RE: ROOT 5.32.02 segs fault on quit

Thanks Axel!

Cheers, Bertrand.

-----Original Message-----
From: Axel Naumann
Sent: 02 April 2012 20:55
To: Bertrand Bellenot
Cc: Fons Rademakers; pcanal_at_fnal.gov; roottalk (Mailing list discussing all aspects of the ROOT system.); Paul Seyfert; Christopher Rob Jones Subject: Re: [ROOT] ROOT 5.32.02 segs fault on quit

Hi,

Thanks for your report! I have added a workaround to the trunk and v5-32-00-patches.

Cheers, Axel.

On 04/02/2012 03:56 PM, Bertrand Bellenot wrote:
> Hi,
>
> So the offending configure option is: --with-cint-longline=4096
> Here is what it produces:
>
> root [0] new TBrowser (class TBrowser*)0xba1220 root [1] .q
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00002aaaad0b553b in malloc_consolidate () from /lib64/libc.so.6
> (gdb) bt #0 0x00002aaaad0b553b in malloc_consolidate () from
> /lib64/libc.so.6 #1 0x00002aaaad0b7bdc in _int_malloc () from
> /lib64/libc.so.6 #2 0x00002aaaad0b9e4e in malloc () from
> /lib64/libc.so.6 #3 0x00002aaaac74a96d in operator new(unsigned
> long) () from
> /afs/cern.ch/sw/lcg/contrib/gcc/4.3.5/x86_64-slc5-gcc34-opt/lib64/libstdc++.so.6
>
>

