp or set through some driver API. For example, I think resolution is a generic driver configuration. I think the general idea was this: there are chip specific configurations that get set in hal_bsp.c. There are “generic driver” configuration that get set through the ADC driver API. Of course, the generic configuration eventually need to call into a bsp or SDK call.
>
> This is the “tricky” part of the code changes and I would need to spend a bit of time on this myself to reconcile all this.
>
> Sorry I am not being the most helpful. If this still does not make sense I can spend a bit more time on it. The problem is, I am not 100% sure about all this.
>
>>
>> I believe I'll also have to add an adc_nrf52dk.h to hw/drivers/adc/adc_nrf52
>>
> This I am not sure of…
>
>> At present there is no hal_adc.h in hw/hal/include/hal where there is a hal_spi.h, etc. so I'm going to assume I need to add one, and this would just be basic stubs for the ADC calls.
>>
>> That's where I lose the plot ...
>>
> There is no hal for ADC and there is no expectation of doing one. I guess one could be added, but I think there was a discussion a ways back regarding the ADC and that coming up with a generic HAL for them, as they can be quite different, was not something we wanted to tackle. Thus, I would not add an ADC. This could be a good topic for the dev list too.
>
--
David G. Simmons
(919) 534-5099
Web <https://davidgs.com/> • Blog <https://davidgs.com/davidgs_blog> • Linkedin <http://linkedin.com/in/davidgsimmons> • Twitter <http://twitter.com/TechEvangelist1> • GitHub <http://github.com/davidgs>
/** Message digitally signed for security and authenticity.
* If you cannot read the PGP.sig attachment, please go to
* http://www.gnupg.com/ <http://www.gnupg.com/> Secure your email!!!
* Public key available at keyserver.pgp.com <http://keyserver.pgp.com/>
**/
♺ This email uses 100% recycled electrons. Don't blow it by printing!
There are only 2 hard things in computer science: Cache invalidation, naming things, and off-by-one errors.