Chris Lattner c277b19e97 Teach valuetracking that byval arguments with a specified alignment are ... aligned. Teach memcpyopt to not give up all hope when confonted with an underaligned memcpy feeding an overaligned byval. If the *source* of the memcpy can be determined to be adequeately aligned, or if it can be forced to be, we can eliminate the memcpy. This addresses PR9794. We now compile the example into: define i32 @f(%struct.p* nocapture byval align 8 %q) nounwind ssp { entry: %call = call i32 @g(%struct.p* byval align 8 %q) nounwind ret i32 %call } in both x86-64 and x86-32 mode. We still don't get a tailcall though, because tailcalls apparently can't handle byval. llvm-svn: 131884		2011-05-23 00:03:39 +00:00
..
2008-02-24-MultipleUseofSRet.ll
2008-03-13-ReturnSlotBitcast.ll
align.ll	fix typo	2011-01-10 02:33:34 +00:00
crash.ll	fix PR8753, eliminating a case where we'd infinitely make a	2010-12-09 07:39:50 +00:00
dg.exp
form-memset.ll	revert 123144, reenabling the rest of memset formation.	2011-01-12 03:25:15 +00:00
loadstore-sret.ll
memcpy-to-memset.ll	MemCpyOpt: Turn memcpys from a constant into a memset if possible.	2010-12-24 21:17:12 +00:00
memcpy.ll	Teach valuetracking that byval arguments with a specified alignment are	2011-05-23 00:03:39 +00:00
memmove.ll
smaller.ll	Actually check memcpy lengths, instead of just commenting about	2011-01-21 22:07:57 +00:00
sret.ll	Implement PR8644: forwarding a memcpy value to a byval,	2010-11-21 00:28:59 +00:00