On Fri, Dec 25, 2009 at 2:11 PM, Satish Balay <span dir="ltr"><<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Are you sure bfort got rebuilt witht he change? I cant get this valgrind output..<br></blockquote><div><br>Yes, I am positive. How do I check this? I make a change, rebuild and install, and<br>when I am in the debugger the line numbers change to reflect my source change.<br>
<br>This is plainly a bug. Its plain C with no library dependencies. It doesn't matter what<br>is on my machine, this should work. If it works somewhere else, that is a happy accident.<br><br> Matt<br> </div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
>>>>>>.<br>
asterix:/home/balay/tmp/spetsc/src/mat/utils>valgrind --tool=memcheck bfort -dir `pwd`/ftn-auto -mnative -ansi -nomsgs -noprofile -anyname -mapptr -mpi -mpi2 -ferr -ptrprefix Petsc -ptr64 PETSC_USE_POINTER_CONVERSION -fcaps PETSC_HAVE_FORTRAN_CAPS -fuscore PETSC_HAVE_FORTRAN_UNDERSCORE -f90mod_skip_header matio.c convert.c gcreate.c freespace.c getcolv.c ptap.c compressedrow.c matstash.c multequal.c axpy.c freespace.h zerodiag.c matstashspace.c<br>
==2289== Memcheck, a memory error detector<br>
==2289== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.<br>
==2289== Using Valgrind-3.5.0 and LibVEX; rerun with -h for copyright info<br>
==2289== Command: bfort -dir /home/balay/tmp/spetsc/src/mat/utils/ftn-auto -mnative -ansi -nomsgs -noprofile -anyname -mapptr -mpi -mpi2 -ferr -ptrprefix Petsc -ptr64 PETSC_USE_POINTER_CONVERSION -fcaps PETSC_HAVE_FORTRAN_CAPS -fuscore PETSC_HAVE_FORTRAN_UNDERSCORE -f90mod_skip_header matio.c convert.c gcreate.c freespace.c getcolv.c ptap.c compressedrow.c matstash.c multequal.c axpy.c freespace.h zerodiag.c matstashspace.c<br>
==2289==<br>
==2289==<br>
==2289== HEAP SUMMARY:<br>
==2289== in use at exit: 5,093 bytes in 50 blocks<br>
==2289== total heap usage: 71 allocs, 21 frees, 17,021 bytes allocated<br>
==2289==<br>
==2289== LEAK SUMMARY:<br>
==2289== definitely lost: 5,093 bytes in 50 blocks<br>
==2289== indirectly lost: 0 bytes in 0 blocks<br>
==2289== possibly lost: 0 bytes in 0 blocks<br>
==2289== still reachable: 0 bytes in 0 blocks<br>
==2289== suppressed: 0 bytes in 0 blocks<br>
==2289== Rerun with --leak-check=full to see details of leaked memory<br>
==2289==<br>
==2289== For counts of detected and suppressed errors, rerun with: -v<br>
==2289== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 6 from 6)<br>
asterix:/home/balay/tmp/spetsc/src/mat/utils><br>
<<<<<<<<<<<br>
<br>
I'll send in the bugrepot to Bill later today - as it gets toggled with the<br>
following change<br>
<div class="im"><br>
<a href="http://petsc.cs.iit.edu/petsc/externalpackages/sowing-1.1.11/rev/e591c037e500" target="_blank">http://petsc.cs.iit.edu/petsc/externalpackages/sowing-1.1.11/rev/e591c037e500</a><br>
<br>
</div>But regarding your machine - somthing changed on it a couple of days<br>
back - thats triggering this issue. You haven't mentioned how I can<br>
reporduce it ['stack smashing detected' mesg etc..]<br>
<div><div></div><div class="h5"><br>
Satish<br>
<br>
On Fri, 25 Dec 2009, Matthew Knepley wrote:<br>
<br>
> Here is the valgrind for your +100 fix:<br>
><br>
> knepley@khan:/PETSc3/petsc/petsc-dev/src/mat/utils$ valgrind<br>
> /PETSc3/petsc/petsc-dev/linux-gnu-cxx-debug/bin/bfort -dir<br>
> /PETSc3/petsc/petsc-dev/src/mat/utils/ftn-auto -mnative -ansi -nomsgs<br>
