[STM32F0] Add asynchronous serial #2415

svastm · 2016-08-10T15:44:38Z

This PR add the asynchronous serial for the following targets:

NUCLEO_F070RB
NUCLEO_F072RB
NUCLEO_F091RC

The struct serial_s move to common_objects.h.

I enable only this ones because the utest require at least 3 uarts to test the implementation. Should we enable non tested (automatically) targets ?

test report

For some reasons the cpputest is failing on IAR and GCC for at least NUCLEO_F070RB and NUCLEO_F072RB.

0xc0170 · 2016-08-11T11:48:59Z

For some reasons the cpputest is failing on IAR and GCC for at least NUCLEO_F070RB and NUCLEO_F072RB.

Same as in the other pull request. Please have a look at failures.

I enable only this ones because the utest require at least 3 uarts to test the implementation.

3 ? you mean stdio + 2 uart (tx to rx) to test?

svastm · 2016-08-16T12:05:38Z

Yes by 3 uart, I mean 2 + stdio.

0xc0170 · 2016-08-24T14:16:44Z

/morph test

0xc0170 · 2016-08-24T14:19:42Z

For some reasons the cpputest is failing on IAR and GCC for at least NUCLEO_F070RB and NUCLEO_F072RB.

What are those failures?

mbed-bot · 2016-08-24T16:18:20Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 699

Build failed!

bridadan · 2016-08-24T16:25:21Z

The build failed for NUCLEO_F070RB and NUCLEO_F072RB for the IAR compiler, looks like some of the tests don't fit into memory

sg- · 2016-08-26T16:57:29Z

@svastm please rebase

svastm · 2016-08-29T08:24:27Z

rebased

sg- · 2016-09-08T22:02:53Z

@mbed-bot: TEST

HOST_OSES=ALL
BUILD_TOOLCHAINS=ALL
TARGETS=ALL

sg- · 2016-09-08T22:03:03Z

/morph test

mbed-bot · 2016-09-08T23:01:13Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 786

Build failed!

mbed-bot · 2016-09-09T01:33:07Z

[Build ${MBED_BUILD_ID}]
FAILURE: Something went wrong when building and testing.

sg- · 2016-09-09T13:22:13Z

Needs further investigation as to why this change causes tests to not fit in the memory of the devices
TESTS-MBED_DRIVERS-C_STRINGS for NUCLEO_F070RB and NUCLEO_F072RB

svastm · 2016-09-09T15:31:36Z

The NUCLEO_F070RB have 16192 Bytes in the RAM section.

without this PR, the build of test-mbed_drivers-c_strings give:

[DEBUG] Output:    IAR ELF Linker V7.40.3.8902/W32 for ARM
[DEBUG] Output:    Copyright 2007-2015 IAR Systems AB.
[DEBUG] Output:
[DEBUG] Output:   35 678 bytes of readonly  code memory
[DEBUG] Output:    1 267 bytes of readonly  data memory
[DEBUG] Output:   15 782 bytes of readwrite data memory
[DEBUG] Output:
[DEBUG] Output: Errors: none
[DEBUG] Output: Warnings: none
[DEBUG] Output:

With the PR

[DEBUG] Output:    IAR ELF Linker V7.40.3.8902/W32 for ARM
[DEBUG] Output:    Copyright 2007-2015 IAR Systems AB.
[DEBUG] Output:
[DEBUG] Output:   35 950 bytes of readonly  code memory
[DEBUG] Output:    1 176 bytes of readonly  data memory
[DEBUG] Output:   16 274 bytes of readwrite data memory
[DEBUG] Output:
[DEBUG] Output: Errors: 1
[DEBUG] Output: Warnings: none
[DEBUG] Output:
[DEBUG] Output: Link time:   0.25 (CPU)   0.26 (elapsed)
[DEBUG] Errors: Error[Lp011]: section placement failed
[DEBUG] Errors:           unable to allocate space for sections/blocks with a total estimated
[DEBUG] Errors:                     minimum size of 0x3f98 bytes (max align 0x8) in
[DEBUG] Errors:                     <[0x200000c0-0x20003fff]> (total uncommitted space 0x3f40).

