TIM6 interrupt in debug/breakpoint

yannick23 · ‎2015-11-20

Posted on November 20, 2015 at 14:22

Hi everyone,

I need some lights on the ability to switch off interrupt of TIM6 when stopping at a breakpoint.

I did call the macro:

__HAL_FREEZE_TIM6_DBGMCU();

But nothing, I still jump into the TIM6_IRQHandler when I step into the code.

Any clue ?

Thanks anyway !

(Coocox, GCC, STM32L0)

#debug-breakpoint-tim

waclawek.jan · ‎2015-11-20

Posted on November 20, 2015 at 19:37

> I did call the macro:

>

> __HAL_FREEZE_TIM6_DBGMCU();

Why? (read: what did you do to find out what to call to disable interrupts?)

JW

yannick23 · ‎2015-11-23

Posted on November 24, 2015 at 07:57

Hi,

Because I want to run my board in debug, so if I want to debug step by step, I need to stop getting interrupt of TIM6 otherwise, when I set a breakpoint, I will go in TIM6_IRQHandler the next step. Which I dont want.

http://www.st.com/web/en/resource/technical/document/reference_manual/DM00108282.pdf

section 27.9.4. page 807

waclawek.jan · ‎2015-11-24

Posted on November 24, 2015 at 09:02

OK and does the timer run or no when you set the breakpoint (i.e. look at the timer's registers and refresh that window).

You can also read and write the DBG_APB1_FZ register directly in the debugger - if it's not in the registers window, you can do it through the memory window.

JW

yannick23 · ‎2015-11-24

Posted on November 24, 2015 at 11:19

the counter still runs.

I did try to set the

__HAL_FREEZE_TIM6_DBGMCU();

Before the timer init, no changes

I try to manually create a pointer at address 0x40015808 and write 0x10,

No changes, I do not see anything written in register DBGMCU

Edit : I did try to write to another register, and it works so it seams to be the problem ( was making sure the way of viewing my register ...) (So 0x10 isn't written in my register)

Edit 2 : I try to manually write 1 at the address with the Memory browser tool (inside coocox ) and it's not setting it (I'm not sure I can do this like that by the way)

waclawek.jan · ‎2015-11-24

Posted on November 24, 2015 at 12:12

It might be the debugger being overtly smart and rewriting it for you. Try a bog standard gdb instead.

JW

waclawek.jan · ‎2015-11-24

Posted on November 24, 2015 at 12:16

It might be the debugger being overtly smart and rewriting it. You might want to search for something like a ''stop timers during debugging'' checkbox or some other similar setting. Or try a bog standard gdb instead.

Also, note that for DBG registers, ''only 32-bit access supported''.

JW

yannick23 · ‎2015-11-25

Posted on November 25, 2015 at 13:41

https://my.st.com/public/STe2ecommunities/mcu/Lists/cortex_mx_stm32/Flat.aspx?RootFolder=https%3a//my.st.com/public/STe2ecommunities/mcu/Lists/cortex_mx_stm32/Freeze%20timers%20in%20breakpoints&FolderCTID=0x01200200770978C69A1141439FE559EB459D7580009C4E14902C3CDE46A77F0FFD06506F5B&currentviews=1190

__DBGMCU_CLK_ENABLE();

Yay ...

Edit : Where is the option to ''mark as solved ?

waclawek.jan · ‎2015-11-25

Posted on November 25, 2015 at 14:21

Ah, RCC_APB2ENR.DBGMCUEN... I should've thought about it when I noticed that - contrary to the 'F4 I am using daily - the 'F0's DBG is not part of core, rather, sits on APB2...

Thanks for coming back with the solution.

JW