Re: [RFC PATCH] tree-ssa-sink: do not sink to in front of setjmp

[Alexander Monakov <amonakov@ispras.ru>]

> I approved the initial sink patch (maybe not clearly enough).

I wasn't entirely happy with that patch. The new version solves this better.

> Can you open
> a bugreport about the missing CFG verification and list the set of FAILs
> (all errors in some passes similar to the one you fixed in sinking I guess)?
> It indeed sounds like something to tackle during next stage1 (unless you
> already narrowed down the culprit to a single offender...)

Most of the failures were related to transactional memory, and the rest are
seemingly solved by forbidding duplication of returns_twice calls. In reply
to this email I'm sending three patches, the first is a revised patch for
tree-ssa-sink, the second forbids duplication of setjmp-like calls, and the
third implements the checks in verify_flow_info:

  tree-ssa-sink: do not sink to in front of setjmp
  tree-cfg: do not duplicate returns_twice calls
  tree-cfg: check placement of returns_twice calls

 gcc/testsuite/gcc.dg/setjmp-7.c | 13 +++++++++++
 gcc/tree-cfg.c                  | 40 +++++++++++++++++++++++++++++++--
 gcc/tree-ssa-sink.c             |  6 +++++
 3 files changed, 57 insertions(+), 2 deletions(-)
 create mode 100644 gcc/testsuite/gcc.dg/setjmp-7.c

-- 
2.33.1