Activity since 2025-07-19

• esp-rs/esp-hal repo issues and pull requests

  • Start i2s circular with borrow #3612 [AstrickHarren]

    This should partially fix #3285 basically we don't take owner ship of i2s when starting transfer. Not sure if this...

  • Use new instablility release, remove rustdoc json hacks #3834 [MabezDev] automation

    ratatui/instability#23 and esp-hal/xtask/src/semver_check.rs Lines 111 to 237 ...

  • Update instability #3836 [bugadani] skip-changelog

    Closes #3834

    • bjoernQ: Happy to see the workaround go away!

  • re-building `esp-riscv-rt` on top of `riscv-rt` 0.13.0 #2390 [romancardenas] package:esp-riscv-rt

    One of the main fresh features of riscv-rt 0.13.0 is that, together with riscv 0.12.0, now it should be possible to...

  • Move irrelevant global symbols away from esp-wifi #3823 [MabezDev] package:esp-wifi

    cc: #2865 mostly, but I'm sure there are others in some of the compat modules. If we really need them, we should try...

  • Examples restructure #3830 [MabezDev] examples

    Our examples are not great. Some are fine to keep, but I think they can be organized better. I propose the following...

  • Panic in exception handler #3838 [bugadani]

    With this change, esp-backtrace can print exceptions even if it is not the panic handler. To be quite honest, maybe...

    • MabezDev: I think the default handlers just panic, so I suppose this change now aligns with that. Whether...

    • bugadani: We can still allow hooking into exception in one way or another, I think, even if that will be...

    • bjoernQ: Probably esp-hal would be the place to handle them 🤔 and have a way to hook into exception...

  • Add SOC config that is hidden from the peri table #3839 [bugadani] skip-changelog

    • bjoernQ

  • Remove all bleps and blocking network stack examples #3831 [MabezDev] examples

    It's pretty clear to me two things: async for any kind of comms is the future troubleBLE and embassy-net are the...

    • playfulFence: We don't seem to have a trouBLE example, so I'll take care of adding it, just so @jessebraham...

    • bjoernQ: We don't seem to have a trouBLE example, so I'll take care of adding it, just so @jessebraham...

    • playfulFence: Oh, correct, that is why we don't have it 😅 Then I think, just adding clear links to the trouBLE...

  • Mark free standing functions in `wifi` module as unstable #3821 [MabezDev] package:esp-wifi

    I'm not convinced wifi_state etc are ready to be stable. With wifi_state, we can definitely miss events, I think we...

  • Make features other than wifi unstable #3820 [MabezDev] package:esp-wifi

    Unlike esp-hal, we probably want to use instabilities unstable-feature flags for the things we're not stabilizing...

  • cleanup `esp-wifi` README #3841 [playfulFence] skip-changelog

    Make it clear that examples of usage with trouBLE might be found in the aforementioned project's repo. Also, some...

    • MabezDev: Thanks!

    • bugadani: Reading two passive sentences is pretty rough. the preferred ... is considered to be is...

    • bugadani: We don't have empty cells.

    • bjoernQ: preemt_builtin is "a bare minimum RISCV and Xtensa round-robin task scheduler" and preemt is...

      • playfulFence: Do we even need that section at all? 😆 Like, who's the potential beneficiary of this? It's A)...

    • bjoernQ: LGTM

  • use riscv 0.14.0 #3842 [bjoernQ] skip-changelog

    Thank you for your contribution! We appreciate the time and effort you've put into this pull request. To help us...

    • bugadani

  • Rng should rate limit to avoid draining entropy before it is added by the hardware #2443 [MabezDev] bug peripheral:rng

    See...

  • Explore a `getrandom` API for RNG (and in the future crypto) #3826 [MabezDev] RFC peripheral:rng

    I'm not convinced that owned crypto and rng make sense in the long term. It's pretty rare for only one thing to want...

  • TRng should not own the RNG peripheral #2372 [nishanthkarthik] peripheral:rng

    Motivations Wifi on esp32c6 requires an Rng instance. TLS requires a Trng instance. But both Rng and Trng consume...

  • Fix GPIO matrix typo #3840 [bugadani] skip-changelog

    This PR fixes an issue where trying to use signals that can't be routed through the GPIO matrix causes a panic.

    • bjoernQ

  • Rewrite RNG #3829 [bugadani]

    This PR decouples the RNG entropy source and the RNG random number generator. It addresses #3826 while at the same...

    • MabezDev

      • MabezDev: What's the TODO? :D

        • MabezDev: Do we know yet? :D

        • bugadani: I know nothing. Based on TRM, we receive 2 bits of entropy when the SAR ADC is enabled, so 16...

