Trouble with assembly files in System Work Bench for STM32

[Wesley Ragle]

Posted on March 06, 2017 at 16:54

Hi folks,

I am attempting to generate an archive file with System Work Bench.  All goes well until the make process hits an assembly file.  It then seems that the command line generated by the work bench tries to process the same assembly file twice resulting in a series of redefinition errors.

Following is what is output to the console window:

------------------------------------------------

'Building file: ../os_cpu_a.s'

'Invoking: MCU GCC Assembler'

C:\1_1_A_F7_WORKSPACE\UCOS109\Debug

arm-none-eabi-as -mcpu=cortex-m7 -mthumb -mfloat-abi=hard -mfpu=fpv5-sp-d16 -g -o 'os_cpu_a.o' '../os_cpu_a.s' '../os_cpu_a.s'

../os_cpu_a.s: Assembler messages:

../os_cpu_a.s:46: Error: symbol `OSCtxSw' is already defined

../os_cpu_a.s:62: Error: symbol `OSIntCtxSw' is already defined

../os_cpu_a.s:78: Error: symbol `SaveStatusReg' is already defined

../os_cpu_a.s:93: Error: symbol `RestoreStatusReg' is already defined

../os_cpu_a.s:107: Error: symbol `OSStartHighRdy' is already defined

../os_cpu_a.s:130: Error: symbol `OSStartHang' is already defined

../os_cpu_a.s:145: Error: symbol `PendSV_Handler' is already defined

../os_cpu_a.s:169: Error: symbol `PendSVHandler_nosave' is already defined

../os_cpu_a.s:153: Error: branch out of range

make: *** [os_cpu_a.o] Error 1

---------------------------------------------------

As you can see, '../os_cpu_a.s'  shows up twice in the command line.

Any help would be greatly appreciated.

Thanks,

Wes

[Tesla DeLorean]

Posted on March 06, 2017 at 19:58

Something you'd probably want to take up directly with them?

http://www.openstm32.org/HomePage

 

[Wesley Ragle]

Posted on March 07, 2017 at 00:36

Thanks,

I just figured that out

Wes 

[Gabriel Staples]

Posted on March 22, 2018 at 02:16

Any luck? I'm having this exact same problem. I can't figure out why it's doing it either. System Workbench seems to have a variety of these little 'gotcha' bugs. I hope STM does something great with Atollis TrueSTUDIO that they just bought and gives me a good reason to switch over to that instead. 

[Andrew Neil]

Posted on March 22, 2018 at 09:03

Presumably, the OP went over to the 

http://www.openstm32.org/HomePage

  and found their own, specific forum?

Which goes to show the importance of including links when cross-posting ...

[Gabriel Staples]

Posted on March 23, 2018 at 00:47

I found a solution: my broken build was due to a file startup_stm32f103xb.s which I put in a folder which was part of the C/C++ Includes. Due to an apparent bug in System Workbench, simply moving the .s file into a folder which was NOT in the Includes fixed it. It compiles now! That makes 5 bugs I've found in System Workbench so far, but I don't know if it's worth reporting them as one of them was reported as early as 2016 and nothing has been done about it. 

[Andrew Neil]

Posted on March 23, 2018 at 01:04

having bought Atollic, I guess TrueStudio is where ST will be focussing their attention going forward ... ?

[Tesla DeLorean]

Posted on March 23, 2018 at 01:20

Does TrueStudio include a bug tracker, or integration with GitHub, etc?

[Andrew Neil]

Posted on March 23, 2018 at 01:37

As it's Eclipse, I guess it inherits the standard Eclipse integrations with Git (which work with GitHub).

As to tracking bugs in TrueStudio itself, that's another question!