gnu optimization bug

Collapse
This topic is closed.
X
X
Collapse
 
  • Page of 1
  • Time
  • Show
Clear All
new posts
Previous template Next
  • Abhishek
    #1

    gnu optimization bug

    Recently i found that my C++ program was running fine with no
    optimization and giving segmentation fault with error code 139 when
    compiled with optimization level 2. I searched somewhat but never find
    any systematic solution. In my recent attempt, i found the problem with
    the optimization tool itself, which iam discussing here.

    I found that the problem of segmentation fault was boiled down to a
    statement of the form A = B, with no pointer or array usage, so forget
    about the illegal memory write problems. When i removed this statement,
    the problem disappers, so concluded that the segmentation fault problem
    is with this statement. A careful study revealed that this statement
    was redundant because A was never used in the program after this
    statement. So i inserted one more statement after the above statement
    which is "X = A", where X is a dummy variable and in that case the
    problem again disappears. Inserting this new statement has made my
    initial statement no more redundant.

    I think the optimization tool had detected that A = B is a redundant
    statement but failed to do its work. So its a clear bug in the
    optimization tool itself.

    The gcc -v output of my system is following:

    Using built-in specs.
    Target: i486-linux-gnu
    Configured with: ../src/configure -v
    --enable-languages=c,c++ ,java,f95,objc, ada,treelang --prefix=/usr
    --enable-shared --with-system-zlib --libexecdir=/usr/lib
    --without-included-gettext --enable-threads=posix --enable-nls
    --program-suffix=-4.0 --enable-__cxa_atexit
    --enable-libstdcxx-allocator=mt --enable-clocale=gnu
    --enable-libstdcxx-debug --enable-java-gc=boehm --enable-java-awt=gtk
    --enable-gtk-cairo
    --with-java-home=/usr/lib/jvm/java-1.4.2-gcj-4.0-1.4.2.0/jre
    --enable-mpfr --disable-werror --enable-checking=releas e i486-linux-gnu
    Thread model: posix
    gcc version 4.0.2 (Debian 4.0.2-2)

    Tags: None
    • Sharad Kala
      #2
      Re: gnu optimization bug


      "Abhishek" <abhishekh.seth @gmail.com> wrote in message

      | Recently i found that my C++ program was running fine with no
      | optimization and giving segmentation fault with error code 139 when
      | compiled with optimization level 2. I searched somewhat but never find
      | any systematic solution. In my recent attempt, i found the problem with
      | the optimization tool itself, which iam discussing here.

      You should take the topic to the gcc mailing list.

      Sharad


        Comment

        • Kai-Uwe Bux
          #3
          Re: gnu optimization bug

          Abhishek wrote:
          [g++ bug report redacted]

          Bug reports for particular compilers are off-topic in this group. All we can
          do is help you determine whether there is actually a bug. Since you did not
          post the code, which may or may not exhibit undefined behavior, we cannot
          help with that either.


          Best

          Kai-Uwe Bux

            Comment

            • Rolf Magnus
              #4
              Re: gnu optimization bug

              Abhishek wrote:
              [color=blue]
              > Recently i found that my C++ program was running fine with no
              > optimization and giving segmentation fault with error code 139 when
              > compiled with optimization level 2. I searched somewhat but never find
              > any systematic solution. In my recent attempt, i found the problem with
              > the optimization tool itself, which iam discussing here.
              >
              > I found that the problem of segmentation fault was boiled down to a
              > statement of the form A = B, with no pointer or array usage, so forget
              > about the illegal memory write problems. When i removed this statement,
              > the problem disappers, so concluded that the segmentation fault problem
              > is with this statement. A careful study revealed that this statement
              > was redundant because A was never used in the program after this
              > statement. So i inserted one more statement after the above statement
              > which is "X = A", where X is a dummy variable and in that case the
              > problem again disappears. Inserting this new statement has made my
              > initial statement no more redundant.
              >
              > I think the optimization tool had detected that A = B is a redundant
              > statement but failed to do its work. So its a clear bug in the
              > optimization tool itself.[/color]

              Without any more code, it's hard to say if you're right. What type are A and
              B? Could you provide a minimal, but complete program that shows the
              behavior. Sometimes, a program crash happens in a place different from the
              one that contains the error.

              As others already mentioned: A g++ bug report would have to go to a
              different place than here, but here, we could tell you if it's actually a
              bug in the compiler or an error in your code. However, for that, you'd have
              to show some code.

                Comment

                Previous template Next
                Working...