https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67755
Jeffrey A. Law <law at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |law at redhat dot com
--- Comment #3 from Jeffrey A. Law <law at redhat dot com> ---
It looks like the code to estimate the on/off path edge data is mucking things
up in the case where we've got a join block.
What's strange here is the way we estimate the probability/counts for the
outgoing edges of the join block's duplicate. ISTM that in the absence of
additional information, the right thing to do is apply the probabilities from
the join block to the duplicate of the join block to get the counts for the
on/off path.
I'm going to shovel some snow and ponder.
