Felipe, questions asked by Tiago are valid questions, which try to underline issues that NuttX development team will face. All these points will have to be processed by someone, take responsibility, and maintain in the long term.
>From my perspective current situation is optimal: NuttX does neither support nor maintain external SDKs, Espressif here is an exception because we have solid support from vendor, Espressif developers are also members of the NuttX development team. What you propose is tremendous amount of additional work, and responsibilities, for a small team of volunteers. My feeling is that big vendor can handle this better, and in fact we always have very good support from Espressif in any case of code, issues, testing, reviews, etc. There is some resistance here, because people did cost-benefit analysis, and they may not be able to handle yet another batch of additional work. Can we try to improve things with current setup? What is the exact problem we want to solve here? Delayed reviews? If so maybe someone from NuttX could get write access to esp-sdk repo and help in reviews? -- CeDeROM, SQ7MHZ, http://www.tomek.cedro.info -- CeDeROM, SQ7MHZ, http://www.tomek.cedro.info On Thu, Aug 20, 2026 at 8:00 PM Felipe Moura Oliveira <[email protected]> wrote: > > Hi Tiago, > > This thread has grown to nearly 100 emails, and the topic has been thoroughly > discussed. > > I will not ask you to go back and reread everything, as I understand you must > be very busy, as you said, you're a PMC member. There are currently several > open issues requesting ESP support, and you have frequently mentioned that > your other priorities prevent you from looking into them. > > Thank you to everyone who participated in this discussion. > > Best regards, > > > Em qui., 20 de ago. de 2026 às 16:40, Tiago Medicci Serrano > <[email protected]> escreveu: >> >> Hi all, >> >> Thanks for the input, Felipe. However, what should we (the NuttX community) >> vote for? This isn't clear to me... >> >> In my humble opinion (talking as a PMC member), we should better define what >> we would expect by hosting HAL repositories under the Apache NuttX umbrella. >> >> - Does it mean NuttX is officially supporting 3rd-party HALs like Zephyr? >> - Who are the maintainers of the repository? >> - Which other repositories would be forked, considering externally provided >> sources used in NuttX (kernel)? >> - What are (and who defines) the rules for the forked repositories? >> - Are there any licensing restrictions? >> - What about CI usage to test automated sync? >> - Where should such policies be documented? >> - Are there any other questions to be asked/answered? >> >> Those (and potentially other) questions should be properly answered (at >> least as a plan) before we move forward towards a voting session. >> >> Best regards, >> >> >> Em qui., 20 de ago. de 2026 às 16:22, Felipe Moura Oliveira >> <[email protected]> escreveu: >>> >>> Hi all, >>> >>> I believe it is valid and mature to open the vote at this stage. >>> >>> The topic has been thoroughly discussed, and regardless of the outcome, it >>> is important to bring this discussion to a conclusion, both for the record >>> and to allow the project to move forward. >>> >>> Best regards, >>> Felipe Moura de Oliveira >>> >>> >>> Em qui., 20 de ago. de 2026 às 15:58, Tiago Medicci Serrano >>> <[email protected]> escreveu: >>>> >>>> Hi all, >>>> >>>> Regarding the latest messages in this conversation, there is something >>>> we'd like to highlight about how `esp-hal-3rdparty` is actually used (and >>>> maintained): as mentioned in its description at >>>> https://github.com/espressif/esp-hal-3rdparty/, it contains both sync and >>>> release branches. The sync branch is automatically derived from ESP-IDF's >>>> sources (`components` folders). On top of that, the release branch is the >>>> branch that NuttX actually uses. It contains changes on top of >>>> `components` folder and the