STM32F3 IRQ remap option and some minor improvements

[salkinium]

Hi, thanks for this awesome library! I'm currently integrating TinyUSB into modm (modm-io/modm#478) and I found some minor things to improve:

1. I added a simple STM32F3_IRQ_REMAP, which can be defined in the config to remap the IRQs. The user must remap the IRQs in their hardware setup (it's just SYSCFG->CFGR1 |= SYSCFG_CFGR1_USB_IT_RMP;).
2. We're using TinyUSB in C++, which doesn't define the __STDC_VERSION__ macro, thus throwing a warning. Just a simple fix.
3. The _MESS_ERR() and _MESS_FAILED() macros use printf instead of tu_printf, which is unfortunate for debugging and also links the newlib version of printf which requires heap, thus pulling a whole lot of unrelated code.
4. To improve debugability I'm overwriting the TU_ASSERT macros with our own fault reporting implementation, and to allow this I just wrapped them in #ifndef. Unfortunately OpenOCD does not by default reset the debugger connected flag on disconnect, thus TU_BREAKPOINT() causes a HardFault, which isn't great to debug.
5. We're using SCons to compile, which has issues resolving #include CFG_TUSB_CONFIG_FILE, thus not recompiling TinyUSB sources when the config changes. So I changed the mechanism to "fallback" to a plain #include "tusb_config.h", which solves this issue.

[hathach]

Thank for the PR, sorry for late response, I have been too busy lately. The PR looks great, your modification and explanation all make sense. I only want to know a bit more about the STM32 remap to see if we could implement it without having an extra DCD specific macro.

[hathach]

This is interesting, I didn't know about this, could you tell me more about the specific MCU that support this, and which section in the reference manual for this. I would like to read this up a bit, to see if we could have a better systematic solution. I kind of not wanting to have custom macro for specific DCD if there is a better way.

[salkinium]

This only exists for the STM32F302 and the larger STM32F303 devices.

See RM0365 Table 40, page 208, footnote 1 (STM32F302) and RM0318 Table 82, page 288, footnote 1 (STM32F303xb/c/d/e):

1. It is possible to remap the USB interrupts (USB_HP, USB_LP and USB_WKUP) on interrupt lines 74, 75 and 76 respectively by setting the USB_IT_RMP bit in the Section 11.1.1: SYSCFG configuration register 1 (SYSCFG_CFGR1) on page 172.

[salkinium]

I guess you could read the bit at runtime and select the right interrupts. Wouldn't be too much overhead.

[hathach]

@salkinium sorry for late response, this is great idea to solve the shared CAN/USB. I like your last suggestion by checking the actual remap bit. The overhead is too small ( a couple of cycle for bit test) and this aren't call at any fast rate either. Would you mind changing it to dynamic check. Also since it only available to selected chips, it should be wrapped with SYSCFG_CFGR1_USB_IT_RMP e.g

#ifdef SYSCFG_CFGR1_USB_IT_RMP
if ( remapped )
{
 use RMP IRQ
}else
#endif
{
  use normal IRQ 
}

[salkinium]

I implemented it like this and remapped the IRQs on the STM32F303 DISCO BSP.

[hathach]

Thanks for the great idea with remap, would you mind taking a bit more time to make it dynamically check of the remapped bit in the SYSCFG_CFGR1, that would be perfect.

[hathach]

@salkinium sorry for late response, this is great idea to solve the shared CAN/USB. I like your last suggestion by checking the actual remap bit. The overhead is too small ( a couple of cycle for bit test) and this aren't call at any fast rate either. Would you mind changing it to dynamic check. Also since it only available to selected chips, it should be wrapped with SYSCFG_CFGR1_USB_IT_RMP e.g

#ifdef SYSCFG_CFGR1_USB_IT_RMP
if ( remapped )
{
 use RMP IRQ
}else
#endif
{
  use normal IRQ 
}

[hathach]

almost perfect, though hand-on testing with f3 disco prove that syscfg clk must be enabled before we could do the remap.

[hathach]

perfect now, thank you very much for your PR. I will merge when ci passed.