[Bug c/114507] New: FR: Randomize order of evluation of function of arguments

[Bug c/114507] New: FR: Randomize order of evluation of function of arguments

[vt at altlinux dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114507

            Bug ID: 114507
           Summary: FR: Randomize order of evluation of function of
                    arguments
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vt at altlinux dot org
  Target Milestone: ---

There is class of bugs when, for example, printf is called with several
functions that output to the same static buffer overwriting each others result.
This could be hard to detect in legacy code, for example there is nested calls
invisible from top level functions, also compiler tend to create the same call
order, so if overwriting is order dependent it could be undetected for years
and appear on some rare architecture when compiler reorder calls [1].

To detect this kind of bugs it would be useful to have compiler option to
reorder function arguments evaluation randomly. With this test suites will have
a _chance_ to detect the bugs even on one architecture.

[1] https://lore.kernel.org/all/20240323091717.dtsioyvjnhl5dtyv@altlinux.org/

[Bug c/114507] FR: Randomize order of evluation of function of arguments

[sjames at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114507

Sam James <sjames at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sjames at gcc dot gnu.org

--- Comment #1 from Sam James <sjames at gcc dot gnu.org> ---
Maybe we should have a metabug for hardening/debugging-style feature requests
like this (PR82179 comes to mind).

[Bug c/114507] FR: Randomize order of evluation of function of arguments

[sjames at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114507

--- Comment #2 from Sam James <sjames at gcc dot gnu.org> ---
That said, I feel as if it's more likely this is better for the analyser.

[Bug c/114507] FR: Randomize order of evluation of function of arguments

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114507

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

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

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
That being said this is a dup of bug 12411 .

*** This bug has been marked as a duplicate of bug 12411 ***

[Bug c/114507] FR: Randomize order of evluation of function of arguments

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114507

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Well the diagnostic part was PR 12411 :) (which I filed years ago and decided
to close as won't fix less than 3 years ago).