Please note that '-fno-tree-ch' is passed to GCC4 only for compiling
op.c, so that 'dyngen' doesn't fail, when it is working on op.o
afterwards! All other source files are still compiled without this flag.
On x86 GCC4 fails during compilation of op.c anyway. But '-fno-tree-
ch' could also he
On Tue, 20 Dec 2005, Daniel Egger wrote:
> On 19.12.2005, at 23:35, Fabrice Bellard wrote:
>
> > Does -fno-tree-ch works on PC ?
>
> Yes. However using it might introduce a performance hit.
I think Fabrice meant does the -fno-tree-ch trick works on x86 with gcc4
too? The answer is no.
__
Hi,
On Tue, 20 Dec 2005, Daniel Egger wrote:
> On 19.12.2005, at 23:35, Fabrice Bellard wrote:
>
> > Does -fno-tree-ch works on PC ?
>
> Yes. However using it might introduce a performance hit.
Why not just test for Darwin *and* for GCC 4, and be done with it?
Ciao,
Dscho
_
On 19.12.2005, at 23:35, Fabrice Bellard wrote:
Does -fno-tree-ch works on PC ?
Yes. However using it might introduce a performance hit.
Servus,
Daniel
PGP.sig
Description: This is a digitally signed message part
___
Qemu-devel mailing list
Does -fno-tree-ch works on PC ?
Fabrice.
Yes. According to the GCC docs[1], -ftree-ch is enabled by -O1 and
above on all architectures.
[1] http://gcc.gnu.org/onlinedocs/gcc-4.0.2/gcc/Optimize-Options.html
- Josh
___
Qemu-devel mailing list
Qem
Does -fno-tree-ch works on PC ?
Fabrice.
Joachim Henke wrote:
I know that you have already discussed this on the list, and that you don't
want to actively support GCC 4. But (at least on Mac OS X) only a fix in
the build configuration is necessary to do the magic: op.c must be compiled
with -fn
I know that you have already discussed this on the list, and that you don't
want to actively support GCC 4. But (at least on Mac OS X) only a fix in
the build configuration is necessary to do the magic: op.c must be compiled
with -fno-tree-ch (as already mentioned on this mailing list).
GCC 4.0 is
