Thread324 32位定时器在PIC32MZ上读取

我猜手工编码的访问是“工作”的，因为它只读取计数器的低端部分。为了读取完整的32位计数器，如果高的部分发生变化，需要重新读取低端部分，所以我会做一些类似的事情（省略索引检查）：PLBIB模板UTI32 32 T TMRYA32 32位寄存器16位寄存器安全（TMRY）。TyrMyReaveT*Re=（（TMRA RealSts**）（index））；TMRA Reavest*tRMR*HOLD＝（（TMRA RealStss*T*）（索引+0x0200 U））；UInt32-T ValLo，ValHuy；DO {ValHe=＝（UTInt3}t）（TMR-Hig & Gt；TMRx）；ValLo=（UTInt3}t）（TMRLU-＆GT；TMRx）；}（ValHy.）=（UTIN 32×T）（TMR-Hyg & Gt；TMRx）；/ /如果高部分改变返回（（ValHy＆Lt；lt；16u）ValLo）重复；}



      以下为原文

    I guess the hand coded access "works" since it only reads the low part of the counter.
To read the full 32 Bit counter the low-part needs to be re-read if the high part changes, so I'd do something like this (omitting index check):
 
PLIB_TEMPLATE uint32_t TMR_Counter32BitGet_In16BitRegister_Safe( TMR_MODULE_ID index )
{
    tmr_registers_t volatile * tmrLow = ((tmr_registers_t *)(index));
    tmr_registers_t volatile * tmrHigh = ((tmr_registers_t *)(index + 0x0200u));
    uint32_t valLow, valHigh;
    do {
        valHigh = (uint32_t)(tmrHigh->TMRx);
        valLow = (uint32_t)(tmrLow->TMRx);
        }
    while(valHigh != (uint32_t)(tmrHigh->TMRx)); // repeat if HighPart changed
    return((valHigh << 16u) | valLow);
    }
 

实际上它做了一个32位的读取并读取完整的计数器！这是没有记录的！数据表声明它应该在上16位返回0。然而，它确实返回正确的高位。Microchip有人能证实吗？最后，我写了以下内容：UTI32 32位计数器32位BIT保存（TMRMuleMexEdID索引）{UInt32t Val1＝0；IF（TMRx），VAR1=（UTIN 32）（（TMRSReistSts:Value*）索引）-& gt；；；}返回（VAL1）；}



      以下为原文

    Actually it does an 32 bit read and reads the complete counter! This is not documented! The datasheet states it should return 0 in the upper 16 bit. However it does return the correct high bits. Can someone from Microchip confirm this?
 
I ended up writing the following:
 
uint32_t Counter32BitGetSave(TMR_MODULE_ID index)
{
   uint32_t val1 = 0;
   if( _TMR_MODULE_ID_IS_EVEN(index) )
   {
      val1 = (uint32_t)((tmr_registers_t volatile *)index)->TMRx;
   }
   else
   {
      PLIB_ASSERT(false, "This Timer instance does not support PLIB_TMR_Counter32BitGet");
   }
return ( val1 );
}

很好的发现。眨眼：



      以下为原文

    nice find. wink:

参见FRM的“32位定时器考虑”第143.2.1节：



      以下为原文

    See section 14.3.2.1, "32-bit Timer Considerations" of the FRM:

好啊！谢谢确认Andersm！



      以下为原文

    OK! Thanks for confirming Andersm!

我认为有人应该向支持报告，所以可能在未来的PLIB实现中得到修复。有志愿者吗？眨眼：



      以下为原文

    I think someone should report it to the support, so maybe it might get fixed in a future PLIB implementation. Any volunteers? wink: