This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Incorrect code on PowerPC


The EGCS-1.1.1 compiler generates incorrect code on Linux/PPC (and I'm
sure most any RS6000 target).  This problem is [again] a register being
incorrectly reused while it is still live and destined for another use.

This bug makes compiling the latest GTK/GNOME packages fail.  I have
extracted a piece of one of the GTK library files and isolated it to
a single function.

The attached file contains the source, ".i", ".s" files, along with
an explanation of why the code is incorrect.  I can send the files from
"-da" separately, but they are really large.

Questions: 
  Is there some document I can read which will help me to better
  understand these [dump] files?  

  Also, a correlation of which file gets generated when and how the 
  various -Ox options affect this would be useful.

  Is there any [portable] way to get some indication in the ".s" file
  of the corresponding C source lines?

Thanks.  Let me know how I can help with this.

gtkbug.tar.gz


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]