[Bug c/22104] New: using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[Bug c/22104] New: using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

I was minding my own business compiling some math code, the profile-generate
build worked and ran fine then I tried a profile-use build and it borked
bitching and all...

-- 
           Summary: using "-fprofile-use" on code that has HUGE amounts of
                    inline asm borks it
           Product: gcc
           Version: 3.4.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: tomstdenis at gmail dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: Linux prescott 2.6.11.11 #1 Tue May 31 09:06:43 EDT 2005
                    i686 In
  GCC host triplet: Linux prescott 2.6.11.11 #1 Tue May 31 09:06:43 EDT 2005
                    i686 In
GCC target triplet: Linux prescott 2.6.11.11 #1 Tue May 31 09:06:43 EDT 2005
                    i686 In


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 14:31 -------
Created an attachment (id=9104)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9104&action=view)
It's the preprocessed code that died.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|Linux prescott 2.6.11.11 #1 |
                   |Tue May 31 09:06:43 EDT 2005|
                   |i686 In                     |
   GCC host triplet|Linux prescott 2.6.11.11 #1 |
                   |Tue May 31 09:06:43 EDT 2005|
                   |i686 In                     |
 GCC target triplet|Linux prescott 2.6.11.11 #1 |i686-pc-linux-gnu
                   |Tue May 31 09:06:43 EDT 2005|
                   |i686 In                     |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug middle-end/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-17 14:36 -------
Can you show how you compiled both with -fprofile-generate and -fprofile-use?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |middle-end


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 14:51 -------
(In reply to comment #2)
> Can you show how you compiled both with -fprofile-generate and -fprofile-use?

cc -march=prescott -fprofile-generate -Wall -W -Wshadow -I./ -O3 -funroll-loops
-fomit-frame-pointer   -c -o fp_set.o fp_set.c

Typical compile statement [just the .c and .o file names change].  I ran my
demo, then did 

rm -f *.[ao]

Then did the build again  switching "-generate" to "-use".

Tom

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|middle-end                  |c


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug middle-end/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-17 14:52 -------
(In reply to comment #3)
One more question, what is the ICE?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |middle-end


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 14:54 -------
(In reply to comment #4)
> (In reply to comment #3)
> One more question, what is the ICE?

What is ICE and how do I figure that out?

Tom

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|middle-end                  |c


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-17 14:55 -------
What is the error comming from GCC with -fprofile-use?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug c/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 15:32 -------
(In reply to comment #6)
> What is the error comming from GCC with -fprofile-use?

cc -march=prescott -fprofile-use -Wall -W -Wshadow -I./ -O3 -funroll-loops
-fomit-frame-pointer   -c -o fp_sqr_comba.o fp_sqr_comba.c
fp_sqr_comba.c: In function `fp_sqr_comba_small':
fp_sqr_comba.c:364: warning: 'sc0' might be used uninitialized in this function
fp_sqr_comba.c:364: warning: 'sc1' might be used uninitialized in this function
fp_sqr_comba.c:364: warning: 'sc2' might be used uninitialized in this function
fp_sqr_comba.c:1872: internal compiler error: in verify_local_live_at_start, at
flow.c:546
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://bugs.gentoo.org/> for instructions.
Preprocessed source stored into /tmp/cc626adC.out file, please attach this to
your bugreport.
make: *** [fp_sqr_comba.o] Error 1


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-17 15:36 -------
(In reply to comment #7)
> See <URL:http://bugs.gentoo.org/> for instructions.

Is there a reason why you reported this to here and not to gentoo or did gentoo tell you to file it here?
Also since this is only reproducible with -fprofile-use, can you attach the full program since we cannot 
reproduce without running it?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c                           |rtl-optimization
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 15:41 -------
Created an attachment (id=9105)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9105&action=view)
Profile data for the file that dies


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[tomstdenis at gmail dot com]

------- Additional Comments From tomstdenis at gmail dot com  2005-06-17 15:44 -------
(In reply to comment #8)
> (In reply to comment #7)
> > See <URL:http://bugs.gentoo.org/> for instructions.
> 
> Is there a reason why you reported this to here and not to gentoo or did
gentoo tell you to file it here?
> Also since this is only reproducible with -fprofile-use, can you attach the
full program since we cannot 
> reproduce without running it?

I've sent the profile data and I've uploaded the code to

http://libtomcrypt.org/tfm.tar.bz2

Build with:
CFLAGS="-march=prescott -fprofile-generate" make && gcc -march=prescott
-fprofile-generate -O3 -fomit-frame-pointer -funroll-loops ecc.c libtfm.a -o ecc

Then run "./ecc"

Then run "rf -f [.][ao]"

Rebuild with:

CFLAGS="-march=prescott -fprofile-use" make && gcc -march=prescott -fprofile-use
-O3 -fomit-frame-pointer -funroll-loops ecc.c libtfm.a -o ecc



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Comment #13 from pinskia at gcc dot gnu dot org  2006-11-20 05:49 -------
No feedback in 3 months for the full source so it is hard to figure out what is
going wrong here so closing.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[pinskia at gcc dot gnu dot org]

------- Comment #12 from pinskia at gcc dot gnu dot org  2006-08-21 05:56 -------
Can you attach the source which is needed to run this program?
All I get trying to reproduce this bug in a new GCC is:
[pinskia@celery ~/Desktop]$ ~/gcc-local/bin/gcc t.c -O2 -fomit-frame-pointer
/usr/lib/crt1.o: In function `_start':
(.text+0x18): undefined reference to `main'
collect2: ld returned 1 exit status


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104

[Bug rtl-optimization/22104] using "-fprofile-use" on code that has HUGE amounts of inline asm borks it

[steven at gcc dot gnu dot org]

------- Comment #11 from steven at gcc dot gnu dot org  2006-04-24 22:51 -------
I'm inclined to say "don't do this" but i can see why this is a problem for the
reporter.  I can't reproduce the issue myself, though.  Tom, do you still see a
problem here?  Sorry for the long delay btw.


-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22104