adapter layer for NuttX. >>>> >>>> We expect that HAL will need less and less changes from its sources >>>> (especially directly in the `components` folder), which means that >>>> contributions will be far apart (they already are) and even less frequent. >>>> That being said, the way HAL is treated would have to be kept exactly the >>>> same to keep the ability to sync with upstream. The moment any commit >>>> breaks this is the moment contribution is no longer feasible. So, while >>>> technically anyone in the community could do anything, they would have to >>>> do it in the same way it is done today. >>>> >>>> Additionally, it's worth mentioning that this is how it is done in Zephyr. >>>> Espressif team constantly checks whether commits don't break the ability >>>> to sync and gather the latest changes from ESP-IDF, which enables us to >>>> support newer SoCs faster and promptly handle bug fixes related to the low >>>> level. >>>> >>>> Best regards, >>>> >>>> Em sáb., 15 de ago. de 2026 às 16:39, Alan C. Assis <[email protected]> >>>> escreveu: >>>>> >>>>> Just an example here: >>>>> >>>>> >>>>> >>>>> The ESP32-S31 is done by someone from the community. >>>>> >>>>> We could add it or wait 6 months from getting it (ESP32-P4 delay more >>>>> than 2 year, and still not complete yet) >>>>> >>>>> BR, >>>>> >>>>> Alan >>>>> >>>>> >>>>> On Sat, Aug 15, 2026 at 12:16 PM Felipe Moura Oliveira >>>>> <[email protected]> wrote: >>>>>> >>>>>> Hi Jehudi, >>>>>> >>>>>> Here are my thoughts on the questions you raised: >>>>>> >>>>>> 1. What are the advantages and consequences of changing the location? >>>>>> It will allow people with write access to NuttX to review and approve >>>>>> PRs, removing the pressure from a few individuals and the vendor's >>>>>> internal >>>>>> queue. >>>>>> >>>>>> 2. What is the goal of changing the location: closer to NuttX and it's >>>>>> style? >>>>>> In addition to the point above, the goal is to prevent NuttX from >>>>>> becoming increasingly dependent on external agents, especially vendors. >>>>>> >>>>>> 3. Who will be responsible? >>>>>> The community, meaning anyone who is interested in contributing. >>>>>> >>>>>> 4. Who will write the code? >>>>>> The community, meaning anyone who is interested in contributing. >>>>>> >>>>>> 5. Who will maintain the code? >>>>>> The community, meaning anyone who is interested in contributing. >>>>>> >>>>>> 6. Who will ensure that newer devices are supported? >>>>>> The community, meaning anyone who is interested in contributing. >>>>>> >>>>>> Best regards, >>>>>> Felipe Moura de Oliveira >>>>>> >>>>>> >>>>>> Em sex., 14 de ago. de 2026 às 12:44, Laczen JMS <[email protected]> >>>>>> escreveu: >>>>>> >>>>>> > Hi Felipe, >>>>>> > >>>>>> > Hosting the repository like zephyr style would just mean moving it >>>>>> > under nuttx but leaving the merge rights to people from espressif. >>>>>> > They understand how this repository is updated and what the work is >>>>>> > when changes are applied. I wouldn't mind this at all, but I don't see >>>>>> > any advantages. In zephyr the repo is there to ensure that whatever >>>>>> > version of zephyr is checked-out it has the correct library, in NuttX >>>>>> > this is achieved in a different way with less work by updating the >>>>>> > commit-sha of the repository. >>>>>> > >>>>>> > Please read the statement from espressif carefully: when the comunity >>>>>> > would decide (according to community rules) that the esp_hal_3rdparty >>>>>> > is to be moved under nuttx (and possibly the merging rights would be >>>>>> > changed) espressif will evaluate whether contribution is feasible. >>>>>> > This could lead to the decision that contribution is no longer >>>>>> > feasible, leaving NuttX behind with less support. Up to now espressif >>>>>> > has reviewing and >>>>>> > creating PR's for NuttX because they understand how the interaction >>>>>> > with esp_hal_3rdparty is happening, expecting espressif to do the same >>>>>> > for a library that is handled by NuttX could be wishfull thinking. >>>>>> > >>>>>> > I would conclude with the following questions: >>>>>> > >>>>>> > 1. What are the advantages and consequences of changing the location? >>>>>> > 2. What is the goal of changing the location: closer to NuttX and it's >>>>>> > style? >>>>>> > 3. Who will be responsible? >>>>>> > 4. Who will write the code? >>>>>> > 5. Who will maintain the code? >>>>>> > 6. Who will ensure that newer devices are supported? >>>>>> > >>>>>> > If any of these questions has an unclear answer it is better too leave >>>>>> > it as is and embrace the work that espressif is doing for NuttX (both >>>>>> > in PR and publicity on their blog). >>>>>> > >>>>>> > Kind regards, >>>>>> > >>>>>> > Jehudi >>>>>> > >>>>>> > >>>>>> > Op vr 14 aug 2026 om 14:56 schreef Felipe Moura Oliveira < >>>>>> > [email protected]>: >>>>>> > >>>>>> > > >>>>>> > > Hi Laczen, >>>>>> > > >>>>>> > > Following our last discussion, and given Espressif's statement that >>>>>> > > they >>>>>> > > will conform to community rules and are waiting for the community to >>>>>> > decide >>>>>> > > first, are you now more comfortable with hosting the repository like >>>>>> > Zephyr >>>>>> > > style? >>>>>> > > >>>>>> > > Is anyone else concerned about this? >>>>>> > > >>>>>> > > Em sex., 14 de ago. de 2026 às 09:01, Tiago Medicci Serrano < >>>>>> > > [email protected]> escreveu: >>>>>> > > >>>>>> > > > Hi Felipe, >>>>>> > > > >>>>>> > > > It may have been lost in this huge thread, but the official >>>>>> > > > position I >>>>>> > have >>>>>> > > > is the one that I already sent to the list: >>>>>> > > > >>>>>> > > > 1) Espressif always said it would conform to community rules; >>>>>> > > > hence, >>>>>> > why we >>>>>> > > > > have been absent. The community must decide first, and then we >>>>>> > > > > can >>>>>> > see >>>>>> > > > > whether the contribution is feasible. >>>>>> > > > > 2) We do not block contributions, and we never saw ourselves as >>>>>> > 'owning' >>>>>> > > > > Espressif support in NuttX. We do have opinions, and we do try to >>>>>> > > > optimize, >>>>>> > > > > and we do listen to others. We actually expect more people >>>>>> > > > > willing to >>>>>> > > > help. >>>>>> > > > > 3) So far, the discussed options were: >>>>>> > > > >>>>>> > > > >>>>>> > > > - >>>>>> > > > >>>>>> > > > Total AI rewrite (I assume at the register level). >>>>>> > > > >>>>>> > > > - >>>>>> > > > >>>>>> > > > Fork, taking it from there - maybe going Zephyr style. >>>>>> > > > >>>>>> > > > - >>>>>> > > > >>>>>> > > > Replacing HAL slowly, piece by piece. >>>>>> > > > >>>>>> > > > >>>>>> > > > >>>>>> > > > Best regards, >>>>>> > > > >>>>>> > > > >>>>>> > > > Em qua., 12 de ago. de 2026 às 08:10, Felipe Moura Oliveira < >>>>>> > > > [email protected]> escreveu: >>>>>> > > > >>>>>> > > > > Hi Tiago, >>>>>> > > > > >>>>>> > > > > Do you have any feedback from the Espressif team regarding this >>>>>> > thread? >>>>>> > > > > Could you check with your team if we could maintain the esp-hal >>>>>> > under the >>>>>> > > > > NuttX GitHub organization? This way, we would be able to add more >>>>>> > > > reviewers >>>>>> > > > > to help with the process. >>>>>> > > > > >>>>>> > > > > Best, >>>>>> > > > > >>>>>> > > > > Em sáb., 8 de ago. de 2026 às 08:33, Alan C. Assis < >>>>>> > [email protected]> >>>>>> > > > > escreveu: >>>>>> > > > > >>>>>> > > > > > Exactly! >>>>>> > > > > > >>>>>> > > > > > And for Zephyr the HAL for ESP32 chips already existed before >>>>>> > Espressif >>>>>> > > > > > started contributing to it. >>>>>> > > > > > >>>>>> > > > > > NuttX also had a native port before Espressif started >>>>>> > > > > > contributing. >>>>>> > > > > > >>>>>> > > > > > When ESP32 chip was created, some engineers from Espressif was >>>>>> > willing >>>>>> > > > to >>>>>> > > > > > use NuttX instead of FreeRTOS and sent some boards to Greg. He >>>>>> > > > > > did >>>>>> > the >>>>>> > > > > > basic