Re: UNIQUE ID (INSN UID) Question

2008-01-26 Thread Richard Guenther
On Jan 26, 2008 1:34 AM, Balaji V. Iyer <[EMAIL PROTECTED]> wrote:
>
> Hello Everyone,
> I have a quick question regarding instruction unique ID in the RTL.
> Is this number unique for the function? or is it unique for the entire
> program that it is compiling?
>
>
> I would like to "mark" instructions and identify them, so can I use this
> value as a unique identifier for program level?

They are unique per function.  In fact, RTL only exists for a single function
at a time.

Richard.


Magic incantation for running multilib testsuite.

2008-01-26 Thread David Daney
I have tried several times (and failed) to run the GCC testsuite on
multilib targets (x86-64 and mips64) and have not been able to figure
out how to get more that a single multilib configuration to be tested. 
The instructions on http://gcc.gnu.org/install/test.html are not working
for me.  I think I want something like:

$ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"

But that only ends up running unix/-m32

I know that it must be possible as I see the results in [EMAIL PROTECTED]

If a kind soul could show the exact command line used to invoke a
combined multilib testsuite run, I would appreciate it.

David Daney


Re: Magic incantation for running multilib testsuite.

2008-01-26 Thread Andreas Schwab
David Daney <[EMAIL PROTECTED]> writes:

> $ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"
>
> But that only ends up running unix/-m32

You need to add more quotes:

$ make -k check  RUNTESTFLAGS="--target_board=unix/\{,-m32\}"

The flags are passed unquoted to runtest, thus the shell will expand the
braces before runtest can see them.

Andreas.

-- 
Andreas Schwab, SuSE Labs, [EMAIL PROTECTED]
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."


Re: Magic incantation for running multilib testsuite.

2008-01-26 Thread Peter Bergner

On Sat, 2008-01-26 at 08:39 -0800, David Daney wrote:
> I have tried several times (and failed) to run the GCC testsuite on
> multilib targets (x86-64 and mips64) and have not been able to figure
> out how to get more that a single multilib configuration to be tested. 
> The instructions on http://gcc.gnu.org/install/test.html are not working
> for me.  I think I want something like:
> 
> $ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"

On powerpc64-linux, I use the following to run both 32-bit and 64-bit tests:
 
  make -k check RUNTESTFLAGS="--target_board=unix'{-m32,-m64}'"

so as Andreas mentioned in another note, you're missing extra qoutes.

Peter





gcc-4.2.3-RC1's gctest hangs on i686-apple-darwin9

2008-01-26 Thread Jack Howarth
   Can someone else please try a test of gcc 4.2.3-RC1 on i686-apple-darwin9?
I am finding that (with the java langauge built) I am getting an apparent hang
in the make check at...

mkdir tests
/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/xgcc 
-B/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/bin/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/lib/ -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/include -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/sys-include -DHAVE_CONFIG_H -I. 
-I../../../gcc-4.2.3-RC-20080125/boehm-gc -I./include -I./include  
-I/sw/src/fink.build/gcc42-4.2.3-1000/gcc-4.2.3-RC-20080125/boehm-gc/include  
-fexceptions -Iinclude -I././targ-include -I.//libc/include -O2 -g -O2  -c -o 
tests/test.o ../../../gcc-4.2.3-RC-20080125/boehm-gc/tests/test.c
/bin/sh ./libtool --mode=link 
/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/xgcc 
-B/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/bin/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/lib/ -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/include -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/sys-include -fexceptions -Iinclude 
-I././targ-include -I.//libc/include -O2 -g -O2   -o gctest -shared-libgcc 
tests/test.o ./libgcjgc.la -lpthread   
/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/xgcc 
-B/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/bin/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/lib/ -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/include -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/sys-include -fexceptions -Iinclude 
-I././targ-include -I.//libc/include -O2 -g -O2 -o .libs/gctest -shared-libgcc 
tests/test.o   ./.libs/libgcjgc.dylib -lpthread
creating gctest
make "AR_FLAGS=rc" "CC_FOR_BUILD=gcc" "CFLAGS=-O2 -g -O2 " "CXXFLAGS=-g -O2 " 
"CFLAGS_FOR_BUILD=-g -O2" "CFLAGS_FOR_TARGET=-O2 -g -O2 " 
"INSTALL=/usr/bin/install -c" "INSTALL_DATA=/usr/bin/install -c -m 644" 
"INSTALL_PROGRAM=/usr/bin/install -c" "INSTALL_SCRIPT=/usr/bin/install -c" 
"LDFLAGS=" "LIBCFLAGS=-O2 -g -O2 " "LIBCFLAGS_FOR_TARGET=-O2 -g -O2 " 
"MAKE=make" "MAKEINFO=makeinfo --split-size=500 --split-size=500  " 
"PICFLAG=" "PICFLAG_FOR_TARGET=" "SHELL=/bin/sh" "EXPECT=expect" 
"RUNTEST=runtest" "RUNTESTFLAGS=" "exec_prefix=/sw/lib/gcc4.2" 
"infodir=/sw/share/info" "libdir=/sw/lib/gcc4.2/lib" "prefix=/sw/lib/gcc4.2" 
"tooldir=/sw/lib/gcc4.2/i686-apple-darwin9" "AR=ar" 
"AS=/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/as" 
"CC=/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/xgcc 
-B/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/bin/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/lib/ -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/include -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/sys-include" 
"CXX=/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/g++ 
-B/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/ -nostdinc++ 
-nostdinc++ 
-I/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/i686-apple-darwin9/libstdc++-v3/include/i686-apple-darwin9
 
