We consider the classical M/G/1 queue with two priority classes and th
e nonpreemptive and preemptive-resume disciplines. We show that the lo
w-priority steady-state waiting-time can be expressed as a geometric r
andom sum of i.i.d. random variables, just like the M/G/1 FIFO waiting
-time distribution. We exploit this structures to determine the asympt
otic behavior of the tail probabilities. Unlike the FIFO case, there i
s routinely a region of the parameters such that the tail probabilitie
s have non exponential asymptotics. This phenomenon even occurs when b
oth service-time distributions are exponential. When non-exponential a
symptotics holds, the asymptotic form tends to be determined by the no
n-exponential asymptotics for the high-priority busy-period distributi
on. We obtain asymptotic expansions for the low-priority waiting-time
distribution by obtaining an asymptotic expansion for the busy-period
transform from Kendall's functional equation. We identify the boundary
between the exponential and non-exponential asymptotic regions. For t
he special cases of an exponential high-priority service-time distribu
tion and of common general service-time distributions, we obtain conve
nient explicit forms for the low-priority waiting-time transform. We a
lso establish asymptotic results for cases with long-tail service-time
distributions. As with FIFO, the exponential asymptotics tend to prov
ide excellent approximations, while the non-exponential asymptotics do
not, but the asymptotic relations indicate the general form. In all c
ases, exact results can be obtained by numerically inverting the waiti
ng-time transform.