[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

xfs,powerpc, gcc (was Re: RaiserFS PPC status)



On Mon, May 07, 2001 at 08:03:57AM -0500, Rahul Jain wrote:
> On Mon, May 07, 2001 at 12:00:34PM +0200, Just a friendly Jedi Knight wrote:
> > On Sun, May 06, 2001 at 09:58:02PM -0500, Rahul Jain wrote:
> > > On Mon, May 07, 2001 at 12:03:43AM +0200, Just a friendly Jedi Knight wrote:
> ...
> > > you should stick to gcc 2.95 for compiling the kernel, and probably the
> > > userspace tools, too.
> >  Most probably you are right.. I used gcc-3.0 because i heard that it
> >  compiles reiserfs code (big endian) correctly. So far i have no problems
> >  with kernel built with gcc-3.0 
> 
> I don't really know of anyone who's using a gcc-3.0-compiled XFS kernel. You
> might want to check on the mailing list as to how well it does with XFS.
 I guess that my mention about using gcc-3.0 was not that good. To clearify a
 bit kernel compiled with gcc-3.0 runs just fine (ok XFS part is not used
 now). The real point thou is that kernel compiles without problems and
 xfsprogs do not compile. 
 Nathan Scott from SGI send me patch but it didn't work. I now figured some
 other way to fix that and it seems it partially worked (now it quits with:
=== mkfile ===
gcc -g -DDEBUG -funsigned-char -Wall  -I../include '-DVERSION="1.2.4"'
-D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DXFS_BIG_FILES=1
-DXFS_BIG_FILESYSTEMS=1   -c -o xfs_mkfile.o xfs_mkfile.c
xfs_mkfile.c: In function `main':
xfs_mkfile.c:144: `O_DIRECT' undeclared (first use in this function)
xfs_mkfile.c:144: (Each undeclared identifier is reported only once
xfs_mkfile.c:144: for each function it appears in.)
make[1]: *** [xfs_mkfile.o] Error 1

(i'm writing this at the same as it's compiling). This should be easier to
solve 

> You need MORE optimization to enable inlining. Inlining reduces debuggability
> and increases executable size for the benefit of not having to set up a whole
> new stack frame and possibly being able to optimize the function in the special
> context of where it's called from. Maybe -O2 will fix it?
 Already tried this.

-- 
 Robert Ramiega | jedi@plukwa.net  IRC: _Jedi_ | Do not underestimate 
 UIN: 13201047  | http://www.plukwa.net/       | the power of  Source



Reply to: