Though this in no way justifies the drug war, it's contribution to the prison population is often overestimated:
Myth No. 2: Low-level drug offenders drive prison population growth. It is popular, perhaps almost mandatory, to blame the boom on the War on Drugs. But it is just not true. Only 20 percent of inmates in prisons (as opposed to jails) are locked up for drug offenses, compared with 50 percent for violent crimes and 20 percent for property offenses; most of the drug offenders are in prison for distribution, not possession. Twenty percent is admittedly much larger than approximately 3 percent, which was the fraction of prisoners serving time on drug charges in the 1970s. But if we were to release every prisoner currently serving time for a drug charge, our prison population would drop only from 1.6 million to 1.3 million. That's not much of a decline, compared with the total number of people in prison in the 1970s—about 300,000.
In fact, the war on drugs does play a role in the prisoner increase. But it's an indirect one. State "predicate felony" laws, for example, impose longer sentences on offenders with prior records: A drug conviction may not send someone to prison, but it will make him serve more time for any future crime he commits. This suggests that simply tackling long drug sentences, as reformers in New York state have done, may miss the real problem.
One interesting point for context: the number of people incarcerated for drug crimes is now approaching the total prison population in the 1970s.