* Re: corrections to recent profile-arcs change
@ 2002-05-11 19:25 John David Anglin
2002-05-12 3:43 ` Jan Hubicka
0 siblings, 1 reply; 8+ messages in thread
From: John David Anglin @ 2002-05-11 19:25 UTC (permalink / raw)
To: gcc; +Cc: rth
There are a bunch on new testsuite regressions under hppa2.0w-hp-hpux11.11
that appear related to profile-arcs changes:
FAIL: gcc.misc-tests/bprob-1.c compilation, -g -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -O0 -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -O1 -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -O2 -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -O3 -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -O3 -g -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-1.c compilation, -Os -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -g -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -O0 -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -O1 -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -O2 -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -O3 -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -O3 -g -DPERFTIME -fprofile-arcs
FAIL: gcc.misc-tests/bprob-2.c compilation, -Os -fprofile-arcs
FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-1.c gcov failed: Could not open data file gcov-1.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-2.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-2.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-2.c gcov failed: Could not open data file gcov-2.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-3.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-3.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-3.c gcov failed: Could not open data file gcov-3.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-4.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-4.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-4.c gcov failed: Could not open data file gcov-4.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-4b.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-4b.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-4b.c gcov failed: Could not open data file gcov-4b.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-5b.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-5b.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-5b.c gcov failed: Could not open data file gcov-5b.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-6.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-6.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-6.c gcov failed: Could not open data file gcov-6.da.
Assuming that all execution counts are zero.
FAIL: gcc.misc-tests/gcov-7.c (test for excess errors)
WARNING: gcc.misc-tests/gcov-7.c compilation failed to produce executable
FAIL: gcc.misc-tests/gcov-7.c gcov failed: Could not open data file gcov-7.da.
Assuming that all execution counts are zero.
FAIL: g++.dg/bprob/bprob-1.C compilation, -g -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -O0 -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -O1 -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -O2 -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -O3 -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -O3 -g -fprofile-arcs
FAIL: g++.dg/bprob/bprob-1.C compilation, -Os -fprofile-arcs
FAIL: g++.dg/gcov/gcov-1.C (test for excess errors)
WARNING: g++.dg/gcov/gcov-1.C compilation failed to produce executable
FAIL: g++.dg/gcov/gcov-1.C gcov failed: Could not open data file gcov-1.da.
Assuming that all execution counts are zero.
FAIL: g++.dg/gcov/gcov-2.C (test for excess errors)
WARNING: g++.dg/gcov/gcov-2.C compilation failed to produce executable
FAIL: g++.dg/gcov/gcov-2.C gcov failed: Could not open data file gcov-2.da.
Assuming that all execution counts are zero.
FAIL: g++.dg/gcov/gcov-3.C (test for excess errors)
WARNING: g++.dg/gcov/gcov-3.C compilation failed to produce executable
FAIL: g++.dg/gcov/gcov-3.C gcov failed: Could not open data file gcov-3.da.
Assuming that all execution counts are zero.
FAIL: g++.robertl/eb83.C (test for excess errors)
FAIL: g77.dg/bprob/bprob-1.f compilation, -g -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -O0 -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -O1 -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -O2 -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -O3 -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -O3 -g -fprofile-arcs
FAIL: g77.dg/bprob/bprob-1.f compilation, -Os -fprofile-arcs
FAIL: g77.dg/gcov/gcov-1.f (test for excess errors)
WARNING: g77.dg/gcov/gcov-1.f compilation failed to produce executable
FAIL: g77.dg/gcov/gcov-1.f gcov failed: Could not open data file gcov-1.da.
Assuming that all execution counts are zero.
In the C fails, the error is:
/var/tmp/ccyMSnnU.c:26: size of variable `*L$PBX0002' is too large
collect2: gcc returned 1 exit status
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-11 19:25 corrections to recent profile-arcs change John David Anglin
@ 2002-05-12 3:43 ` Jan Hubicka
2002-05-12 8:50 ` John David Anglin
0 siblings, 1 reply; 8+ messages in thread
From: Jan Hubicka @ 2002-05-12 3:43 UTC (permalink / raw)
To: John David Anglin; +Cc: gcc, rth
> There are a bunch on new testsuite regressions under hppa2.0w-hp-hpux11.11
> that appear related to profile-arcs changes:
>
> FAIL: gcc.misc-tests/bprob-1.c compilation, -g -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -O0 -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -O1 -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -O2 -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -O3 -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -O3 -g -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-1.c compilation, -Os -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -g -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -O0 -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -O1 -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -O2 -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -O3 -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -O3 -g -DPERFTIME -fprofile-arcs
> FAIL: gcc.misc-tests/bprob-2.c compilation, -Os -fprofile-arcs
I am just looking for proper fix for these. The problem is that C++ fronend gets
called to create alias set and is not able to grok the structures constructed by
our code. I hope to get some help from C++ folks here, as I am quite lost about
how to construct them C++ friendly.
> FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
> WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
> FAIL: gcc.misc-tests/gcov-1.c gcov failed: Could not open data file gcov-1.da.
> Assuming that all execution counts are zero.
What hapepnds to the compilation? Why the executable is not produced?
Honza
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-12 3:43 ` Jan Hubicka
@ 2002-05-12 8:50 ` John David Anglin
2002-05-12 9:44 ` Jan Hubicka
0 siblings, 1 reply; 8+ messages in thread
From: John David Anglin @ 2002-05-12 8:50 UTC (permalink / raw)
To: Jan Hubicka; +Cc: gcc, rth
> > FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
> > WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
> > FAIL: gcc.misc-tests/gcov-1.c gcov failed: Could not open data file gcov-1.da.
> > Assuming that all execution counts are zero.
>
> What hapepnds to the compilation? Why the executable is not produced?
All the C failures generated errors similar to this:
Executing on host: /xxx/gnu/gcc-3.2/objdir/gcc/xgcc -B/xxx/gnu/gcc-3.2/objdir/gc
c/ /xxx/gnu/gcc-3.2/gcc/gcc/testsuite/gcc.misc-tests/gcov-1.c -fprofile-arcs -
ftest-coverage -lm -o ./gcov-1.exe (timeout = 300)
/var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
collect2: gcc returned 1 exit status
compiler exited with status 1
output is:
/var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
collect2: gcc returned 1 exit status
FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
Excess errors:
/var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
collect2: gcc returned 1 exit status
WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
This appears to be a problem with the .c file being generated by collect2.
There aren't any errors in the initial compilation and we have the
following definition for L$PBX0002:
.align 8
L$PBX0002
.blockz 40
I see collect2 tries to compile the code to init constructors as follows:
../xgcc -x c -c -o /var/tmp//ccqqBDR3.o -B../ -fprofile-arcs -ftest-coverage -fno-exceptions -w /var/tmp//ccYuin0B.c
and this is where the above error occurs. Here is what ccYuin0B.c looks like:
#ifdef __cplusplus
extern "C" {
#endif
typedef void entry_pt();
extern entry_pt x4 __asm__ ("_GLOBAL__I_noopGCOV");
entry_pt * __CTOR_LIST__[] = {
(entry_pt *) 1,
x4,
0
};
entry_pt * __DTOR_LIST__[] = {
(entry_pt *) 0,
0
};
extern entry_pt __main;
entry_pt *__main_reference = __main;
#ifdef __cplusplus
}
#endif
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-12 8:50 ` John David Anglin
@ 2002-05-12 9:44 ` Jan Hubicka
2002-05-12 10:26 ` John David Anglin
2002-05-12 21:54 ` John David Anglin
0 siblings, 2 replies; 8+ messages in thread
From: Jan Hubicka @ 2002-05-12 9:44 UTC (permalink / raw)
To: John David Anglin; +Cc: Jan Hubicka, gcc, rth
> > > FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
> > > WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
> > > FAIL: gcc.misc-tests/gcov-1.c gcov failed: Could not open data file gcov-1.da.
> > > Assuming that all execution counts are zero.
> >
> > What hapepnds to the compilation? Why the executable is not produced?
>
> All the C failures generated errors similar to this:
>
> Executing on host: /xxx/gnu/gcc-3.2/objdir/gcc/xgcc -B/xxx/gnu/gcc-3.2/objdir/gc
> c/ /xxx/gnu/gcc-3.2/gcc/gcc/testsuite/gcc.misc-tests/gcov-1.c -fprofile-arcs -
> ftest-coverage -lm -o ./gcov-1.exe (timeout = 300)
> /var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
> collect2: gcc returned 1 exit status
> compiler exited with status 1
> output is:
> /var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
> collect2: gcc returned 1 exit status
>
> FAIL: gcc.misc-tests/gcov-1.c (test for excess errors)
> Excess errors:
> /var/tmp/ccd0q6hV.c:26: size of variable `*L$PBX0002' is too large
> collect2: gcc returned 1 exit status
>
> WARNING: gcc.misc-tests/gcov-1.c compilation failed to produce executable
>
> This appears to be a problem with the .c file being generated by collect2.
