T7571 with WAY=llvm fails, but not WAY=optllvm

Test case (#7571 (closed) just will not die, it seems)

$ make TEST=T7571 WAY=llvm
...
====> Scanning ./all.T
=====> T7571(llvm) 6 of 7 [0, 0, 0]
cd . && '/home/a/ghc/ghc-pristine/inplace/bin/ghc-stage2' -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-ghci-history -c T7571.cmm -fllvm  -no-hs-main   >T7571.comp.stderr 2>&1
Actual stderr output differs from expected:
--- /dev/null	2013-08-07 22:18:14.132934017 -0500
+++ ./T7571.comp.stderr	2013-08-14 01:26:33.942103723 -0500
@@ -0,0 +1 @@
+WARNING: Non constant alignment value given for memcpy! Please report to GHC developers
*** unexpected failure for T7571(llvm)

However, this test case does not fail under WAY=optllvm, which I think is worth investigating probably:

$ make TEST=T7571 WAY=optllvm
...
====> Scanning ./all.T
=====> T7571(optllvm) 6 of 7 [0, 0, 0]
cd . && '/home/a/ghc/ghc-pristine/inplace/bin/ghc-stage2' -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-ghci-history -c T7571.cmm -O -fllvm  -no-hs-main   >T7571.comp.stderr 2>&1

OVERALL SUMMARY for test run started at Wed Aug 14 01:25:42 CDT 2013
       7 total tests, which gave rise to
      28 test cases, of which
      27 were skipped

       0 had missing libraries
       1 expected passes
       0 expected failures

       0 caused framework failures
       0 unexpected passes
       0 unexpected failures

However, there is another way to trigger this bug in *both* cases:

#include "Cmm.h"

testMemcpy (W_ dst, W_ src, W_ l, W_ sz)
{
  prim %memcpy(dst, src, l, sz);
  return ();
}

will fail both ways.

Trac metadata

Trac field	Value
Version	7.7
Type	Bug
TypeOfFailure	OtherFailure
Priority	high
Resolution	Unresolved
Component	Compiler
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system	Unknown/Multiple
Architecture