stm32f4Cube integration with Ride 7?

[mike2399]

Posted on June 07, 2014 at 13:18

I use the Ride 7 IDE, and have just updated it so I can migrate from STM32f1 to STM32f4 - but cannot find how to integrate Cube into the IDE.  It is supposed to be compatible but I can find no installation instructions.  Please help!

[Nickname306_O]

Posted on June 27, 2014 at 11:49

Hi Mike,

The

STM32CubeF4 FW Package

support EWARM,MDK-ARM and TrueSTUDIO toolchains. Which example you need to port to Ride7? Best Regards, Nouha

[mike2399]

Posted on October 27, 2014 at 11:31

Hi, sorry about the delay in responding!

What I really need to do is build my own application on top of the HAL/standard peripheral libraries.  I already use the RIDE7 IDE for all my previous STM32 development and don't want to change. As for starting with an example project - the one I would probably wish to start with, and then modify to insert my own functions would be one of the LwIP demonstrators.  I currently use a mix of ''httpwebserver'' and ''client/server'' elements in my project.  Although current requirement for STM32F4xx is just to use the various I2S and TDM ports to move audio samples around.

[Tesla DeLorean]

Posted on October 27, 2014 at 11:39

A more generic GNU/GCC+MAKE output from Cube has been suggested to ST at several levels.

I found the RIDE code/examples in the SPL to be the cleanest port to generic GNU/GCC. You could perhaps tweak the Atollic compatible output, likely replacing the start up code and linker scripts with something more RIDE appropriate.

[mike2399]

Posted on October 27, 2014 at 12:04

I have to say I would be happy to use the SPL, as I don't have time to get familiar with the Cube-generated HAL.  Last time I looked for SPL for STM32F427 I couldn't find one?

Otherwise, am I right that you think setting the Cube for Atollic IDE will give the best results?  Is it then just the linker script and startup files (presumably in the CMSIS folder?) the need changing? Many thanks,

[frankmeyer9]

Posted on October 27, 2014 at 12:58

I have to say I would be happy to use the SPL, as I don't have time to get familiar with the Cube-generated HAL.  Last time I looked for SPL for STM32F427 I couldn't find one?

Neither have I found one for the F334, the MCU on one of the F3-Nucleo boards. I have the strong suspicion ST does not plan to release SPL's for new MCUs, in stark contrast to statements made by an ST employee in another thread here. Apart from the obvious immaturity of the CubeMX software (just count bug reports and help requests), I'm not sure what audience ST targets with Cube. To paraphrase from Murphy's Laws here: ''A software written to be simple and fool-proof will only be used by simpletons and fools.''

[Tesla DeLorean]

Posted on October 27, 2014 at 13:14

Doesn't the DSP library version of the SPL cover the F427?

http://www.st.com/web/catalog/tools/FM147/CL1794/SC961/SS1743/PF257901

[Tesla DeLorean]

Posted on October 27, 2014 at 14:15

''I have the strong suspicion ST does not plan to release SPL's for new MCUs, in stark contrast to statements made by an ST employee in another thread here.''

Indeed. I haven't seen anything for the L0, and I'm looking for competitive Cortex-M7 solutions on the assumption there will be no SPL or FPU-D. The reason the SPL has been so popular/successful is that it is thin abstraction that strikes the right balance. It hides a lot of the register level stuff, while not trying to do too much, or force a particular implementation model. I can, and have done, register level code, I chose not too because it is an inefficient use of developer time, and makes things unnecessarily non-portable. The burden of moving between STR7, STR9 and assorted STM32 has been pretty low. The SPL doesn't need a lot of maintenance because it has a very low defect density, and has so since inception.

[mike2399]

Posted on October 27, 2014 at 18:03

Yes, it does, thanks.  I have now found the copy I downloaded a few months ago - but it is on ST's website, you just have to use a fairly specific search term to find it!

[Nickname306_O]

Posted on October 31, 2014 at 18:49

Hi Mike,

Please find below  in

red 

my answer : Otherwise, am I right that you think setting the Cube for Atollic IDE will give the best results? Is it then just the linker script and startup files (presumably in the CMSIS folder?) the need changing?

Ride7 and TrueSTUDIO toolchains are both GNU compilers. They use the same Linker and startup files .

Only one change should be done : comment out ''bl __libc_init_array'' from ride7 startupfile, because it doesn't call __libc_init_array.  Best Regards, Nouha