背靠背bset命令存在已知问题

我刚刚开始使用PPL24 Explorer 16 EVARE板，MPLAB X IDE V3.35和APIC24FJ256GB110处理器模块。我从以前的项目中导入了一组代码，基本上有一个“永远”的循环。在这个循环中，我添加了以下代码：查看“if”子句内容的程序集代码，我看到（反汇编的简化版本）：注意，对于同一个寄存器（PORTA），有两个背靠背的“BSET”指令。在Eval板上，我期望LEDD9和D10同时发光，并且发光速率相同（尽管由于连续的汇编代码“BSET”指令，间隙很小）。但是，当我在调试器中“运行”代码时，我只看到两个LED中的一个亮了。另一条指令短暂地打开（240ns），然后随着第二条指令的打开而关闭（显然，我有一个o范围来告诉它）。就好像第二条BSET指令在某种程度上同时对要设置的原始位执行BCLR一样。如果我在IF子句本身碰到断点，然后通过I/EL子句指令单步执行，它就完美地工作了。此外，我可以重写代码中的两种方法之一，并使它像预期的那样运行（当“运行”）。这两种方式是：或者：关于问题可能有什么想法？PIC24对同一个寄存器进行背对背“BSET”指令是否真的存在一个已知问题？这似乎不太可能，但这正是我所目睹的。谢谢。



      以下为原文

    I'm just starting out with the PIC24 Explorer 16 Eval board with MPLAB X IDE v3.35 and a PIC24FJ256GB110 processor module.  I've imported a bunch of code from a previous project and have basically a "while forever" loop.  Inside this loop, I've added the following code:
if (toggle == 0) {
toggle = 1;
PORTAbits.RA7 = 1;
PORTAbits.RA6 = 1;
} else {
toggle = 0;
PORTAbits.RA7 = 0;
PORTAbits.RA6 = 0;
}Looking at the assembly code for the contents of the 'if' clause, I see (shortened version of disassembly):

toggle = 1; MOV #0x1, W0
MOV W0, [W14+2]
PORTAbits.RA7 = 1; BSET PORTA, #7
PORTAbits.RA6 = 1; BSET PORTA, #6

Note that there are two back-to-back "BSET" instructions to the same register (PORTA).  On the Eval board, I expect both LEDs D9 and D10 to illuminate at the same time and rate (though with a small gap due to consecutive assembly code "BSET" instructions).  However, when I "run" the code in the debugger, I see only one of the two LEDs lit.  The other one turns on briefly (240ns) and then turns off as the second one turns on (obviously I have an o-scope to tell this).

It is as though the second BSET instruction is somehow simultaneously doing a BCLR on the original bit I wanted to set.  If I hit a breakpoint at the if clause itself and then single-step through the if/else clause instructions, it works perfectly.  Further, I can rewrite the code one of two ways and make it work as expected (when "running").  Those two ways are:
if (toggle == 0)
{
PORTAbits.RA7 = 1;
toggle = 1;
PORTAbits.RA6 = 1;
}
else
{
PORTAbits.RA7 = 0;
toggle = 0;
PORTAbits.RA6 = 0;
}Or:
if (toggle == 0)
{
toggle = 1;
PORTA = PORTA | 0x00C0;
}
else
{
toggle = 0;
PORTA = PORTA & ~0x00C0;
}
Any thoughts on what the problem could be?  Is there REALLY a known issue with the PIC24 doing back-to-back 'BSET' instructions to the same register?  This seems highly unlikely but is what I appear to be witnessing.

Thanks.