So the readwrite data memory increased by 492 bytes.

I did not do the math precisely but I found:

changed a unique global UART_HandleTypeDef to an array of 4. (+ ~360 bytes)
increase serial_t size (+ ~32 bytes)

The number does not look huge to me.

0xc0170 · 2016-09-12T13:59:40Z

We could optimize here?

From the diff:

+struct serial_s {
 +    UARTName uart;
 +    int index; // Used by irq
 +    uint32_t baudrate;
 +    uint32_t databits;
 +    uint32_t stopbits;
 +    uint32_t parity;
 +    PinName pin_tx;
 +    PinName pin_rx;
 +#if DEVICE_SERIAL_ASYNCH
 +    uint32_t events;
 +#endif
 +#if DEVICE_SERIAL_FC
 +    uint32_t hw_flow_ctl;
 +    PinName pin_rts;
 +    PinName pin_cts;
 +#endif
 +}

index could be uint8_t. Baudrate could be uint16_t. databits could be uint8_t, same for stopbits and parity ?

svastm · 2016-09-14T08:57:25Z

With your suggestion I get this:

[DEBUG] Output:    IAR ELF Linker V7.40.3.8902/W32 for ARM
[DEBUG] Output:    Copyright 2007-2015 IAR Systems AB.
[DEBUG] Output:
[DEBUG] Output:   35 734 bytes of readonly  code memory
[DEBUG] Output:    1 108 bytes of readonly  data memory
[DEBUG] Output:   16 250 bytes of readwrite data memory

So we save 24 bytes. We could look for more static RAM optimizations but I think we will stay close to the limit. Maybe we should reduce heap and stack size.

For now, heap take 1/4 of RAM and stack 1/8.

0xc0170 · 2016-09-14T12:49:07Z

Maybe we should reduce heap and stack size.

How much are those 2 now for this target?

Can you please resolve conflicts?

svastm · 2016-09-20T11:23:26Z

Rebased.
With the #2745 which come just at the right time. The size of the c_strings test decrease to

36 174 bytes of readonly  code memory
   1 331 bytes of readonly  data memory
  15 824 bytes of readwrite data memory

So there is 368 bytes left for an other feature upgrade. Should I still reduce heap and stack for IAR ?

0xc0170 · 2016-09-20T13:55:37Z

Should I still reduce heap and stack for IAR ?

We could answer if you share how much is currently allocated for this target.

svastm · 2016-09-20T14:27:01Z

The reserved size is

heap:  0x1000 -> 4096 bytes
stack: 0x0800 -> 2048 bytes

0xc0170 · 2016-09-22T08:48:16Z

368 bytes is not much, shall we reduce at least heap?

svastm · 2016-09-22T13:04:26Z

I'm a bit confuse. I may have test on the wrong branch. I did another test with the PR merged and

mbed test -m NUCLEO_F070RB -t IAR -n mbed-os-tests-mbed_drivers-c_strings -v
...
[DEBUG] Output:    IAR ELF Linker V7.60.1.11101/W32 for ARM
[DEBUG] Output:    Copyright 2007-2016 IAR Systems AB.
[DEBUG] Output:
[DEBUG] Output:   35 922 bytes of readonly  code memory
[DEBUG] Output:    1 254 bytes of readonly  data memory
[DEBUG] Output:   13 772 bytes of readwrite data memory
[DEBUG] Output:

2420 bytes free 😂

sg- · 2016-09-22T13:53:17Z

@mbed-bot: TEST

HOST_OSES=ALL
BUILD_TOOLCHAINS=ALL
TARGETS=ALL

sg- · 2016-09-22T13:55:17Z

/morph test

mbed-bot · 2016-09-22T21:00:57Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 927

All builds and test passed!