> -noprofile -anyname -mapptr -mpi -mpi2 -ferr -ptrprefix Petsc -ptr64<br>
> PETSC_USE_POINTER_CONVERSION -fcaps PETSC_HAVE_FORTRAN_CAPS -fuscore<br>
> PETSC_HAVE_FORTRAN_UNDERSCORE -f90mod_skip_header matio.c convert.c<br>
> gcreate.c freespace.c getcolv.c ptap.c compressedrow.c matstash.c<br>
> multequal.c axpy.c freespace.h zerodiag.c matstashspace.c<br>
> ==20868== Memcheck, a memory error detector.<br>
> ==20868== Copyright (C) 2002-2007, and GNU GPL'd, by Julian Seward et al.<br>
> ==20868== Using LibVEX rev 1804, a library for dynamic binary translation.<br>
> ==20868== Copyright (C) 2004-2007, and GNU GPL'd, by OpenWorks LLP.<br>
> ==20868== Using valgrind-3.3.0-Debian, a dynamic binary instrumentation<br>
> framework.<br>
> ==20868== Copyright (C) 2000-2007, and GNU GPL'd, by Julian Seward et al.<br>
> ==20868== For more details, rerun with: -v<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C2BB: PrintBody (bfort.c:1362)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C293: PrintBody (bfort.c:1363)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C5E7: PrintBody (bfort.c:1384)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C396: PrintBody (bfort.c:1385)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C3CE: PrintBody (bfort.c:1387)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C3E9: PrintBody (bfort.c:1387)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Conditional jump or move depends on uninitialised value(s)<br>
> ==20868== at 0x804C589: PrintBody (bfort.c:1406)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Use of uninitialised value of size 4<br>
> ==20868== at 0x40239D8: strlen (mc_replace_strmem.c:242)<br>
> ==20868== by 0x4198127: fputs (in /lib/tls/i686/cmov/<a href="http://libc-2.7.so" target="_blank">libc-2.7.so</a>)<br>
> ==20868== by 0x804C5BE: PrintBody (bfort.c:1408)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== Invalid read of size 1<br>
> ==20868== at 0x40239D8: strlen (mc_replace_strmem.c:242)<br>
> ==20868== by 0x4198127: fputs (in /lib/tls/i686/cmov/<a href="http://libc-2.7.so" target="_blank">libc-2.7.so</a>)<br>
> ==20868== by 0x804C5BE: PrintBody (bfort.c:1408)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868== Address 0x0 is not stack'd, malloc'd or (recently) free'd<br>
> ==20868==<br>
> ==20868== Process terminating with default action of signal 11 (SIGSEGV)<br>
> ==20868== Access not within mapped region at address 0x0<br>
> ==20868== at 0x40239D8: strlen (mc_replace_strmem.c:242)<br>
> ==20868== by 0x4198127: fputs (in /lib/tls/i686/cmov/<a href="http://libc-2.7.so" target="_blank">libc-2.7.so</a>)<br>
> ==20868== by 0x804C5BE: PrintBody (bfort.c:1408)<br>
> ==20868== by 0x804A622: OutputRoutine (bfort.c:575)<br>
> ==20868== by 0x804A0B2: main (bfort.c:475)<br>
> ==20868==<br>
> ==20868== ERROR SUMMARY: 72 errors from 9 contexts (suppressed: 17 from 1)<br>
> ==20868== malloc/free: in use at exit: 1,056 bytes in 3 blocks.<br>
> ==20868== malloc/free: 6 allocs, 3 frees, 2,112 bytes allocated.<br>
> ==20868== For counts of detected errors, rerun with: -v<br>
> ==20868== searching for pointers to 3 not-freed blocks.<br>
> ==20868== checked 243,864 bytes.<br>
> ==20868==<br>
> ==20868== LEAK SUMMARY:<br>
> ==20868== definitely lost: 0 bytes in 0 blocks.<br>
> ==20868== possibly lost: 0 bytes in 0 blocks.<br>
> ==20868== still reachable: 1,056 bytes in 3 blocks.<br>
