Coverage measures the amount of testing done of a certain type. It is always in relation to some kind of measure, for example: statement coverage, branch coverage, etc. (Not restricted to code coverage).
Cem Kaner's "The Impossibility of Complete Testing" is helpful for anyone being asked for "100% test coverage": http://www.kaner.com/pdfs/impossible.pdf
Also worth reading is his "Software Negligence and Testing Coverage": http://www.badsoftware.com/coverage.htm (the appendix lists 101 different coverage measures).
Brian Marick has some wise warnings in "How to misuse code coverage": http://www.exampler.com/testing-com/writings/coverage.pdf
And Michael Bolton on: "Finding bugs vs coverage" http://www.developsense.com/presentations/FindingBugsVersusCoverage.pdf