[Bug optimization/11975] New: gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[Bug optimization/11975] New: gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[tausq at debian dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975

           Summary: gcc miscompiles natClassLoader.cc with -fnon-call-
                    exceptions
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: tausq at debian dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: hppa-unknown-linux
  GCC host triplet: hppa-unknown-linux
GCC target triplet: hppa-unknown-linux

While trying to build libjava for hppa-linux, i ran into a few cases where gcc
built code that segfaults when run. One example is attached -- this code was
extracted from natClassLoader.cc, when built with -fnon-call-exceptions, gcc
emits an insn:

       ldw RT'loaded_classes(%r25),%r25

but r25 is never initialized. It looks like the gcse pass eliminated the
initialization of r25

The complete build command was:
gcc -fno-rtti -fdollars-in-identifiers -Wswitch-enum -fnon-call-exceptions -O2
-g -fPIC -S bug.ii

With -fno-gcse, the code looks like:
        .loc 1 48 0
        addil LT'loaded_classes,%r19
        ldw RT'loaded_classes(%r1),%r1

without -fno-gcse
        .loc 1 48 0
        ldw RT'loaded_classes(%r25),%r25

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[tausq at debian dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975



------- Additional Comments From tausq at debian dot org  2003-08-19 05:12 -------
Created an attachment (id=4618)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=4618&action=view)
test case

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975



------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-28 16:21 -------
Most likely related to bug 11767.

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[danglin at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975



------- Additional Comments From danglin at gcc dot gnu dot org  2003-08-29 00:58 -------
I did some investigation of the problem and found that the gcc pass
eliminated the initialization of the instruction in question.  It
probably should have eliminated the insn itself.  Possibly, it didn't
because it contains an unspec construct.

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-10-11 03:02 -------
Does this happen any more, there was a fix for -fgcse recently to fix a problem like this?

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[dhazeghi at yahoo dot com]

------- Additional Comments From dhazeghi at yahoo dot com  2003-12-11 00:30 -------
Is this bug still occurring on mainline? I tried taking a look, but my hppa-linux cross-compiler 
outputted assembly somewhat different from what you included :-)

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975

[Bug optimization/11975] gcc miscompiles natClassLoader.cc with -fnon-call-exceptions

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-12-28 03:08 -------
No feedback in 3 months (T-15 days).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11975