> ==20868== suppressed: 0 bytes in 0 blocks.<br>
> ==20868== Rerun with --leak-check=full to see details of leaked memory.<br>
> Segmentation fault<br>
><br>
> The problem is that argument lists are just not parsed correctly for<br>
> gcreate.c. You can send that to Bill.<br>
><br>
> Matt<br>
><br>
> On Fri, Dec 25, 2009 at 10:55 AM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>> wrote:<br>
><br>
> > One more thing. If I remove this patch from sowing - the valgrind log<br>
> > is clean.<br>
> ><br>
> ><br>
> > <a href="http://petsc.cs.iit.edu/petsc/externalpackages/sowing-1.1.11/rev/e591c037e500" target="_blank">http://petsc.cs.iit.edu/petsc/externalpackages/sowing-1.1.11/rev/e591c037e500</a><br>
> ><br>
> > Perhaps you can find the bug in this change. If not - I'll send a bug<br>
> > report to Bill.<br>
> ><br>
> > Satish<br>
> ><br>
> > On Fri, 25 Dec 2009, Satish Balay wrote:<br>
> ><br>
> > > Can you send me the valgrind.log - with the patch applied to the<br>
> > > unmodified sowing-1.1.11-a.tar.gz?<br>
> > ><br>
> > > Also the command you are using to generate this log?<br>
> > ><br>
> > ><br>
> > > I've used the following:<br>
> > > valgrind --tool=memcheck -q --log-file=valgrind.log bfort -dir<br>
> > `pwd`/ftn-auto -ansi -nomsgs -noprofile -anyname -mapptr -mpi -mpi2 -ferr<br>
> > -ptrprefix Petsc -ptr64 PETSC_USE_POINTER_CONVERSION -fcaps<br>
> > PETSC_HAVE_FORTRAN_CAPS -fuscore PETSC_HAVE_FORTRAN_UNDERSCORE matrix.c<br>
> > ><br>
> > > Satish<br>
> > ><br>
> > > On Fri, 25 Dec 2009, Matthew Knepley wrote:<br>
> > ><br>
> > > > Valgrind is not clean for me with the change.<br>
> > > ><br>
> > > > Matt<br>
> > > ><br>
> > > > On Fri, Dec 25, 2009 at 10:28 AM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > wrote:<br>
> > > ><br>
> > > > > Well - normally the first step with detecting the bugs is to report<br>
> > > > > them to the author - and ask for a fix..<br>
> > > > ><br>
> > > > > Satish<br>
> > > > ><br>
> > > > > On Fri, 25 Dec 2009, Matthew Knepley wrote:<br>
> > > > ><br>
> > > > > > I can try, but I still think replacement is the only real<br>
> > alternative.<br>
> > > > > This<br>
> > > > > > is not<br>
> > > > > > able to be debugged, or you would not recommend sticking in random<br>
> > > > > numbers<br>
> > > > > > in malloc() and I would be able to see where an SEGV occurs with<br>
> > gdb.<br>
> > > > > ><br>
> > > > > > Matt<br>
> > > > > ><br>
> > > > > > On Fri, Dec 25, 2009 at 10:16 AM, Satish Balay <<a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > > > > wrote:<br>
> > > > > ><br>
> > > > > > > BTW: What linux are you using? ubuntu version? i686 or x86_64?<br>
> > etc...<br>
> > > > > > ><br>
> > > > > > > also try:<br>
> > > > > > ><br>
> > > > > > > arg->name = (char *)MALLOC( strlen(p) + 100 );<br>
> > > > > > ><br>
> > > > > > > satish<br>
> > > > > > ><br>
> > > > > > ><br>
> > > > > > > On Fri, 25 Dec 2009, Satish Balay wrote:<br>
> > > > > > ><br>
> > > > > > > > Did my suggested change not work for you?<br>
> > > > > > > ><br>
> > > > > > > > Satish<br>
> > > > > > > ><br>
> > > > > > > > On Thu, 24 Dec 2009, Matthew Knepley wrote:<br>
> > > > > > > ><br>
> > > > > > > > > I spent a bunch of time on this today. This shit is<br>
> > hopelessly<br>
> > > > > broken.<br>
> > > > > > > It<br>
> > > > > > > > > sucks completely.<br>
> > > > > > > > > I cannot get it to run, nor see why it is causing stack<br>
> > overruns<br>
