STM32H5: ADC circular GPDMA so battery/current track live values - #15419
Merged
haslinghuis merged 1 commit intoJul 16, 2026
Merged
Conversation
|
Do you want to test this code? You can flash it directly from the Betaflight App:
WARNING: It may be unstable. Use only for testing! |
Contributor
WalkthroughSTM32H5 ADC initialization now stores per-device GPDMA nodes and queues, builds an ADC-to-buffer transfer node, enables circular linked-list mode, and attaches the queue through the HAL linked-list APIs. ChangesSTM32H5 ADC DMA
Estimated code review effort: 3 (Moderate) | ~20 minutes Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
This was referenced Jul 14, 2026
blckmn
approved these changes
Jul 14, 2026
sugaarK
approved these changes
Jul 15, 2026
Member
Author
|
@coderabbitai full review - for reference see also datasheet and reference manual |
Contributor
|
✅ Action performedFull review finished. |
haslinghuis
force-pushed
the
fix/stm32h5-gpdma-adc-ledstrip
branch
from
July 15, 2026 18:32
9634951 to
b7b2e3c
Compare
nerdCopter
approved these changes
Jul 16, 2026
On STM32H5 the ADC DMA channel was configured as a one-shot linear transfer (DMA_NORMAL). H5's GPDMA has no plain circular mode, so HAL_ADC_Start_DMA() took its one-shot branch and transferred a single scan: the conversion buffer was filled once at boot and then frozen forever, so VBAT/current/RSSI never tracked the real inputs. Rebuild the ADC DMA as a circular linked-list queue (HAL_DMAEx_List_* with DMA_LINKEDLIST_CIRCULAR, one static node/queue per ADC device) so HAL_ADC_Start_DMA() takes the linked-list branch and continuously refreshes the buffer. The conversion buffer (destination) is allocated to GPDMA master PORT1: both ports reach all memories on H5, PORT1 just has the zero-latency fast bus multiplexer to SRAM (RM0481 fig. 1), matching the DShot/WS2811 port allocation.
haslinghuis
force-pushed
the
fix/stm32h5-gpdma-adc-ledstrip
branch
from
July 16, 2026 15:16
b7b2e3c to
94297e0
Compare
This was referenced Jul 16, 2026
gwlim
pushed a commit
to gwlim/betaflight
that referenced
this pull request
Jul 25, 2026
…aflight#15419) fix(STM32H5): ADC circular GPDMA so battery/current track live values On STM32H5 the ADC DMA channel was configured as a one-shot linear transfer (DMA_NORMAL). H5's GPDMA has no plain circular mode, so HAL_ADC_Start_DMA() took its one-shot branch and transferred a single scan: the conversion buffer was filled once at boot and then frozen forever, so VBAT/current/RSSI never tracked the real inputs. Rebuild the ADC DMA as a circular linked-list queue (HAL_DMAEx_List_* with DMA_LINKEDLIST_CIRCULAR, one static node/queue per ADC device) so HAL_ADC_Start_DMA() takes the linked-list branch and continuously refreshes the buffer. The conversion buffer (destination) is allocated to GPDMA master PORT1: both ports reach all memories on H5, PORT1 just has the zero-latency fast bus multiplexer to SRAM (RM0481 fig. 1), matching the DShot/WS2811 port allocation.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Scope change: the WS2811 LED-strip fix has been split out to #15438 — a deep review against RM0481 found an additional bug there (TIM handle state latch-up freezing the strip after the first frame) that needs an on-hardware retest, while this ADC fix is already hardware-verified. This PR now carries only the ADC change. A comment-only follow-up for the DShot driver's GPDMA rationale is #15439.
On STM32H5 the ADC DMA channel was configured as a one-shot linear transfer (
DMA_NORMAL). H5's GPDMA has no plain circular mode, soHAL_ADC_Start_DMA()took its one-shot branch and transferred a single scan: the conversion buffer was filled once at boot and then frozen forever, so VBAT/current/RSSI never tracked the real inputs.Rebuild the ADC DMA as a circular linked-list queue (
HAL_DMAEx_List_*withDMA_LINKEDLIST_CIRCULAR, one static node/queue per ADC device) soHAL_ADC_Start_DMA()takes the linked-list branch and continuously refreshes the buffer.The conversion buffer (destination) is allocated to GPDMA master PORT1. Note the rationale was corrected during review: both GPDMA master ports reach all memories on H5 (RM0481 §2.1.5, fig. 1) — PORT1 simply has the zero-latency fast bus multiplexer to SRAM, matching the DShot/WS2811 port allocation. The linked-list construction was verified against the vendored H5 HAL:
HAL_ADC_Start_DMA()overwrites the head node's source/destination/length (converting samples to bytes itself), and the circular CLLR update mask refetches CTR1/CTR2/CBR1/CSAR/CDAR on every loop, so the buffer genuinely refreshes each scan.Verified working on hardware: VBAT/current/RSSI track live values.
make STM32H563builds clean with zero warnings.