-I/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/i686-apple-darwin9/libstdc++-v3/include
 
-I/sw/src/fink.build/gcc42-4.2.3-1000/gcc-4.2.3-RC-20080125/libstdc++-v3/libsupc++
 
-I/sw/src/fink.build/gcc42-4.2.3-1000/gcc-4.2.3-RC-20080125/libstdc++-v3/include/backward
 
-I/sw/src/fink.build/gcc42-4.2.3-1000/gcc-4.2.3-RC-20080125/libstdc++-v3/testsuite/util
 
-L/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/i686-apple-darwin9/libstdc++-v3/src
 
-L/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/i686-apple-darwin9/libstdc++-v3/src/.libs
 -B/sw/lib/gcc4.2/i686-apple-darwin9/bin/ 
-B/sw/lib/gcc4.2/i686-apple-darwin9/lib/ -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/include -isystem 
/sw/lib/gcc4.2/i686-apple-darwin9/sys-include" 
"LD=/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/collect-ld" 
"LIBCFLAGS=-O2 -g -O2 " 
"NM=/sw/src/fink.build/gcc42-4.2.3-1000/darwin_objdir/./gcc/nm" "PICFLAG=" 
"RANLIB=ranlib -c" "DESTDIR=" check-TESTS
Switched to incremental mode
Emulating dirty bits with mprotect/signals

I see one gctest process with five threads and four gnumake processes. None of 
these appear to have any
cpu usage. If I sample gctest in the Activity Monitor application, I see...

Sampling process 56569 for 3 seconds with 1 millisecond of run time between 
samples
Sampling completed, processing symbols...
Analysis of sampling gctest (pid 56569) every 1 millisecond
Call graph:
1988 Thread_2503
  1988 start
1988 _start
  1988 main
1988 run_one_test
  1988 GC_init_gcj_malloc
1988 GC_lock
  1988 pthread_mutex_lock
1988 semaphore_wait_signal_trap
  1988 semaphore_wait_signal_trap
1988 Thread_2603
  1988 thread_start
1988 _pthread_start
  1988 GC_mprotect_thread
1988 exc_server
   

Re: Magic incantation for running multilib testsuite.

2008-01-26 Thread David Daney
Andreas Schwab wrote:
> David Daney <[EMAIL PROTECTED]> writes:
>
>   
>> $ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"
>>
>> But that only ends up running unix/-m32
>> 
>
> You need to add more quotes:
>
> $ make -k check  RUNTESTFLAGS="--target_board=unix/\{,-m32\}"
>
> The flags are passed unquoted to runtest, thus the shell will expand the
> braces before runtest can see them.
>   
Indeed, it is obvious in hindsight.

Thanks,
David Daney


Re: Magic incantation for running multilib testsuite.

2008-01-26 Thread Matthias Klose
David Daney writes:
> I have tried several times (and failed) to run the GCC testsuite on
> multilib targets (x86-64 and mips64) and have not been able to figure
> out how to get more that a single multilib configuration to be tested. 
> The instructions on http://gcc.gnu.org/install/test.html are not working
> for me.  I think I want something like:
> 
> $ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"

when running with make -j, the testsuite is still run sequentially for
each pass; is there a way to run the passes in parallel?

  Matthias


Re: Magic incantation for running multilib testsuite.

2008-01-26 Thread Kaveh R. GHAZI
On Sun, 27 Jan 2008, Matthias Klose wrote:

> David Daney writes:
> > I have tried several times (and failed) to run the GCC testsuite on
> > multilib targets (x86-64 and mips64) and have not been able to figure
> > out how to get more that a single multilib configuration to be tested.
> > The instructions on http://gcc.gnu.org/install/test.html are not working
> > for me.  I think I want something like:
> >
> > $ make -k check  RUNTESTFLAGS="--target_board=unix/{,-m32}"
>
> when running with make -j, the testsuite is still run sequentially for
> each pass; is there a way to run the passes in parallel?
>   Matthias

I believe yes, I haven't tried it but see the end of item 0.2 here:
http://gcc.gnu.org/install/test.html

It says this feature is only supported in the gcc subdirectory.  (I don't
know if that's current or not.)  I suppose that means the top level target
library tests are not handled in parallel for their multilibs.

I'd be interested in your results if you try it out.

--Kaveh
--
Kaveh R. Ghazi  [EMAIL PROTECTED]


Re: GCC 4.3 target deprecation proposals

2008-01-26 Thread Hans-Peter Nilsson
On Thu, 24 Jan 2008, DJ Delorie wrote:
>
> > message was truncated because of the massive number of failures.
>
> Or massive number of multilibs :-)

Let me humbly and pragmatically suggest testing with just the
default multilib (or a much smaller subset than all you do) once
in a while.

Fixing test_summary to prune identical results for multilibs
would also be a solution.

brgds, H-P