> > > > > and<br>
> > > > > > > SEGVs.<br>
> > > > > > > > > If anyone does<br>
> > > > > > > > > not think it is hopeless, speak up now. This is a complete<br>
> > fucking<br>
> > > > > > > > > embarrassment.<br>
> > > > > > > > ><br>
> > > > > > > > > Matt<br>
> > > > > > > > ><br>
> > > > > > > > > On Mon, Dec 21, 2009 at 4:42 PM, Matthew Knepley <<br>
> > > > > <a href="mailto:knepley@gmail.com">knepley@gmail.com</a>><br>
> > > > > > > wrote:<br>
> > > > > > > > ><br>
> > > > > > > > > > This does not make any sense to me because it would be a<br>
> > heap<br>
> > > > > > > violation,<br>
> > > > > > > > > > not a stack smash.<br>
> > > > > > > > > ><br>
> > > > > > > > > > Matt<br>
> > > > > > > > > ><br>
> > > > > > > > > ><br>
> > > > > > > > > > On Mon, Dec 21, 2009 at 4:30 PM, Satish Balay <<br>
> > <a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a><br>
> > > > > ><br>
> > > > > > > wrote:<br>
> > > > > > > > > ><br>
> > > > > > > > > >> [I don't know the correct fix for this - but ] The<br>
> > following<br>
> > > > > change<br>
> > > > > > > is<br>
> > > > > > > > > >> getting rid of valgrind messages for me. Maybe you can use<br>
> > this,<br>
> > > > > > > build<br>
> > > > > > > > > >> sowing separately - and continue..<br>
> > > > > > > > > >><br>
> > > > > > > > > >> Satish<br>
> > > > > > > > > >><br>
> > > > > > > > > >> ----------<br>
> > > > > > > > > >><br>
> > > > > > > > > >> diff -r dbe25084c0e4 src/bfort/bfort.c<br>
> > > > > > > > > >> --- a/src/bfort/bfort.c Mon Dec 15 22:20:58 2008 -0600<br>
> > > > > > > > > >> +++ b/src/bfort/bfort.c Mon Dec 21 16:29:09 2009 -0600<br>
> > > > > > > > > >> @@ -2157,7 +2157,7 @@<br>
> > > > > > > > > >><br>
> > > > > > > > > >> /* Current token is name */<br>
> > > > > > > > > >> arg->has_star = (nstar > 0);<br>
> > > > > > > > > >> - arg->name = (char *)MALLOC( strlen(p) + 1 );<br>
> > > > > > > > > >> + arg->name = (char *)MALLOC( strlen(p) + 10 );<br>
> > > > > > > > > >> strcpy( arg->name, p );<br>
> > > > > > > > > >><br>
> > > > > > > > > >> /* We can't output the name just yet, because if it is<br>
> > > > > > > > > >><br>
> > > > > > > > > >><br>
> > > > > > > > > >><br>
> > > > > > > > > >><br>
> > > > > > > > > >> On Mon, 21 Dec 2009, Matthew Knepley wrote:<br>
> > > > > > > > > >><br>
> > > > > > > > > >> > The problem appears to be in OutputRoutine() in bfort.c,<br>
> > but<br>
> > > > > that<br>
> > > > > > > code<br>
> > > > > > > > > >> is<br>
> > > > > > > > > >> > impossible<br>
> > > > > > > > > >> > to debug. I can't see where something is getting<br>
> > overwritten,<br>
> > > > > and<br>
> > > > > > > it<br>
> > > > > > > > > >> looks<br>
> > > > > > > > > >> > like the check<br>
> > > > > > > > > >> > only happens when the routine returns. bfort is such<br>
> > crap.<br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> > Matt<br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> > On Mon, Dec 21, 2009 at 3:25 PM, Matthew Knepley <<br>
> > > > > > > <a href="mailto:knepley@gmail.com">knepley@gmail.com</a>><br>
> > > > > > > > > >> wrote:<br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> > > On Mon, Dec 21, 2009 at 3:21 PM, Satish Balay <<br>
> > > > > > > <a href="mailto:balay@mcs.anl.gov">balay@mcs.anl.gov</a>><br>
> > > > > > > > > >> wrote:<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > >> On Mon, 21 Dec 2009, Lisandro Dalcín wrote:<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > >> > On Mon, Dec 21, 2009 at 5:37 PM, Matthew Knepley <<br>