    • MabezDev: Can we use anything from https://docs.rs/riscv/latest/riscv/? Looks like the csr macros...

      • MabezDev: Oops, you already are :D

    • MabezDev: I think this turned out really nicely, and will end up closing at least three issues! Thank you!

  • Encode AES key lengths & modes in metadata #3843 [bugadani] skip-changelog

    • bjoernQ: nitpick but other variants have a nice description

    • bjoernQ

  • Sha algos as metadata #3844 [bugadani] skip-changelog

    • JurajSadel

      • JurajSadel: nit: missing docstring

      • JurajSadel: Just curious, if this would work Suggested change ...

        • bugadani: Unlikely, quote! {} creates a TokenStream, not an option

  • Restore printing exception context #3850 [bugadani] skip-changelog

    This was accidentally removed in #3838

    • bjoernQ: LGTM

  • Add HW DTR support (port of #2994) #3796 [i404788]

    Submission Checklist 📝 I have updated existing examples or added new ones (if applicable). I have used cargo...

  • Bump `embassy-usb` version #3848 [HaoboGu]

    Thank you for your contribution! We appreciate the time and effort you've put into this pull request. To help us...

    • bugadani: Thank you!

  • ESP32S3: there are multiple different versions of crate `embassy_usb_driver` in the dependency graph #3846 [HaoboGu]

    embassy-usb and embassy-usb-driver have released a new version, when using it with esp32s3, the USB module broken...

    • bugadani: Sooo... could you use the previous embassy-usb version?

    • HaoboGu: Yeah, the embassy-usb v0.4 works, but it would be great if esp-hal could be compatible with the...

    • bjoernQ: We'll update it at some point, and it will be available in the release after that. However, the...

    • HaoboGu: Thanks for the reply, I've made a PR for it. Let's wait for a new release :)

    • HaoboGu: embassy-usb-synopsys-otg was just released, so I marked my PR ready for review now 😄

  • Yield in usleep #3852 [bugadani] skip-changelog

    This PR lets esp-wifi tasks yield to other tasks while sleeping. The scheduler would preempt longer sleep times...

    • bjoernQ: LGTM

  • panic timed eventq_get not yet supported when central l2cap device disconnects #3760 [Szybet] package:esp-wifi status:awaiting-response

    Bug description I'm talking about this line: esp-hal/esp-wifi/src/ble/npl.rs Line...

    • Szybet: INFO - Starting advertising and GATT service ERROR - [adv] error:...

    • bugadani: Your application crashes because you advertise with a name that is too long. That's hardly our...

    • Szybet: As mentioned here I know it's hacky, but it gets the job done and triggers the same error....

    • bugadani: Alright, I'll try again, honestly your changes got lost among the reformatting you've done in the diff.

    • bugadani: Thanks I was able to reproduce the issue. It's somewhat random, I must have been lucky last...

  • Implement NPL timed dequeue #3762 [bugadani]

    Closes #3760 Untested as I can't reprouce the original issue.

  • Fix ble crash #3854 [bugadani]

    This PR fixes the second problem reported in #3760 by inserting a short sleep before deinitializing the BLE...

    • JurajSadel: Thanks

  • feat(esp-wifi): add Country Code #3837 [denysvitali]

    NoteFeature was "vibe-coded" with Claude Sonnet 4. Tested and it works for me. Submission Checklist 📝 I have...

    • copilot-pull-request-reviewer: Pull Request Overview This PR adds support for extracting and returning country codes from WiFi...

      • copilot-pull-request-reviewer: The magic number 2 is used multiple times for country code length validation. Consider defining...

      • copilot-pull-request-reviewer: The magic number 3 for the country code array size should be defined as a constant to improve...

    • JurajSadel

    • bugadani

      • bugadani: Wasting 12 bytes AND a heap allocation for something so rarely needed is not the greatest...

      • bugadani: This example is really not necessary. If you think it is, then it can be part of the...

  • add embassy #[main] callback passing #3813 [dragonnn]

    Submission Checklist 📝 I have updated existing examples or added new ones (if applicable). I have used cargo...

    • bugadani: I think Scott meant we shouldn't complicate the macro with magic like this, but we have no...

    • dragonnn: While I would mostly agree I think they is a valid use case for it: esp-wifi could provide a...

  • Don't try to write FUNC_IN_SEL_CFG of non-matrix signal #3849 [bugadani] skip-changelog

    This PR stops after selecting the digital function, for signals that can't be routed through the GPIO matrix....

