Should we not round the end of Linux modules to page boundaries?

_From [bruen...@google.com](https://code.google.com/u/109494838902877177630/) on July 19, 2013 17:21:15_

this was PR 307599

Currently when providing events to clients and storing in our module list we round the end of the modules to the then next page. We could instead use the exact end (we can figure it out from the program headers and it's the size the loader uses for its first map). On the other hand the maps file rounds to page size (so we'd have to adjust all asserts on maps file walking) and when the loader does the anonymous .bss section map it rounds it up to page size. Of course on unmap the loader again uses the exact size. Rounding to page size is probably what we want, but documenting the decision here.

_Original issue: http://code.google.com/p/dynamorio/issues/detail?id=1222_