#4 0x00002aaaab0e4e18 in TCint::CallFunc_Factory() const ()
> from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #5 0x00002aaaab07ba80 in TQSlot::TQSlot(TClass*, char const*, char
> const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #6 0x00002aaaab07c2ee in TQSlotPool::New(TClass*, char const*,
> char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #7 0x00002aaaab07cada in TQConnection::TQConnection(TClass*,
> void*, char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #8 0x00002aaaab081038 in TQObject::ConnectToClass(TQObject*, char
> const*, TClass*, void*, char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #9 0x00002aaaab081225 in TQObject::Connect(TQObject*, char const*,
> char const*, void*, char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #10 0x00002aaaab0b47ef in TTimer::SingleShot(int, char const*,
> void*, char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #11 0x00002aaab0bccfdf in TGFrame::DeleteWindow() () ---Type
> <return> to continue, or q <return> to quit--- from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libGui.so
> #12 0x00002aaab0c997b9 in TRootBrowser::CloseTabs() () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libGui.so
> #13 0x00002aaaab047333 in TBrowser::~TBrowser() () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #14 0x00002aaaab0d2b64 in TCollection::GarbageCollect(TObject*) ()
> from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #15 0x00002aaaab0d8ed5 in TList::Delete(char const*) () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #16 0x00002aaaab08856f in TROOT::~TROOT() () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #17 0x00002aaaad0787de in __cxa_finalize () from /lib64/libc.so.6
> #18 0x00002aaaab038016 in __do_global_dtors_aux () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #19 0x00000000007862b0 in ?? () #20 0x00007fffffffb740 in ?? () #21
> 0x00002aaaab491e21 in _fini () from
> /afs/cern.ch/user/b/bellenot/scratch0/root.v5.32.02/lib/libCore.so
> #22 0x00007fffffffb740 in ?? () #23 0x00002aaaaaab8a44 in _dl_fini
> () from /lib64/ld-linux-x86-64.so.2 Backtrace stopped: previous
> frame inner to this frame (corrupt stack?)
>
> So if some CINT guru wants to try, he's welcome...
>
> Cheers, Bertrand.
>
> -----Original Message----- From: owner-roottalk_at_root.cern.ch
> [mailto:owner-roottalk_at_root.cern.ch] On Behalf Of Bertrand
> Bellenot Sent: 02 April 2012 15:41 To: Fons Rademakers; Axel
> Naumann; pcanal_at_fnal.gov Cc: roottalk (Mailing list discussing all
> aspects of the ROOT system.); Paul Seyfert; Christopher Rob Jones
> Subject: RE: [ROOT] ROOT 5.32.02 segs fault on quit
>
> Hi,
>
> Interesting enough, it doesn't crash anymore if those options are
> not passed to the configure script: --with-cint-maxstruct=36000
> --with-cint-maxtypedef=36000 --with-cint-longline=4096
>
> Thanks Axel for suggesting it! Then I'll investigate which one is
> actually the real responsible of the crash. Stay tuned...
>
> Cheers, Bertrand.
>
> -----Original Message----- From: owner-roottalk_at_root.cern.ch
> [mailto:owner-roottalk_at_root.cern.ch] On Behalf Of Paul Seyfert
> Sent: 30 March 2012 14:43 To: Christopher Rob Jones Cc: roottalk
> (Mailing list discussing all aspects of the ROOT system.) Subject:
> Re: [ROOT] ROOT 5.32.02 segs fault on quit
>
> Hi,
>
> for the record 1. I could reproduce the crash with self compiled
> 5.30/04 2. The crash can either be avoided by instantiation without
> new: pseyfert_at_robusta ~$ root
> ******************************************* *
> * * W E L C O M E to R O O T * *
> * * Version 5.30/04 8 November 2011 * *
> * * You are welcome to visit our Web site * *
> http://root.cern.ch * *
> * *******************************************
>
> ROOT 5.30/04 (tags/v5-30-04_at_41803, Jan 02 2012, 16:06:00 on
> linuxx8664gcc)
>
> CINT/ROOT C/C++ Interpreter version 5.18.00, July 2, 2010 Type ?
> for help. Commands must be C++ statements. Enclose multiple
> statements between { }. root [0] TBrowser b root [1] .q
> pseyfert_at_robusta ~$
>
> or by deleting the browser before quitting:
>
> pseyfert_at_robusta ~$ root
> ******************************************* *
> * * W E L C O M E to R O O T * *
> * * Version 5.30/04 8 November 2011 * *
> * * You are welcome to visit our Web site * *
> http://root.cern.ch * *
> * *******************************************
>
> ROOT 5.30/04 (tags/v5-30-04_at_41803, Jan 02 2012, 16:06:00 on
> linuxx8664gcc)
>
> CINT/ROOT C/C++ Interpreter version 5.18.00, July 2, 2010 Type ?
> for help. Commands must be C++ statements. Enclose multiple
> statements between { }. root [0] TBrowser* b = new TBrowser() root
> [1] delete b root [2] .q pseyfert_at_robusta ~$
>
> I also saw that root segfaults when I quit it after running some
> bad coding practice macros where other objects were instantiated
> but not deleted.
>
> Cheers, Paul
>
> Am 30.03.2012 14:18, schrieb Chris Jones:
>> Hi,
>
>> I'm running the LCG version of ROOT 5.32.02.
>
>> I've noticed if I just start root, start a TBrowser and then
>> immediately quit, root always seg. faults.
>
>> pciy ~ > which root
>> /afs/cern.ch/sw/lcg/app/releases/ROOT/5.32.02/x86_64-slc5-gcc43-opt/root/bin/root
>
>> pciy ~ > root ******************************************* *
>> * * W E L C O M E to R O O T * *
>> * * Version 5.32/02 27 March 2012 * *
>> * * You are welcome to visit our Web site * *
>> http://root.cern.ch * *
>> * *******************************************
>
>> ROOT 5.32/02 (tags/v5-32-02_at_43514, Mar 27 2012, 22:51:10 on
>> linuxx8664gcc)
>
>> CINT/ROOT C/C++ Interpreter version 5.18.00, July 2, 2010 Type ?
>> for help. Commands must be C++ statements. Enclose multiple
>> statements between { }. root [0] new TBrowser(); root [1] .q
>
>> *** Break *** segmentation violation
>
>> It just hangs there. Eventually I have to kill it by hand.
>
>> The same thing happens with the LHCb ROOT installation under
>> /cvmfs/lhcb.cern.ch
>
>> cheers Chris
>
>
>
Received on Tue Apr 03 2012 - 09:27:07 CEST