stm32/uart: Add support for LPUART1 on L0, L4, H7 and WB MCUs. #5396

chrismas9 · 2019-12-08T04:26:30Z

Add LPUART1 as a standard UART. No low power features are supported.
LPUART1 is enabled as the next available UART after the standard U(S)ARTs.
STM32WB: LPUART1 = UART(2)
STM32L0: LPUART1 = UART(6)
STM32L4: LPUART1 = UART(6)
STM32H7: LPUART1 = UART(9)

LPUART1 is enabled by defining MICROPY_HW_LPUART1_TX and MICROPY_HW_LPUART1_RX in mpconfigboard.h

chrismas9 · 2019-12-08T04:48:51Z

Motivation for this PR was changing from F030 to L073 to get EEPROM. L073 does not have USART3, but has LPUART1 on same pin (PB10/11). It also adds a user UART to NUCLEO_WB55 which uses its only USART for REPL. Gives more pin choices for UART when normal UARTS unavailable due to other use of pins for I2C, ADC, Eth, etc.

Testing: LPUART1 as UART(6) tested with external loopbak and baud rate check on NUCLEO_L073RZ and NUCLEO_L452RE. Flow control not tested. I don't have hardware to test WB55 or H7. Checked UASART6 as UART(6) on PYBV3. NUCLEO_WB55 and NUCLEO_H743ZI ports compile with LPUART1 enabled. PYBV3 and NUCLEO_F767ZI compile without error. Review of all register differences between U(S)ART and LPUART. Code inspection of all direct access to STM32 registers. Handle difference in buadrate (BRR register). There are several interrupt enables in CR1 and CR2 not implemented in LPUART. These must be kept at reset state (0). They are defined, but only cleared, so have not changed the code.

chrismas9 · 2019-12-08T04:52:06Z

I have another PR coming to reduce size of L073 port. Then I will update mpconfigport.h for various boards to enable the extra UARTs - LPUART1 and UASART4/5 on L0.

dpgeorge · 2019-12-09T06:06:52Z

Thanks for this.

LPUART1 is enabled as the next available UART after the standard U(S)ARTs.

I'm not sure this is the right approach because it could get confusing with what the last normal UART number is (the STM32's have a big range in available UARTs).

I don't think it makes sense to have a machine.LPUART class because it's very stm32-specific. Maybe something like machine.UART('LP1') to select the low-power peripherals? Also need to think how the low-power functionality would be exposed, possibly just additional keyword args to .init().

chrismas9 · 2019-12-09T15:15:41Z

I don't think it makes sense to have a machine.LPUART class

I agree, I didn't want to make any changes to the UART class, that's why I did it this way.

Maybe something like machine.UART('LP1')

Good idea. Done

Also need to think how the low-power functionality would be exposed, possibly just additional keyword args to .init().

I think the only low power feature that would be useful is to set the clock to 32kHz (9600 buad max) and enable wake on interrupt. I don't have a need of this and I don't know how popular it would be. I am not sure I would be confident implementing this.

H-Grobben · 2021-02-01T07:56:42Z

ports/stm32/machine_uart.c

-        mp_printf(print, "UART(%u)", self->uart_id);
+        #ifdef LPUART1
+        if (self->uart_id == PYB_LPUART_1) {
+            mp_printf(print, "UART('MP1')");


Shouldn't this be LP1, or LPUART1?

I think it should be 'LP1", also line 182. I will review, rebase and retest this week

H-Grobben · 2021-02-01T10:49:56Z

I will merge this with my G4 setup ( #6808 ), and also with our custom WB board, and try testing later this week.

dpgeorge · 2021-02-01T12:25:48Z

I still think naming the UART with "LP1" and "LP2" etc is the right approach (rather than numbering them after all other UARTs).

@christmas9

… micopython specifies MICROPY_HW_MAX_UART in mpconfigboard_common.h (is max also max number of uarts, or is it related to internal UART numbering (e.g. [UART1, UART2, UART3, UART6] -> max = 6?) All credits go to @christmas9 for pull request micropython#5396

H-Grobben · 2021-02-01T21:05:35Z

Tested succesfully it in our repo, only for G4 series Nucleo board:
AEMICS@ba929ae
(and some more commits for the checks for code formatting etc)
With this pull request, we don't need the additional wires on the Nucleo board!

chrismas9 · 2021-02-02T07:14:30Z

I still think naming the UART with "LP1" and "LP2" etc is the right approach (rather than numbering them after all other UARTs).

Yes, I agree now, especially if low power features are added later, in which case 'LP1', etc identify that these UARTs are different. The UARTs are enumerated in uart.h so the next number is used in the driver with 'LP1' handled as a special case. Implementing LPUART without using the next number would need more driver changes than I could do.

chrismas9 · 2021-02-02T13:00:28Z

Rebased, fixed typo 'MP1' -> 'LP1', updated code formatting

dpgeorge · 2021-02-02T13:16:01Z

ports/stm32/machine_uart.c

+        if (self->uart_id == PYB_LPUART_1) {
+            mp_printf(print, "UART('LP1')");
+        } else {
+        #endif


simpler to use the following pattern:

#ifdef LPUART1 if (...) { } else #endif { mp_printf(...); }

dpgeorge · 2021-02-02T13:21:09Z

ports/stm32/mpconfigboard_common.h

 #define MICROPY_HW_MAX_I2C (4)
 #define MICROPY_HW_MAX_TIMER (17)
-#define MICROPY_HW_MAX_UART (8)
+#define MICROPY_HW_MAX_UART (9)


might be clearer to do something like this:

#define MICROPY_HW_MAX_LPUART (1) #define MICROPY_HW_MAX_UART (8)

(ie last line there unchanged), and then in mpconfigport.h:

struct _pyb_uart_obj_t *pyb_uart_obj_all[MICROPY_HW_MAX_UART + MICROPY_HW_MAX_LPUART]; \

dpgeorge · 2021-02-02T13:22:05Z

ports/stm32/uart.h

+    PYB_LPUART_1 = 9,
+    #else
    PYB_UART_9 = 9,
+    #endif


all this could be simplified to:

#ifdef LPUART1 PYB_LPUART_1 = MICROPY_HW_MAX_UART, #endif

chrismas9 · 2021-02-02T13:37:43Z

Thanks for the review and suggestions. Once this is cleaned up I can add LPUART1 to mpconfigboard.h for some boards and update the docs.

dpgeorge · 2021-02-16T14:08:57Z

@chrismas9 I'd like to merge this before the STM32G4 addition. If you can't get to updating the PR just let me know and I can do it.

H-Grobben · 2021-02-16T16:08:01Z

I've added this into the STM32G4 (#6911), but it would also apply to other families of the STM32, just didn't test it, as I don't have the hardware to test it. I did build all STM32 families (ci script nucleo and pyb boards)

Add LPUART1 as a standard UART. No low power features are supported. LPUART1 is enabled as the next available UART after the standard U(S)ARTs. STM32WB: LPUART1 = UART(2) STM32L0: LPUART1 = UART(6) STM32L4: LPUART1 = UART(6) STM32H7: LPUART1 = UART(9) or all ports: LPUART1 = UART('LP1') LPUART1 is enabled by defining MICROPY_HW_LPUART1_TX and MICROPY_HW_LPUART1_RX in mpconfigboard.h Signed-off-by: Chris Mason <c.mason@inchipdesign.com.au>

dpgeorge · 2021-02-21T04:54:21Z

Thanks for updating the PR. Merged in 9d674cf with some minor changes:

fixed so it raises an exception creating an LP1 UART (without specifying a baudrate or any other params) if the TX/RX pins are not defined
reverted stray changes to some config values in mpconfigboard_common.h
made AF_FN_LPUART distinct from AF_FN_UART to avoid changing mp_hal_pin_config_alt

I also enabled LPUART1 on NUCLEO_WB55 in a separate commit, 03a64f2

Fix a critical "typo"

dpgeorge added the port-stm32 label Dec 9, 2019

chrismas9 force-pushed the LPUART branch from 21a36df to a42c5a2 Compare December 9, 2019 15:04

chrismas9 force-pushed the LPUART branch from a42c5a2 to e47691b Compare December 9, 2019 15:35

H-Grobben reviewed Feb 1, 2021

View reviewed changes

dpgeorge mentioned this pull request Feb 1, 2021

STM32G4 series added #6808

Closed

chrismas9 force-pushed the LPUART branch 2 times, most recently from acd1f46 to d086bba Compare February 2, 2021 12:43

dpgeorge reviewed Feb 2, 2021

View reviewed changes

H-Grobben mentioned this pull request Feb 16, 2021

STM32G4: add support for G4 series, and LPUART #6911

Closed

chrismas9 force-pushed the LPUART branch from d086bba to 64b31b4 Compare February 17, 2021 05:12

chrismas9 force-pushed the LPUART branch from 64b31b4 to 82281e4 Compare February 17, 2021 09:27

dpgeorge closed this Feb 21, 2021

chrismas9 deleted the LPUART branch February 24, 2021 14:06

tannewt pushed a commit to tannewt/circuitpython that referenced this pull request Oct 13, 2021

Merge pull request micropython#5396 from Fe7n/patch-1

5cbfac6

Fix a critical "typo"

Uh oh!

stm32/uart: Add support for LPUART1 on L0, L4, H7 and WB MCUs. #5396

stm32/uart: Add support for LPUART1 on L0, L4, H7 and WB MCUs. #5396

Uh oh!

Conversation

chrismas9 commented Dec 8, 2019

Uh oh!

chrismas9 commented Dec 8, 2019

Uh oh!

chrismas9 commented Dec 8, 2019

Uh oh!

dpgeorge commented Dec 9, 2019

Uh oh!

chrismas9 commented Dec 9, 2019

Uh oh!

H-Grobben Feb 1, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

chrismas9 Feb 1, 2021

Choose a reason for hiding this comment

Uh oh!

H-Grobben commented Feb 1, 2021

Uh oh!

dpgeorge commented Feb 1, 2021

Uh oh!

H-Grobben commented Feb 1, 2021

Uh oh!

chrismas9 commented Feb 2, 2021

Uh oh!

chrismas9 commented Feb 2, 2021

Uh oh!

dpgeorge Feb 2, 2021

Choose a reason for hiding this comment

Uh oh!

dpgeorge Feb 2, 2021

Choose a reason for hiding this comment

Uh oh!

dpgeorge Feb 2, 2021

Choose a reason for hiding this comment

Uh oh!

chrismas9 commented Feb 2, 2021

Uh oh!

dpgeorge commented Feb 16, 2021

Uh oh!

H-Grobben commented Feb 16, 2021

Uh oh!

dpgeorge commented Feb 21, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

H-Grobben Feb 1, 2021 •

edited

Loading