    • Dominaezzz

  • esp-wifi: remove bunch of FIXMEs and clippy cleanup #3851 [JurajSadel] skip-changelog

    • bugadani: I'm not in love with this. Internal would probably be enough, but Interface needs a better name,...

    • bugadani: Thanks

  • esp-hal: i2s: add a way to stop I2sWriteDmaTransferAsync #3847 [anarsoul]

    Motivations Currently there is no way to stop I2sWriteDmaTransferAsync Solution Implement stop() method, like in...

    • Dominaezzz: This might be fixed by #3639

  • esp32s3 wifi station mode can scan but found nothing after wakeup from sleep #3751 [vult0306] bug low power package:esp-wifi

    Case 1: Start esp32s3 wifi dhcp example, can scan and connect to wifi normally. Case 2: When device is up, put...

    • TuEmb: I also have the same problem with esp32s3

    • TuEmb: I (24) boot: ESP-IDF v6.0-dev-1002-gbfe5caf58f-dirty 2nd stage bootloader I (24) boot: compile...

  • I2C: low level transaction builder #2598 [bugadani] peripheral:i2c 1.0 non-breaking

    We should expose a low level interface, that passes a number of commands to the driver. This allows users to freely...

    • Dominaezzz: Would this be different from the Operation enum? The ack settings could be added to that. Unless...

    • Dominaezzz: Also, is this issue different from #2505?

  • sdio: initial SDIO HAL implementation #3503 [rmsyn]

    Thank you for your contribution! We appreciate the time and effort you've put into this pull request. To help us...

    • rmsyn: I added an initial integration test for sdio to test out some ideas for the static DMA...

    • bugadani: SDIO_CLK cannot be routed through the GPIO matrix #3840

    • rmsyn: #3840 Thank you, that fixed the issue with SDIO_CLK. The test panics now because the SDIO_*...

    • rmsyn: What are your thoughts on deviating from the TRMs, and creating a pseudo-mapping for the...

    • bugadani: so maybe we could use those for SDIO_[CLK, CMD]? We shouldn't do any such hackery. These...

    • rmsyn: We shouldn't do any such hackery. Yeah, the more I thought about it, the more this looked like...

    • rmsyn: Could you please check if #3849 helps you? Otherwise, could you please paste the relevant panic...

    • bugadani: This fixed the esp32c6 error! Awesome! It's fine to ignore ESP32 for now, we can figure that...

    • rmsyn: It's fine to ignore ESP32 for now, we can figure that chip out once C6 works as you want it to....

  • enabling psram feature causes panic after updating to esp-hal 1.0.0-beta.1 and espflash 4.0 #3765 [kendra-pdx] bug peripheral:psram

    Bug description Some combination of espflash (4.0.1) and esp_hal (1.0 beta 1) seem to break psram on my ESP32S3...

    • kendra-pdx: re-tested today on: ESP32-S3 esp-hal 1.0.0-rc.0` esp-bootloader-esp-idf 0.2.0 espflash...

  • Stop using `ok` in esp-wifi/esp-ieee802154 crates codebase. #3853 [playfulFence] skip-changelog

    closes #2755 I'm not entirely sure if that's it for resolving the issue though. there's another usage of .ok() in...

    • bugadani: Use the unwrap! macro instead. TBH we might want a variant of check_error that panics with a...

    • bugadani: Should be fine after a rustfmt run

  • `ok()` used for error handling #2755 [ede1998] bug package:esp-ieee802154 package:esp-wifi

    Thanks for maintaining and developing ESP HAL. It was really fun when I developed my previous project with it. Bug...

  • Use riscv-rt's startup code #3857 [bjoernQ] skip-changelog

    Thank you for your contribution! We appreciate the time and effort you've put into this pull request. To help us...

  • Move RSA operand sizes to metadata (including AES fallout) #3845 [bugadani]

    • bugadani: RSA quickly proved that this wasn't the best idea, lol, we don't want 200 symbols just for the...

    • bugadani: This is bad, I'll be looking for alternatives.

    • bjoernQ: LGTM

  • Fold ieee801215 into esp-rf #3817 [MabezDev] package:esp-ieee802154 package:esp-wifi

    Whilst the ieee driver is pure rust, both share the same phy init blobs, therefore I think we should merge these...

  • Rename `esp-wifi-sys` to `esp-radio-sys` #3859 [playfulFence] RFC

    As an extension of #3815

    • MabezDev: We might not be able to do this. We got special permission to have increased crate size for...

    • playfulFence: I don't mind it at all either (especially given the fact this crate exists exclusively for...

    • bjoernQ: Ah good point about the size limit exception .... totally forgot about it 🤦‍♂️ If someone wants...

    • bugadani: One package per MCU sounds good to me even if some wouldn't fit the size limit. Downloading...

