[Octave-bug-tracker] [bug #64729] Octave segfault when running the fft r2c transform

[Hg200]

URL:
  <https://savannah.gnu.org/bugs/?64729>

                 Summary: Octave segfault when running the fft r2c transform
                   Group: GNU Octave
               Submitter: hg200
               Submitted: Sat 30 Sep 2023 01:27:22 PM UTC
                Category: Octave Function
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: None
                  Status: None
             Assigned to: None
         Originator Name: hg200
        Originator Email: 
             Open/Closed: Open
                 Release: other
         Discussion Lock: Any
        Operating System: Any
           Fixed Release: None
         Planned Release: None


    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Sat 30 Sep 2023 01:27:22 PM UTC By: Hg200 <hg200>
Hello all.

Running the attached script segfaults octave. I strongly suspect an
insufficient memory allocation as root cause, since "howmany" can be greater
than one. Changing following statement solves the problem for me: The command


OCTAVE_LOCAL_BUFFER (double, itmp, nn + 32);


should be rather


OCTAVE_LOCAL_BUFFER (double, itmp, nn * howmany + 32);


It can be found at:

https://hg.savannah.gnu.org/hgweb/octave/file/43d010974a89/liboctave/numeric/oct-fftw.cc#l369

Note that this instance exists two times: There is one implementation for
double as well as single (float) precision.








    _______________________________________________________
File Attachments:


-------------------------------------------------------
Date: Sat 30 Sep 2023 01:27:22 PM UTC  Name: test_fft.m  Size: 286B   By:
hg200

<http://savannah.gnu.org/bugs/download.php?file_id=55180>

    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?64729>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/