> > > > > > > > > >> <a href="mailto:knepley@gmail.com">knepley@gmail.com</a>><br>
> > > > > > > > > >> > >> wrote:<br>
> > > > > > > > > >> > >> > ><br>
> > > > > > > > > >> > >> > > It says there is a stack smash and no other info.<br>
> > This<br>
> > > > > is<br>
> > > > > > > > > >> completely<br>
> > > > > > > > > >> > >> fucking<br>
> > > > > > > > > >> > >> > > my development right now.<br>
> > > > > > > > > >> > >> > ><br>
> > > > > > > > > >> > >> ><br>
> > > > > > > > > >> > >> > Any chance bfort was built with -fstack-protector<br>
> > flag?<br>
> > > > > This<br>
> > > > > > > > > >> failure<br>
> > > > > > > > > >> > >> > could could be signaling an actual old bug in<br>
> > bfort... I<br>
> > > > > > > would<br>
> > > > > > > > > >> > >> > re-build bfort with debug and re-run under<br>
> > valgrind...<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > >> That must be it.<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > >> I just ran my build [which is without<br>
> > -fstack-protector] -<br>
> > > > > and<br>
> > > > > > > > > >> > >> valgrind does flag a bunch of things with bfort.<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > > 1) That flag is nowhere in my build.<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > > 2) Something changed<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > > Matt<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > >> I normally install sowing separately and have it in<br>
> > my PATH<br>
> > > > > -<br>
> > > > > > > so that<br>
> > > > > > > > > >> > >> it doesn't have to be rebuilt each time I build<br>
> > petsc.<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > >> I guess we should sync up [our patches] with latest<br>
> > sowing<br>
> > > > > and<br>
> > > > > > > make<br>
> > > > > > > > > >> > >> sure its valgrind clean aswell.<br>
> > > > > > > > > >> > >><br>
> > > > > > > > > >> > >> Satish<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> > > --<br>
> > > > > > > > > >> > > What most experimenters take for granted before they<br>
> > begin<br>
> > > > > their<br>
> > > > > > > > > >> > > experiments is infinitely more interesting than any<br>
> > results<br>
> > > > > to<br>
> > > > > > > which<br>
> > > > > > > > > >> their<br>
> > > > > > > > > >> > > experiments lead.<br>
> > > > > > > > > >> > > -- Norbert Wiener<br>
> > > > > > > > > >> > ><br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> ><br>
> > > > > > > > > >> ><br>
> > > > > > > > > >><br>
> > > > > > > > > ><br>
> > > > > > > > > ><br>
> > > > > > > > > ><br>
> > > > > > > > > > --<br>
> > > > > > > > > > What most experimenters take for granted before they begin<br>
> > their<br>
> > > > > > > > > > experiments is infinitely more interesting than any results<br>
> > to<br>
> > > > > which<br>
> > > > > > > their<br>
> > > > > > > > > > experiments lead.<br>
> > > > > > > > > > -- Norbert Wiener<br>
> > > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > > ><br>
> > > > > > > ><br>
> > > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > > ><br>
> > > > ><br>
> > > ><br>
> > > ><br>
> > > ><br>
> > > ><br>
> > ><br>
> ><br>
><br>
><br>
><br>
> </div></div></blockquote></div><br><br clear="all"><br>-- <br>What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>
-- Norbert Wiener<br>