  • Clean up NPL a bit #3764 [bugadani] skip-changelog

    This PR attempts to clean up NPL a bit by making the following changes: Take mut pointers where applicable. The C...

    • bugadani: I think these changes should generally be okay, but they aren't the different I was hoping for.

  • Strip `EspRadio` prefixes across the `esp-radio` codebase #3862 [playfulFence] package:esp-wifi

  • Some modules have interesting descriptions in docs #3856 [bjoernQ]

    Some modules show unhelpful information as their description on the index page

    • bugadani: This issue is caused by our weird module structure. pub use-ing modules unstably will generate...

    • bjoernQ: yeah that's the problem I guess - definitely an eyesore

  • Move out the builtin scheduler to esp-preempt crate #3818 [MabezDev] package:esp-wifi

    This is hopefully simpler than it sounds, and given that we already have another scheduler impl used with esp-wifi...

  • Create `esp-preempt-driver` #3819 [MabezDev]

    In here will live the current Scheduler trait (might need some name bikeshedding), and impl macro. This crate is...

  • Split out preempt driver and default impl #3855 [bugadani]

    Closes #3818 Closes #3819 Additional changes include: usleep is now part of the scheduler implementation added a...

    • bjoernQ: This came out quite nice! I was thinking if there should be a token from the scheduler init to...

    • bugadani: I suspect the yield_now in esp_wifi::init might be a problem point. usleep also shouldn't yield...

  • esp-config: potential license compatibility issue #3822 [kaspar030]

    esp-config depends on evalexpr, which unfortunately changed its license to AGPLv3 starting with version 12.0.0. As...

  • Replace evalexpr #3860 [bugadani] skip-changelog

    Somni is a very young language at this point, but the expression evaluator should be good enough to replace...

    • bjoernQ: Thanks! somni looks nice

    • bugadani: Thanks, tbh it still needs quite a bit of polish but at least we have something 🙃

  • Remove raw interrupt map address #3863 [bugadani] skip-changelog

    • bjoernQ: What about C2?

    • bugadani: Either zed didn't find it, or I didn't notice :)

    • bjoernQ: LGTM

  • Move TIMG clock source defaults to metadata #3864 [bugadani] skip-changelog

    This PR isn't perfect, but at least we don't need to store constants in code. Ideally, we'd describe the available...

  • Hide C types from public API #3824 [MabezDev] package:esp-wifi

    We have to be especially careful with our From/Into impls, these can leak these types into he public API easily.

  • Move REF_TICK and RC_FAST_CLK to metadata #3866 [bugadani] skip-changelog

  • Rename esp-wifi to esp-rf #3815 [MabezDev]

    I hope this is not too controversial. For a while, esp-wifi has had more things in it other than WiFi. Imo we should...

    • playfulFence: There's something I don't like about the rf thing here... As minimum, it might look like esp-rs...

    • MabezDev: esp-hal-rf 😛 (since it depends on esp-hal) I don't think we really want to do this. The only...

  • Should esp-wifi take ownership of ADC2? #2150 [Frostie314159] package:esp-wifi chip:esp32

    Motivations On the ESP32, you can't use the ADC2 peripheral, during Wi-Fi operations, which the current API, this is...

  • Rename `esp-wifi` to `esp-radio` (part 1) #3858 [playfulFence] skip-changelog

    Closes #3815 Second PR will rename esp-wifi-sys to esp-radio-sys, but it is more of an internal thing, so I assume...

    • bugadani: This probably should say esp_wifi

    • bugadani

      • bugadani: The rename should be mentioned here somehow, as technically there is no esp_wifi::init anymore

      • bugadani: lol let's just remove the whole EspWifi prefix

        • playfulFence: Agreed, but let's do it in a separate PR, just so we only have "renaming" changes here And we...

    • playfulFence: Skip changelog because of GitHub complaining about esp-hal changes, but none of them are user-facing

    • playfulFence: This is ugly, but otherwise rustfmt does this to me

    • bugadani: Why was this changed?

      • playfulFence: No reason for that, thank you!

    • bugadani: Welp, now this PR has a few conflicts. Sorry 'bout that.

    • Frostie314159: Do you think it would be possible to make the phy initialization generic (ideally in a separate...

    • bugadani

      • bugadani: Suggested change ...

      • bugadani: Suggested change + esp-radio =...

      • bugadani: I'm not sure we want to be this pedantic 🤔

        • playfulFence: Done

    • bugadani: We can drop these links I think

    • bugadani

  • esp-wifi: Add unstable feature, mark `ble`, `esp-now`, `csi`, `sniffer`, and `smoltcp` features and APIs as unstable #3865 [JurajSadel]

    cc #3821, #3820, #3825 This will most likely have a follow up PR focusing on esp-wifi feature (#3700 (comment)).