port, but because their internal FreeRTOS port was more >>>>>> > advanced, >>>>>> > > > > > they decided to keep using it. >>>>>> > > > > > >>>>>> > > > > > BR, >>>>>> > > > > > >>>>>> > > > > > Alan >>>>>> > > > > > >>>>>> > > > > > On Saturday, August 8, 2026, Laczen JMS <[email protected]> >>>>>> > wrote: >>>>>> > > > > > >>>>>> > > > > > > Hi Felipe, >>>>>> > > > > > > >>>>>> > > > > > > Indeed, just like your PR was committed (merged) into >>>>>> > > > esphal_3rdparty. >>>>>> > > > > > The >>>>>> > > > > > > hal_espressif for zephyr is maintained by espressif. The only >>>>>> > > > > difference >>>>>> > > > > > > between NuttX and zephyr is where it is located, not who >>>>>> > maintains >>>>>> > > > it. >>>>>> > > > > > > >>>>>> > > > > > > Kind regards, >>>>>> > > > > > > >>>>>> > > > > > > Jehudi >>>>>> > > > > > > >>>>>> > > > > > > On Sat, Aug 8, 2026, 09:31 Felipe Moura Oliveira < >>>>>> > > > [email protected]> >>>>>> > > > > > > wrote: >>>>>> > > > > > > >>>>>> > > > > > > > Hi Laczen, >>>>>> > > > > > > > >>>>>> > > > > > > > I think you might be confusing the author with the >>>>>> > > > > > > > committer. >>>>>> > > > > > > > >>>>>> > > > > > > > The change was indeed committed by someone from Espressif >>>>>> > > > > > > > who >>>>>> > has >>>>>> > > > > write >>>>>> > > > > > > > access to the repository, but the actual author of the code >>>>>> > was an >>>>>> > > > > > > external >>>>>> > > > > > > > contributor using a @gmail >>>>>> > > > > > > > >>>>>> > > > > > > > >>>>>> > > > > > > > *--Felipe Moura de Oliveira* >>>>>> > > > > > > > Linkedin >>>>>> > > > > > > > <https://www.linkedin.com/in/felipe-oliveira-75a651a0 >>>>>> > > >>>>>> > > > > > > > >>>>>> > > > > > > > On Sat, 8 Aug 2026 at 02:31 Laczen JMS >>>>>> > > > > > > > <[email protected]> >>>>>> > > > wrote: >>>>>> > > > > > > > >>>>>> > > > > > > > > Alan, >>>>>> > > > > > > > > >>>>>> > > > > > > > > That person is from espressif, and the commit of your >>>>>> > previous >>>>>> > > > > > > reference >>>>>> > > > > > > > > was also done by someone from espressif. >>>>>> > > > > > > > > >>>>>> > > > > > > > > Kind regards, >>>>>> > > > > > > > > >>>>>> > > > > > > > > Jehudi >>>>>> > > > > > > > > >>>>>> > > > > > > > > On Fri, Aug 7, 2026, 22:51 Alan C. Assis >>>>>> > > > > > > > > <[email protected]> >>>>>> > > > > wrote: >>>>>> > > > > > > > > >>>>>> > > > > > > > > > Ok, so this one the also the maintainer of something >>>>>> > > > > > > > > > else >>>>>> > and >>>>>> > > > > this >>>>>> > > > > > is >>>>>> > > > > > > > why >>>>>> > > > > > > > > > is has a commit there: >>>>>> > > > > > > > > > >>>>>> > > > > > > > > > >>>>>> > > > > > > > > > >>>>>> > > > > > > > > >>>>>> > > > > > > > https://github.com/zephyrproject-rtos/hal_espressif/commit/ >>>>>> > > > > > > 34795d7847426426c618002789af2667cc262f6d >>>>>> > > > > > > > > > >>>>>> > > > > > > > > > There are many commits from people outside Espressif, >>>>>> > > > > > > > > > what >>>>>> > you >>>>>> > > > > said >>>>>> > > > > > > is >>>>>> > > > > > > > > (to >>>>>> > > > > > > > > > avoid saying I'm accusing you) "not accurate". >>>>>> > > > > > > > > > >>>>>> > > > > > > > > > BR, >>>>>> > > > > > > > > > >>>>>> > > > > > > > > > Alan >>>>>> > > > > > > > > > >>>>>> > > > > > > > > >>>>>> > > > > > > > >>>>>> > > > > > > >>>>>> > > > > > >>>>>> > > > > >>>>>> > > > > >>>>>> > > > > -- >>>>>> > > > > >>>>>> > > > > *--Felipe Moura de Oliveira* >>>>>> > > > > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0> >>>>>> > > > > >>>>>> > > > >>>>>> > > >>>>>> > > >>>>>> > > -- >>>>>> > > >>>>>> > > *--Felipe Moura de Oliveira* >>>>>> > > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0> >>>>>> > >>>>>> >>>>>> >>>>>> -- >>>>>> >>>>>> *--Felipe Moura de Oliveira* >>>>>> Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0> >>> >>> >>> >>> -- >>> -- >>> Felipe Moura de Oliveira >>> Linkedin > > > > -- > -- > Felipe Moura de Oliveira > Linkedin
