CmmParse: Don't force alignment in memcpy-ish operations
authorBen Gamari <ben@smart-cactus.org>
Mon, 3 Aug 2015 13:31:03 +0000 (15:31 +0200)
committerBen Gamari <ben@smart-cactus.org>
Mon, 3 Aug 2015 15:21:52 +0000 (17:21 +0200)
commit64b6733e217f32feb8c4872096749c6f417d0c39
tree1cab4eb2dbeaa5cf40789a19a65453a1395eeed5
parent7ec6ffc4e8ccf5c24149d9ab34c3619516dc3081
CmmParse: Don't force alignment in memcpy-ish operations

This was initially made in 681973c31c614185229bdae4f6b7ab4f6e64753d.
Here I wanted to enforce that the alignment passed to %memcpy was a
constant expression, as this is required by LLVM. However, this breaks
the knot-tying done in `loopDecls`, causing T8131 to hang.

Here I remove the `seq` and mark T8131 as `expect_broken` in the case
of the NCG, which doesn't force the alignment in this case.

Fixes #10664.
compiler/cmm/CmmParse.y
testsuite/tests/codeGen/should_fail/all.T