cancel
Showing results for 
Search instead for 
Did you mean: 

Open source tools for the STM32

lanchon
Associate III
Posted on September 19, 2009 at 01:43

Open source tools for the STM32

66 REPLIES 66
mcatudal
Associate II
Posted on May 17, 2011 at 12:17

Wrong on OpenOCD, the maintainer of the Cortex code told me that he has

no plan for it. That is where I come in. I will make the support.

I need to add hardware support for it also as well as the driver.

They call it single wire debug but it is actually 2 or 3 wires. Two wires are allways needed, one for data and one for the clock. The third wire is optional and is the trace wire.

It can be very valuable if you are running out of pins and don't want to move to a bigger package.

sjo
Associate II
Posted on May 17, 2011 at 12:17

Magnus may have no plans, but as i am also a developer on openocd i have plans.

The first thing we are working on is a new jtag dongle, probably using the stm32. The ftdi type device are too limited.

I am fully aware of what the swd is, i have been developing with it for about a year.

Cheers

sjo

sudazqb
Associate
Posted on May 17, 2011 at 12:17

I am using CodeSourcery now, and I have use it to build several program.

CodeSourcery is nice for me, for I do not need a IDE, debugler. So I just use to build the project and then use ST's ISP software to download the binary and then run.

I have managed to generate a correct makefile for the library and the application, so does the .ld file. (but I am not sure it is absolutly right).

My first demo using codesourcery is at here(a lcd demo of EK-STM32F):

http://www.ouravr.com/bbs/bbs_content.jsp?bbs_sn=874394

the second:(I use a VS1003 chip, and migrate my mp3 player(with lyric) to STM32)

http://bbs.21ic.com/club/bbs/bbsview.asp?essenceID=9020

Or see the attachment:

Sorry for that the web is in Chinese, and sorry for my poor English.

Hope this can be useful to you.

mcatudal
Associate II
Posted on May 17, 2011 at 12:17

Interesting to know that I am not the only one working on SW debugging.

I do agree that bit banging from the PC has its issues.

I am also working on a new approach for the dongle.

planeetmaa
Associate II
Posted on May 17, 2011 at 12:17

Hi!

About that new jtag dongle sjo mentioned. Are you going to release it as open source or planning so sell those? What is the schedule for new dongle? Im just interested about cortex and that 1 wire dubgging thing. But just decided to wait until tools are ready. Of course I can use that ''normal'' jtag debugging, but in case switching to cortex, I would like to use that 1 serial wire stuff as there are always few pins short anyway...

Madis

lanchon
Associate III
Posted on May 17, 2011 at 12:17

@sjo,

@mcatudal,

Hi, is there a place where you're discussing:

-problems with FTDI JTAG dongles,

-objectives of the new dongle,

-firmware and software interface of the new dongle?

@sjo,

Do you think there ar
lanchon
Associate III
Posted on May 17, 2011 at 12:17

@sudazqb,

if you're unsure about your linker file, see:

http://www.st.com/mcu/forums-cat-6304-23.html

cheers!

pandoraems
Associate II
Posted on May 17, 2011 at 12:17

Sorry to bring up old thread, but..

is there any info/update on the SWD plans for openOCD with a stm32 based dongle?

Is there an ongoing open effort on this?

Cheers

sjo
Associate II
Posted on May 17, 2011 at 12:17

currently no update, not enough time - volunteers welcome however.

Cheers

Spen

andreas2
Associate II
Posted on May 17, 2011 at 12:17

I spent the weekend trying to adapt OpenOCD to talk SWD through a FT2232 based dongle (actually a JTAGkey with an external tristate buffer). I managed to get it working to the amount i had time testing. I think there are some reset/halting issues, since it seems neccessary to do a soft_reset_halt before using commands that upload algorithms. But otherwise it works just as well as with JTAG. And I can boost the clock higher than in JTAG mode before getting overruns, don't know why, maybe higher protocol overhead. I'm getting something like 30kB/s upload to SRAM with 3MHz clock.

The main problem is that it's a rather invasive patch to OpenOCD, and I'm not sure it still works in JTAG mode. It was kinda hard to fit the protocol to the existing structure and it could probably be much better integrated on another level. I will look into it the coming days if/when I get time.