Marc Glisse wrote:
I noticed that in some cases we were failing to find aliasing
information because we were only looking at an SSA_NAME variable,
missing the fact that it was really an ADDR_EXPR. The attached patch
passes bootstrap+testsuite, does it make sense? (I am a bit afraid of
losing some type information for instance)
I didn't investigate the 2 tests where I had to remove dg-bogus, because
removing dg-bogus sounds like a bonus...
I wonder why you are seeing failures with those test cases. I tired your
patch (w/o modifying the test cases) and they passed.
The idea of those test cases is to ensure that there is no output like
"note: loop versioned for vectorization because of possible aliasing"
Because that output would be a hint that "#pragma GCC ivdep" doesn't work.
What kind of output do you see? Seemingly not the one above as you kept
the "version" dg-bogus.
Tobias
