On Tue, Dec 06, 2011 at 09:16:22PM +0000, Richard Sandiford wrote: > Showing my ignorance here, but is that the usual behaviour for this kind > of thing? I wouldn't have expected versions to apply to internally-created > symbols. Normally this sort of symbol isn't dynamic, at least nowadays. Earlier versions of GNU ld made many symbols like _GLOBAL_OFFSET_TABLE_ dynamic unnecessarily, so it might be just a case of old mips code following even older practice. -- Alan Modra Australia Development Lab, IBM