Marc Glisse <marc.gli...@inria.fr> writes: > Hello, > > this patch lets gcc know that if a pointer existed before the call to > malloc, the result of malloc cannot alias it. This is a bit ad hoc and > fragile. A small improvement would be, when the 2 statements are in the > same bb but in the wrong order, to check if there is any statement in > between that might prevent from reordering them. But that's more > complicated, and the patch as it is already does help. > > I expect people may not like the call to a function from > tree-ssa-loop-niter too much, but it is convenient. And if someone > improves it, they will likely have to rewrite something not quite > equivalent to stmt_dominates_stmt_p.
It has linear complexity for statements in the same block though. (reassoc_stmt_dominates_stmt_p avoids that, but relies on uids being up-to-date.) > The testcase uses libstdc++ quite a bit. I thought about putting it in the > libstdc++ testsuite, but it does not know scan-tree-dump, and in the > assembly it may be too late, memcpy may get expanded to something > target-specific. So it is in g++.dg. I could write a more artificial > testcase, but the behavior of std::vector is really what I want to test... > > Bootstrap+regtest on x86_64-pc-linux-gnu. > > 2019-05-13 Marc Glisse <marc.gli...@inria.fr> > > gcc/ > * tree-ssa-loop-niter.c (stmt_dominates_stmt_p): Handle NULL > basic block. > * tree-ssa-alias.c: Include tree-ssa-loop-niter.h. > (ptr_derefs_may_alias_p): Detect malloc and an older pointer. > > gcc/testsuite/ > * g++.dg/tree-ssa/ldist-2.C: New file.
