Subtitution of a DIV by a MUL, why does this work?

Question

Consider the following code:

unsigned f( unsigned u )
{
    return u / 10;
}

Surprisingly the compiler generates the following code from this:

mov     eax, 0xcccccccd
mul     ecx
shr     edx, 3
mov     eax, edx
ret

So why does this work, i.e. why aren't there rounding errors ?