> There aren't any errors in the initial compilation and we have the
> following definition for L$PBX0002:
>
> .align 8
> L$PBX0002
> .blockz 40
I see, this is underflow for maximal value of zero sized array.
Does the attached patch solve your problem?
Sat May 12 18:20:36 CEST 2001 Jan Hubicka <jh@suse.cz>
* final.c (end_final): Do not output profile_arcs constructor, when
no functions are instrumented.
*** final.c.old Sun May 12 19:35:30 2002
--- final.c Sun May 12 19:35:41 2002
*************** void
*** 252,258 ****
end_final (filename)
const char *filename;
{
! if (profile_arc_flag)
{
char name[20];
tree char_pointer_type, char_array_type, string_type, string_cst;
--- 252,258 ----
end_final (filename)
const char *filename;
{
! if (profile_arc_flag && profile_info.count_instrumented_edges)
{
char name[20];
tree char_pointer_type, char_array_type, string_type, string_cst;
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-12 9:44 ` Jan Hubicka
@ 2002-05-12 10:26 ` John David Anglin
2002-05-12 21:54 ` John David Anglin
1 sibling, 0 replies; 8+ messages in thread
From: John David Anglin @ 2002-05-12 10:26 UTC (permalink / raw)
To: Jan Hubicka; +Cc: jh, gcc, rth
> Does the attached patch solve your problem?
>
> Sat May 12 18:20:36 CEST 2001 Jan Hubicka <jh@suse.cz>
> * final.c (end_final): Do not output profile_arcs constructor, when
> no functions are instrumented.
Yes, it appears to resolve the problem. I've started a complete bootstrap
and check with the patch.
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-12 9:44 ` Jan Hubicka
2002-05-12 10:26 ` John David Anglin
@ 2002-05-12 21:54 ` John David Anglin
2002-05-13 4:30 ` Jan Hubicka
1 sibling, 1 reply; 8+ messages in thread
From: John David Anglin @ 2002-05-12 21:54 UTC (permalink / raw)
To: Jan Hubicka; +Cc: jh, gcc, rth
> Sat May 12 18:20:36 CEST 2001 Jan Hubicka <jh@suse.cz>
> * final.c (end_final): Do not output profile_arcs constructor, when
> no functions are instrumented.
This fixes all the C and fortran errors with no regressions that I report.
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
2002-05-12 21:54 ` John David Anglin
@ 2002-05-13 4:30 ` Jan Hubicka
0 siblings, 0 replies; 8+ messages in thread
From: Jan Hubicka @ 2002-05-13 4:30 UTC (permalink / raw)
To: John David Anglin; +Cc: Jan Hubicka, gcc, rth
> > Sat May 12 18:20:36 CEST 2001 Jan Hubicka <jh@suse.cz>
> > * final.c (end_final): Do not output profile_arcs constructor, when
> > no functions are instrumented.
>
> This fixes all the C and fortran errors with no regressions that I report.
Thanks. I guess we should supress output of the constructor on the other place
as well, so I send patch once testing finishes.
Honza
>
> Dave
> --
> J. David Anglin dave.anglin@nrc.ca
> National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: corrections to recent profile-arcs change
[not found] <no.id>
@ 2002-05-11 20:28 ` John David Anglin
0 siblings, 0 replies; 8+ messages in thread
From: John David Anglin @ 2002-05-11 20:28 UTC (permalink / raw)
To: John David Anglin; +Cc: gcc, rth
> FAIL: g++.dg/bprob/bprob-1.C compilation, -O2 -fprofile-arcs
This one a SIGSEGV:
(gdb) r -fprofile-arcs -O2 bprob-1.ii
...
Program received signal SIGSEGV, Segmentation fault.
0x0010c574 in ok_to_generate_alias_set_for_type (t=0x402dda10)
at ../../gcc/gcc/cp/cp-lang.c:214
214 if (! CLASSTYPE_NON_POD_P(t))
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2002-05-13 10:18 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-05-11 19:25 corrections to recent profile-arcs change John David Anglin
2002-05-12 3:43 ` Jan Hubicka
2002-05-12 8:50 ` John David Anglin
2002-05-12 9:44 ` Jan Hubicka
2002-05-12 10:26 ` John David Anglin
2002-05-12 21:54 ` John David Anglin
2002-05-13 4:30 ` Jan Hubicka
[not found] <no.id>
2002-05-11 20:28 ` John David Anglin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).