Darwin 3.4 make check results with -mcpu=G5

Darwin 3.4 make check results with -mcpu=G5

[Bradley Lucier]

Make check results with and without -mcpu=G5 for 
powerpc-apple-darwin7.2.0 are reported at

http://gcc.gnu.org/ml/gcc-testresults/2004-01/msg00906.html

The biggest differences appear in the gcc tests:

		=== gcc Summary for unix/-mcpu=G5 ===

# of expected passes		23725
# of unexpected failures	379
# of unexpected successes	2
# of expected failures		68
# of unresolved testcases	236
# of untested testcases		7
# of unsupported tests		295

		=== gcc Summary for unix ===

# of expected passes		24268
# of unexpected failures	36
# of unexpected successes	2
# of expected failures		68
# of unresolved testcases	40
# of untested testcases		7
# of unsupported tests		295

There are a few patches dealing with 64-bit powerpc problems now being 
proposed, but it would appear that we have a way to go before having 
good enough 64-bit support with the darwin 32-bit/64-bit api.

Brad

Re: Darwin 3.4 make check results with -mcpu=G5

[Andrew Pinski]

On Jan 21, 2004, at 10:44, Bradley Lucier wrote:
> There are a few patches dealing with 64-bit powerpc problems now being 
> proposed, but it would appear that we have a way to go before having 
> good enough 64-bit support with the darwin 32-bit/64-bit api.

Okay I think it is not that far away because the current number of 
patches, 3, should fix most of the differences because they look like 
they are all the same problem.  And these 3
patches should also fix the bootstrap problem when compiling with 
-mcpu=G5.


The 3 patches:

Fix PR target/13785, ICE when using a function pointer that returns a 
long long:
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg02209.html

Fix PR middle-end/13779, wrong initialization of 'long long' with 
constant
(or an ICE when using a constant that would not fit in 32bits):
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg01497.html

Fix target/13674, ICE in reload_cse_simplify_operands, reload and 
ld/std:
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg02107.html



Thanks,
Andrew Pinski
pinskia@physics.uc.edu
apinski@apple.com