0xc0170

Code LGTM

mbed-bot · 2016-09-23T18:19:24Z

[Build 971]
SUCCESS: Building succeeded and tests were run! Be sure to check the test results

Ports for Upcoming Targets 2669: Added u-blox C029 target ARMmbed/mbed-os#2669 2707: [EFM32] Add IAR support for remaining Silicon Labs targets ARMmbed/mbed-os#2707 2819: MultiTech xDot platform support - 09.26.2016 ARMmbed/mbed-os#2819 2827: include MultiTech xDot in mbed 5 releases ARMmbed/mbed-os#2827 Fixes and Changes 2522: Add CThunk for CM7 ARMmbed/mbed-os#2522 2518: Enable uvisor on Beetle ARMmbed/mbed-os#2518 2571: STM32F7 - Add asynchronous serial ARMmbed/mbed-os#2571 2616: STM32F3xx - Add Serial Flow Control pins + enable it ARMmbed/mbed-os#2616 2619: NUCLEO_L152RE - Add Serial Flow Control ARMmbed/mbed-os#2619 2620: NUCLEO_F429ZI - Add SERIAL_FC macro ARMmbed/mbed-os#2620 2666: [EFM32] Microsecond ticker optimization ARMmbed/mbed-os#2666 2681: STM32F0xx - Add support of ADC internal channels ARMmbed/mbed-os#2681 2687: [NRF5] Add fs_data symbol in data secton for gcc ARMmbed/mbed-os#2687 2696: Add device_has to all nrf51 devices ARMmbed/mbed-os#2696 2703: TARGET_NRF5: Changed 'serial_baud' implementation to support special baud rates. ARMmbed/mbed-os#2703 2704: DISCO_L476VG: add SPI nicknames ARMmbed/mbed-os#2704 2723: KSDK serial_api.c: Fix assertion error for ParityEven ARMmbed/mbed-os#2723 2463: [STM32L0] Add asynchronous serial ARMmbed/mbed-os#2463 2572: Fix STM32F407VG target name and LPC11U6X linker errors ARMmbed/mbed-os#2572 2698: DELTA_DFBM_NQ620 target ARMmbed/mbed-os#2698 2542: Dev spi asynch stm32f4 ARMmbed/mbed-os#2542 2650: STM32F3 - Add low power timer ARMmbed/mbed-os#2650 2415: [STM32F0] Add asynchronous serial ARMmbed/mbed-os#2415 2585: Added support for ADC only pins in LPC43xx ARMmbed/mbed-os#2585 2622: [STM32F4] Add asynchronous I2C ARMmbed/mbed-os#2622 2719: Updated ARM linker scripts for Kinetis platforms that use SDK 2.0 ARMmbed/mbed-os#2719 2728: Added ethernet and enabled IPV4 feature for the EVK-ODIN-W2/C029 target ARMmbed/mbed-os#2728 2747: [LPC11U68] Fix pin interrupt select offset ARMmbed/mbed-os#2747 2751: STM32L0xx - Add Serial Flow Control ARMmbed/mbed-os#2751 2753: [NUCLEO_F767ZI] Add CAN capability ARMmbed/mbed-os#2753 2759: STM32F0 - Add low power timer ARMmbed/mbed-os#2759 2763: STM32L1 - Add low power timer ARMmbed/mbed-os#2763 2764: STM32L4 - Add low power timer ARMmbed/mbed-os#2764 2771: STM32L4 - Update deepsleep implementation ARMmbed/mbed-os#2771 2775: Update KSDK SDHC driver for K64F & K66F ARMmbed/mbed-os#2775 2792: [NUCLEO_F303ZE] MBED-OS5 capability ARMmbed/mbed-os#2792 2762: STM32L0 - Add low power timer ARMmbed/mbed-os#2762 2761: STM32F7 - Add low power timer ARMmbed/mbed-os#2761

0xc0170 added needs: work needs: review needs: CI labels Aug 11, 2016

svastm force-pushed the serial_asynch_f0 branch from 14a5705 to 30d41a9 Compare August 11, 2016 12:28

svastm force-pushed the serial_asynch_f0 branch from 30d41a9 to a61c425 Compare August 23, 2016 14:52

