Here's the ticket:
http://hackage.haskell.org/trac/ghc/ticket/2917
Lennart's patch added allocaBytesAligned to Foreign.Marshal.Alloc, but we'll have to do a library proposal to add that.
If you can reproduce the assertion failure, please submit a bug report.
Unfortunately this assertion failure, though it will always happen eventually, has proven to be a real pain to track down. I filed a bug report a long time ago, closed it when I mistakenly thought I had fixed the problem, and now I'm waiting until I've narrowed things down some more. I'm no longer so suspicious of allocaBytes since it still happened after I replaced it with alloca, which, contrary to the trac ticket, does appear to use allocaBytesAligned. And I don't really see how it could cause memory corruption anyway...