[Arm-netbook] STM32F blinky lights

[joem]

On Thu, 2013-06-13 at 10:10 +0100, luke.leighton wrote:
> On Thu, Jun 13, 2013 at 9:17 AM, joem <joem at martindale-electric.co.uk> wrote:
> > On Wed, 2013-06-12 at 22:31 +0100, luke.leighton wrote:
> >> well amazing i am so happy to have managed to make an LED flash on the
> >> little waveshare STM32F103RCT6.  it is such an achievement.  no,
> >> really: a blinking light means a working compiler, working upload
> >> mechanism over the serial port (stm32sprog), and picking the correct
> >> GPIO location to make the flashy-blinkey thing go wiggle, wiggle.
> >
> > 100% correct thing to do and a big achievement
> > if non-techies are reading this!!
> >
> > Interesting CPU.
> >
> > Care to share what is the IDE
> 
>  IDE?  IDE??? what do you take me for - a muppet?? :)  naah mate - the
> only reason i can't bludgeon or nuke IDEs is because they're just
> programs.
> 
> > and compiler,
> 
>  arm-none-eabi-gcc - i got it frooomm... errr....  ah!  here it is:
>  https://pixhawk.ethz.ch/px4/dev/toolchain_installation_lin
> 
> 
> > and programmer for this?
> 
>  stm32sprog - worked a treat, just needed to short boot0, apply power,
> and run it:
>     http://code.google.com/p/stm32sprog/

Thank you very much, I go and investigate.


> > And is there a Linux version?
> 
>  indeed there is - would i use anything else? :)
> 
> > I searched waveshare http://www.wvshare.com and could not find any
> > pointers.
> 
>  naah they're going with the ST board support software.  it's all
> pre-prepared as shite proprietary compiler project structure so i made
> a beeline for libopencm3.

That seems very interesting.
http://libopencm3.org/wiki/Main_Page
But I still believe it should be ARM that should supply
the main header files for the CPUs it releases giving names
to registers and bit fields.
The implementators of ARM should then customize it with
physical locations and actual bit field values to fill
in the details specific to their implementation.
That way the names for all the bit fields and registers
are consistent for every programmer that uses the header files.
At the moment, there is no such thing. Libopencm3 competes with
the half baked arm CMSIS library:
http://www.arm.com/products/processors/cortex-m/cortex-microcontroller-software-interface-standard.php
(The lib that goes around naming an RS232 port
 UART0 in Cortex M0 and USART0 in Cortex M3 - its a complete guide 
 software idiocy - They even managed to keep a straight face and
 put this on there -

 "Creation of software is a major cost factor in the embedded industry.
By standardizing the software interfaces across all Cortex-M silicon
vendor products, especially when creating new projects or migrating
existing software to a new device, means significant cost reductions"
 
 - sheeessh!!!)

> > So far, I got me LPC11xx series, LPC13xx, and LPC17xx NXP's cortex 0 and
> > 3 series working with Code Red http://www.code-red-tech.com/ IDE
> > which is derivative of Eclipse and the cheap LPCexpresso
> > http://uk.farnell.com/nxp/om13012-598/eval-lpcxpresso-lpc11c2x/dp/2251821
> > programmer (which is one half of the board that can be broken off).
> > The compiler software is free up to code size of 128k and the
> > lpcexpresso programmer is only about $20.
> 
>  why? :) 

The ones I use don't have any boot options - have to use SWD method or
JTAG method to program it.


>  i used a USB-to-Serial port which was 8 quid and that's
> quite a lot for a USB-to-serial port, and all the software is entirely
> software (libre)
> 
> > If you remember, I already do a SoM board for 2 x LPC1764 chip
> > http://uk.farnell.com/nxp/lpc1764fbd100/mcu-32bit-arm-cortex-m3-100lqfp/dp/1718546
> > The chip is 100 pin and that means the SoM board has around 180 GPIO
> > pins available for use. The KiCAD files are free:
> > http://www.gplsquared.com/SoM1/SoM1.html
> > And it is built and working.
> 
>  oo i do have to look at alternatives to the STM32F, but pricing needs
> to be competitive (around $2.40)

They were competitive - $0.6 not so long ago for the cheapest options.
Now everything seems to have doubled, tripled and gone further up.
I'm thinking of evaluating Nuvoton chips, but coocox that make software
is so anti-Linux support despite porting open source gcc and eclipse to
run on windows. http://www.coocox.org/index.html
Resistance is slowly changing from reading the forums there.