[Bug analyzer/106620] New: Incorrectly thinks execution can continue after a return statement

[Bug analyzer/106620] New: Incorrectly thinks execution can continue after a return statement

[nrk at disroot dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106620

            Bug ID: 106620
           Summary: Incorrectly thinks execution can continue after a
                    return statement
           Product: gcc
           Version: 12.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: analyzer
          Assignee: dmalcolm at gcc dot gnu.org
          Reporter: nrk at disroot dot org
  Target Milestone: ---

Created attachment 53457
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53457&action=edit
sample code

Sample code (also attached) : https://godbolt.org/z/ez556Yd39

It thinks that memchr() is being called with null. But if `hay` was null, then
the function returns, so that's not possible. Removing the `|| nlen == 1`
silences the warning.

[Bug analyzer/106620] Incorrectly thinks execution can continue after a return statement

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106620

David Malcolm <dmalcolm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |WORKSFORME

--- Comment #1 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Thanks for filing this.

Looking at the Compiler Explorer link, the false positive occurs with gcc 11.1
at -O3.

With gcc 11.1 and gcc 11.3 the false +ve occurs at -O1 and above.

With gcc 12.1 and trunk, the false +ve does not occur.

I've made lots of implementation changes to -fanalyzer in gcc 12 in beyond that
I don't intend to backport to gcc 11, so I'm going to mark this as resolved
with "WORKSFORME".

Feel free to reopen it if you're able to reproduce it with gcc 12 or later;
thanks again for filing this.

[Bug analyzer/106620] Incorrectly thinks execution can continue after a return statement

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106620

--- Comment #2 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
(In reply to David Malcolm from comment #1)

> I've made lots of implementation changes to -fanalyzer in gcc 12 in beyond
"and beyond", I meant to say

[Bug analyzer/106620] Incorrectly thinks execution can continue after a return statement

[nrk at disroot dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106620

--- Comment #3 from nrk at disroot dot org ---
> Feel free to reopen it if you're able to reproduce it with gcc 12 or later

Huh, I linked to godbolt specifically because I wanted to test gcc 12 and see
if the bug still persists or not since my local version is at 11.2.1. But looks
like I messed it up.

In any case, thanks for fixing it!