0xc0170 removed the needs: work label Aug 24, 2016

svastm force-pushed the serial_asynch_f0 branch from a61c425 to 5357782 Compare August 29, 2016 08:24

sg- removed the needs: review label Sep 8, 2016

sg- added the needs: work label Sep 9, 2016

sg- removed the needs: CI label Sep 16, 2016

svastm added 3 commits September 20, 2016 12:57

[STM32F0] Add asynchronous serial

12210bc

[STM32F0] Enable asynchronous serial

a132415

[STM32F0] Add tests for asynchronous serial

7e3d733

svastm force-pushed the serial_asynch_f0 branch from 5357782 to 7e3d733 Compare September 20, 2016 11:23

sg- added needs: CI and removed needs: work labels Sep 22, 2016

0xc0170 approved these changes Sep 23, 2016

View reviewed changes

sg- added ready for merge and removed needs: CI labels Sep 23, 2016

sg- merged commit 297597c into ARMmbed:master Sep 24, 2016

sg- added release-version: 5.1.5 and removed ready for merge labels Sep 24, 2016

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[STM32F0] Add asynchronous serial #2415

[STM32F0] Add asynchronous serial #2415

svastm commented Aug 10, 2016

0xc0170 commented Aug 11, 2016

svastm commented Aug 16, 2016

0xc0170 commented Aug 24, 2016

0xc0170 commented Aug 24, 2016

mbed-bot commented Aug 24, 2016

bridadan commented Aug 24, 2016

sg- commented Aug 26, 2016

svastm commented Aug 29, 2016

sg- commented Sep 8, 2016

sg- commented Sep 8, 2016

mbed-bot commented Sep 8, 2016

mbed-bot commented Sep 9, 2016

sg- commented Sep 9, 2016

svastm commented Sep 9, 2016 •

edited

Loading

0xc0170 commented Sep 12, 2016

svastm commented Sep 14, 2016 •

edited

Loading

0xc0170 commented Sep 14, 2016

svastm commented Sep 20, 2016

0xc0170 commented Sep 20, 2016

svastm commented Sep 20, 2016 •

edited

Loading

0xc0170 commented Sep 22, 2016

svastm commented Sep 22, 2016

sg- commented Sep 22, 2016

sg- commented Sep 22, 2016

mbed-bot commented Sep 22, 2016

0xc0170 left a comment

mbed-bot commented Sep 23, 2016

[STM32F0] Add asynchronous serial #2415

[STM32F0] Add asynchronous serial #2415

Conversation

svastm commented Aug 10, 2016

0xc0170 commented Aug 11, 2016

svastm commented Aug 16, 2016

0xc0170 commented Aug 24, 2016

0xc0170 commented Aug 24, 2016

mbed-bot commented Aug 24, 2016

Result: FAILURE

Output

bridadan commented Aug 24, 2016

sg- commented Aug 26, 2016

svastm commented Aug 29, 2016

sg- commented Sep 8, 2016

sg- commented Sep 8, 2016

mbed-bot commented Sep 8, 2016

Result: FAILURE

Output

mbed-bot commented Sep 9, 2016

sg- commented Sep 9, 2016

svastm commented Sep 9, 2016 • edited Loading

0xc0170 commented Sep 12, 2016

svastm commented Sep 14, 2016 • edited Loading

0xc0170 commented Sep 14, 2016

svastm commented Sep 20, 2016

0xc0170 commented Sep 20, 2016

svastm commented Sep 20, 2016 • edited Loading

0xc0170 commented Sep 22, 2016

svastm commented Sep 22, 2016

sg- commented Sep 22, 2016

sg- commented Sep 22, 2016

mbed-bot commented Sep 22, 2016

Result: SUCCESS

Output

0xc0170 left a comment

Choose a reason for hiding this comment

mbed-bot commented Sep 23, 2016

svastm commented Sep 9, 2016 •

edited

Loading

svastm commented Sep 14, 2016 •

edited

Loading

svastm commented Sep 20, 2016 •

edited

Loading