    • bugadani: Does this (and the rest of the function pointer tables) have to be public?

    • bugadani: This will result in something like #3856. Let's remove the asynch module instead.

  • Get rid of `EspRadio` prefix in `esp-radio` structs #3869 [playfulFence]

    closes #3862

    • bugadani

      • bugadani: Suggested change ## `EspRadio`...

      • bugadani: Suggested change - Rename `esp-wifi`...

  • Reorganize PSRAM #3867 [bugadani] skip-changelog

    This PR moves PSRAM code around to fix part of the rustdoc issue in #3856

  • Reorganize the efuse module #3868 [bugadani] skip-changelog

    Closes #3856

  • Fold `esp-ieee802154` package into `esp-radio` #3861 [playfulFence]

    closes #3817 this PR is based on #3858, so it can't be merged before it

    • bugadani: wtf is pub(self) :D is it just private?

    • bugadani: Suggested change #[cfg(all(feature =...

    • bugadani: Might just be github UI but the random spaces are random.

    • bugadani: Should these be gated by the ieee802154 feature?

    • bugadani: Why are you copying over old changelogs?

      • playfulFence: Just to have its history as part of esp-radio now. I was big time hesitant about this, so if you...

    • bugadani: This won't be the case forever. You can add another column I guess, and also change Coex to Coex...

    • bugadani: Don't remove the test data, use a different package

  • Use move based api for I2S #3639 [AstrickHarren]

    This is to fix #3285, as in #3612 @Dominaezzz laid out reasons for a move-based i2s API so I thought I may help a...

    • AstrickHarren: Think this may be ready. Added the TxStreamBuf that basically works symmetrically with the...

    • Dominaezzz: I tried to utilize DescriptorEmpty interrupt as a sign for DMA rx completion but I couldn't get...

    • Dominaezzz: First batch of comments. Overall looks good, I'm excited to land this

      • Dominaezzz: I don't understand this comment

      • Dominaezzz: Both are correct but the one in the comment is certainly clearer.

        • AstrickHarren: I'm not sure why, shouldn't descriptors.len() be like very small compared to chunk_size? I could...

        • Dominaezzz: It should probably be, but regardless of that, the excess is always going to be smaller than the...

      • Dominaezzz: There needs to be a way to push some data into the buffer before the transfer is started,...

        • AstrickHarren: One way to do this is to put the head of the actual data before creating the buffer, as is done...

        • Dominaezzz: One way to do this is to put the head of the actual data before creating the buffer You could...

      • Dominaezzz: Feel free to delete these structs from the dma module too, they won't be needed anymore

      • Dominaezzz: These too

      • Dominaezzz: Yes, DmaRxBuffer guarantees this. You can drop this comment,

      • Dominaezzz: For consistency, call this I2sTxDmaTransfer

      • Dominaezzz: You have to stop the DMA TX channel too

      • Dominaezzz: This should more explicit and say Waits for the [DmaTxInterrupt::Done] interrupt.

    • Dominaezzz: A small downside to this PR is we've lost support of u16, u32, etc. but that can be brought back later.

    • AstrickHarren: Thanks a lot for PR reviewing @Dominaezzz appreciated!

    • AstrickHarren: Which esp device are you unsure of? I meant esp32c3 and similar devices in the sense that they...

    • Dominaezzz: You can now delete the prepare_transfer_without_start function from the dma module too

    • Dominaezzz: Do you actually need to listen to this here? This should be up to the user

    • Dominaezzz: You can remove DescriptorChain too

    • Dominaezzz: This example can (and probably should) be updated to use a DmaLoopBuf instead. I can take care...

      • AstrickHarren: Not entirely sure if we still want loop buf? I mean I don't really see the advantages of making...

      • Dominaezzz: That specific qa-test plays a repeated pattern indefinitely. With DmaLoopBuf, you can set the...

      • AstrickHarren: Where is that quite from? I could be wrong but I think when I read the code few weeks ago...

      • Dominaezzz: Yes, it's not meant to be modified once the transfer starts

      • anarsoul: you can set the pattern once, start the transaction and it'll play indefinitely without CPU...

      • Dominaezzz: you can set the pattern once, start the transaction and it'll play indefinitely without CPU...

      • anarsoul: For transferring a repeating pattern of data I mean for the real-world application

      • Dominaezzz: Take a look at the qa-test application this thread is on.

created at 2025-07-26 02:07 by dinghy 0.13.1.