[Bug bootstrap/26175] New: In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[Bug bootstrap/26175] New: In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[malitzke at metronets dot com]

Manifest itself in make check.


-- 
           Summary: In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex
                    undefined
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: malitzke at metronets dot com
 GCC build triplet: powerpc-linux-gnu
  GCC host triplet: powerpc-linux-gnu
GCC target triplet: powerpc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug bootstrap/26175] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[rguenth at gcc dot gnu dot org]

------- Comment #1 from rguenth at gcc dot gnu dot org  2006-02-08 16:43 -------
You really need recent enough kernel (2.6.x with futex support) and binutils to
use openmp.  I guess we could come up with more configure check inflation,
tough.

You also should provide more information about your system (powerpc-linux-gnu
isn't really exhaustive).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|bootstrap                   |libgomp
           Keywords|                            |build
            Summary|In gcc-4.2.0                |[4.2 Regression] In gcc-
                   |libgomp/.../powerpc/futex.h |4.2.0
                   |SYS_futex undefined         |libgomp/.../powerpc/futex.h
                   |                            |SYS_futex undefined
   Target Milestone|---                         |4.2.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2006-02-10 06:36:32
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2006-05-21 20:19 -------
Any news on this one?  Shouldn't we have better autoconf for libgomp to dect
the case where SYS_futex does not exist?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[jakub at gcc dot gnu dot org]

------- Comment #3 from jakub at gcc dot gnu dot org  2006-05-22 08:31 -------
Users using prehistoric kernels and/or glibcs just need to configure
with --disable-linux-futex.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[malitzke at metronets dot com]

------- Comment #4 from malitzke at metronets dot com  2006-05-23 00:49 -------
No need to be offensive. 
At the time I was using the the latest kernel available (something like
2.6.15.x or 16.x) I was still able to compile on that dual G4 MAC glibc, as
available from gcc.gnu.org/pub/glibc/snapshots. Now, I cannot even get glibc to
compile because the weird way glibc_pic.so is derived from glibc_pic.a because
of some discarded...linkonce garbage. (see
sources.redhat.com/bugzilla/.../2672) 
Just casting aspersions on other parts of the GNU community or the user
community does not help the cause of _free_/open-source software.  

Richard Guenther's remark was helpful and perhaps should given more prominence
in the configure documentation for both gcc and binutils regarding openmp.
Example:
Using built-in specs.
Target: powerpc-unknown-linux-gnu
Configured with: ../gcc-4.2.0/configure --prefix=/usr 
--enable-version-specific-runtime-libs --with-java-home=/usr 
--infodir=/usr/share/gcc-42 --datadir=/usr/share/gcc-42 
--mandir=/usr/share/gcc-42 --program-suffix=-42 --enable-shared 
--enable-gomp --enable-mudflap --enable-libgfortran 
--enable-threads=posix --enable-__cxa_atexit --enable-libgcc-math 
--disable-checking --disable-multilib --disable-nls 
--disable-werror --with-gnu-ld --with-mpfr-dir=/src/src/mpfr-2.2.0 
--with-mpfr=/usr/lib --with-gmp-dir=/src/src/gmp-4.2 --with-gmp=/usr/lib 
--with-cpu=7450 --enable-languages=c,c++,fortran,obj-c++,java 
--enable-clocale=gnu
Thread model: posix
gcc version 4.2.0 20060521 (experimental)

BTW It is gcc that always wants to revert to __powerpc-unknown-linux-gnu__.
More descriptive alternatives cause too much hassle.

Also, I believe that gcc, binutils and glibc should be considered as one ball
of wax and be put under one umbrella. Really, glibc is now part of the C
__hosted__ specification. Also Drepper is already part of the g++ sub-empire. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |blocker


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[mmitchel at gcc dot gnu dot org]

------- Comment #5 from mmitchel at gcc dot gnu dot org  2006-06-04 17:35 -------
Yes, this is a critical problem.  OpenMP is a major feature in 4.2.  It would
be embarassing for it to break builds for users.  We should fix the configury
so as not to build the library on unsupported systems.


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[malitzke at metronets dot com]

------- Comment #6 from malitzke at metronets dot com  2006-06-06 03:59 -------
Good to see the GCC release manager looking at things from a user perspective,
and not just looking at an individual leaf in a forest.

Regarding the powerpc specific issues; I was able for just one day to compile a
complete glibc package. Now, after major work on mainline binutils (the one
with daily snapshots on ftp:gcc.gnu.org) It can not even pass a significant
portion of its own testsuite (while the ix86 build does fine). After having
been repeatedly told that the user community should not bother the
_illustrious_ glibc developers with build problems I have refrained from
contacting the binutils folks.

The steering committee, wisely, has recognized that the gcc group can __not__
contnuously and exhaustively test all software changes on the __vast__ array of
hardware configurations and thus has sought help from the user community. The
operating principle being that ""All bugs are shallow if there are thousands of
eyes looking "" (imperfect quote). 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[jakub at gcc dot gnu dot org]

------- Comment #7 from jakub at gcc dot gnu dot org  2006-06-20 09:55 -------
Subject: Bug 26175

Author: jakub
Date: Tue Jun 20 09:55:42 2006
New Revision: 114809

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=114809
Log:
        PR libgomp/26175
        PR libgomp/26477
        * configure.ac: If neither --enable-linux-futex nor
        --disable-linux-futex is passed, determine the default by checking
        for compiling and/or running against NPTL.  With --enable-linux-futex,
        check if SYS_gettid and SYS_futex are defined.
        * configure: Rebuilt.

Modified:
    trunk/libgomp/ChangeLog
    trunk/libgomp/configure
    trunk/libgomp/configure.ac


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175

[Bug libgomp/26175] [4.2 Regression] In gcc-4.2.0 libgomp/.../powerpc/futex.h SYS_futex undefined

[pinskia at gcc dot gnu dot org]

------- Comment #8 from pinskia at gcc dot gnu dot org  2006-06-23 04:35 -------
Fixed.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26175