typedef unsigned char tChar;
typedef unsigned int  tInt;
typedef unsigned long tLong;

Using your own typedefs like this, rather than just char, int, etc, is a good idea as it allows you to "encapsulate" compiler-dependencies.
However, your names are not very informative; "tChar" tells you nothing explicitly about the size or signed-ness of the type.
Therefore, I prefer names like:
U8 - Unsigned, 8-bits;
S8 - Signed, 8-bits;
U16 - Unsigned, 16-bits;
S16 - Signed, 8-bits;
etc, etc.
These explicitly & succinctly indicate both size & signed-ness.
You may prefer something slightly less terse, but you get the idea...

void msDelay(tInt msec)
{

tInt cntr_1, cntr_2;

for (cntr_1 = 0; cntr_1 <= msec; ++cntr_1 )
    {
      for ( cntr_2 = 0; cntr_2 <= 5000; ++cntr_2);
    }
}

You should define cntr_1, cntr_2 as volatile - otherwise the compiler may completely optimise this code away!
Also, beware that you have absolutely no guarantee whatsoever as to what specific machine code will be generated - so any change to build options, and certainly a compiler update, could very easily change the timing - possibly drastically.
Therefore, you need to keep an eye on the generated assembler.
However, since you need to keep an eye on the generated assembler, you might as well write it in assembler in the first place!
This is definitely one of those places where assembler is better than 'C'!