stm32 spi driver adds the frame format

[FRASTM]

this PR includes the support of the Format (TI vs MOTOROLA) of the SPI frame
The spi frame format is a property added to the spi device through an overlay for the led_strip sample on ws2812
$ west build -p auto -b nucleo_l476rg samples/drivers/led_ws2812/

After the new #39990

fixes #38683

Signed-off-by: Francois Ramu [email protected]

[FRASTM]

stm32F1xx and some of the stm32L1 MCUs do not support the frame format.

[FRASTM]

includes overlay for nucleo_g071rb board

[erwango]

otherwise LGTM

[erwango]

Rather than silently converting property to 0 for these cases, it should rather be notified at binding level,
and then generate an error if an unsupported case is configured

[FRASTM]

#else
--> raising a compilation error 

[ABOSTM]

LGTM

[FRASTM]

@tbursztyka
Since the PR #39990, the frame_format property is missing in the spi_config structure as follows:
.frame_format = DT_PROP(node_id, frame_format),
However there is a bit [15] in the .operation bitfield
* format [ 15 ] - Motorola or TI frame format (optional).
but it seems not immediately compatible with the frame-format as described in the dts/bindings/spi/spi-device.yaml

[FRASTM]

rebase on f882d43

Some stm32L1x mcus (cat 1 or 2) do not have the LL function to set the frame format because they do not have the SPI TI mode (their SPI Control reg 2 bit FRF remains at reset value)

[tbursztyka]

@tbursztyka Since the PR #39990, the frame_format property is missing in the spi_config structure as follows: .frame_format = DT_PROP(node_id, frame_format), However there is a bit [15] in the .operation bitfield * format [ 15 ] - Motorola or TI frame format (optional). but it seems not immediately compatible with the frame-format as described in the dts/bindings/spi/spi-device.yaml

there isn't a property match between dts and spi_config. in dts it's frame_format, and in in spi_config.operation it's a bit there. Where is the issue?

Let me check the bit values, they must be the same (15th bit set or not)
[edit] they seem to match for me. spi_config.operation |= DT_PROP(node_id, frame_format) should be the same as spi_config.operation |= <SPI_FRAME_FORMAT_MOTOROLA/SPI_FRAME_FORMAT_TI>

[tbursztyka]

@FRASTM looks like I forgot to update SPI_CONFIG_DT() macro. see pr #41641

[FRASTM]

rebase
get the PR #41641
--> the stm32 drivers takes the frame_format from the DTS property : operation bitfield

[FRASTM]

rebase

• changed to have else flagged by #if defined(LL_SPI_PROTOCOL_MOTOROLA) && defined(SPI_CR2_FRF)

[FRASTM]

rebase on bb00120

[codecov-commenter]

Codecov Report

Merging #40860 (bb00120) into main (bb00120) will not change coverage.
The diff coverage is n/a.

❗ Current head bb00120 differs from pull request most recent head b6efb3f. Consider uploading reports for the commit b6efb3f to get more accurate results

@@           Coverage Diff           @@
##             main   #40860   +/-   ##
=======================================
  Coverage   51.45%   51.45%           
=======================================
  Files         616      616           
  Lines       72357    72357           
  Branches    16651    16651           
=======================================
  Hits        37234    37234           
  Misses      29003    29003           
  Partials     6120     6120           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bb00120...b6efb3f. Read the comment docs.