FreeRTOS Support Archive
The FreeRTOS support forum is used to obtain active support directly from Real
Time Engineers Ltd. In return for using our top quality software and services for
free, we request you play fair and do your bit to help others too! Sign up
to receive notifications of new support topics then help where you can.
This is a read only archive of threads posted to the FreeRTOS support forum.
The archive is updated every week, so will not always contain the very latest posts.
Use these archive pages to search previous posts. Use the Live FreeRTOS Forum
link to reply to a post, or start a new support thread.
[FreeRTOS Home] [Live FreeRTOS Forum] [FAQ] [Archive Top] [June 2009 Threads] Hint: STM32 HD stack problemPosted by Clemens Gerlach on June 16, 2009 Hi All,
I just searched a while so I share it here before someone else is wasting time...
I am using a STM32 High-Density device with FreeRTOS and the ST FWlib 3.0 and had system crashes. Reason: The supplied startup code in the FWlib sets the stack pointer in the vector table (1st entry) to the fixed address 0x20000400. After initializing the external memory interface (if any) it sets the MSP register to the desired value (_estack defined in the linker script). Normaly this should be ok but FreeRTOS resets the stack pointer (MSP) to the value stored in the vector table (when staring the first thread). This is ok and saves stack space.
The too ideas are nice but do not play together very well...
My solution was to edit the startup code (in $(FWLIB)/CMSIS/Core/CM3/startup) to place &_estack into the vector table (I am not using external RAM).
Clemens
RE: Hint: STM32 HD stack problemPosted by Richard on June 16, 2009 Hmm, interesting. All CM3 devices are supposed to have the stack location in the first position within the vector table, FreeRTOS locates the vector table through the system control registers, and resets the stack back to its start position. Are you saying the STM32 startup code defines a different stack in the linker script to that pointed to by the vector table? I will need to look at that in more detail - thanks for the heads up. What is at 0x20000400?
This must be resurged in the linker script?
Regards.
RE: Hint: STM32 HD stack problemPosted by Clemens Gerlach on June 16, 2009 Hi Richard,
the problem only occurs with the STM32 startup code for the high-density devices. Since firmware lib 3.0.0 they are included. I think their intention for using this scheme was to provide a stack for the startup code (RAM starts at 0x20000000), setup the external memory controller and then move the stack to the end of the ram (which may be inside the external memory).
I think FreeRTOS behavior is correct. It just collides with the (example) routines from ST. But if someone (for whatever reason) wants to put the stack into external RAM FreeRTOS' way of resetting the stack may be a problem.
From $(fwlib)/Libraries/CMSIS/Core/CM3/startup/gcc/startup_stm32f10x_hd.c:
#define Initial_spTop 0x20000400 <- this is in internal ram (cg)
[...]
__attribute__ ((section(".isr_vector"))) void (* const g_pfnVectors[])(void) = { (void *)Initial_spTop, /* The initial stack pointer */ Reset_Handler, ....
[...]
void Reset_Handler(void) { SystemInit_ExtMemCtl();
/* restore original stack pointer */ asm(" LDR r0, =_estack"); <- _estack defined by linker script asm(" MSR msp, r0");
/* Initialize data and bss */ __Init_Data();
/* Call the application's entry point.*/ main(); }
Maybe FreeRTOS' stack reset behavior sould be documented in the port information. Also the use of the dual stack pointers is not mentioned there...
BTW: Thank's for the great OS!
Clemens
Copyright (C) Amazon Web Services, Inc. or its affiliates. All rights reserved.
|