[Bug tree-optimization/126815] [17 Regression] internal compiler error: verify_flow_info failed caused by r17-3204-ge02ce3b8c4574c

"pinskia at gcc dot gnu.org via Gcc-bugs" <[email protected]>
Newsgroups gmane.comp.gcc.bugs
Message-ID <[email protected]/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126815

--- Comment #8 from Drea Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Macleod from comment #7)
> (In reply to Drea Pinski from comment #4)
> > (In reply to Andrew Macleod from comment #3)
> > > yes, either leave the call, or I also have a patch in the queue which puts
> > > the _9 = 0 before the call instead of after.   This should be safe as there
> > > are known to be no other uses of _9
> > 
> > Before wont work for calls that returns-twice. So it is best not to change
> > it. Plus it saves slight memory and time .
> 
> eh? why wouldn't that work? There are zero uses of _9, so it doesn't matter
> if its the return value or not.
> 
> Regardless, your patch is fine.

Because returns_twice functions are required by the verifier to be the first
stmt of a basic block. So if you sre removing the lhs of a returns twice
functions inserting before the call will also ice. I didnt come up with a
testcase but I can add one if needed.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.