|
From: Harry M. <hj...@ta...> - 2005-08-09 23:31:48
|
For some reason, the set of apps I'm currently dealing with (the netcdf operators - aka nco) require valgrind & friends to be tweaked more than other apps that I've used it on. V3 is no different, unfortunately: on a PIII coppermine laptop (IBM a22p), 512M, Debian unstable, app compiled with gcc 3.4.2. ANy ideas? TIA Harry 557 $ valgrind --tool=cachegrind ncks -h -O -v PO2 /home/hjm/nco_bm/skn_lgs.nc /home/hjm/nco_bm/foo.nc ==26895== Cachegrind, an I1/D1/L2 cache profiler. ==26895== Copyright (C) 2002-2005, and GNU GPL'd, by Nicholas Nethercote et al. ==26895== Using LibVEX rev 1313, a library for dynamic binary translation. ==26895== Copyright (C) 2004-2005, and GNU GPL'd, by OpenWorks LLP. ==26895== Using valgrind-3.0.0, a dynamic binary instrumentation framework. ==26895== Copyright (C) 2000-2005, and GNU GPL'd, by Julian Seward et al. ==26895== For more details, rerun with: -v ==26895== @@ don't know what type 'n' is @@ parsing ndom_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;; gave NULL type (dom_access_iterator_tag,bool,ptrdiff_t,bool*,bool&>:(91,4),0,64;_M_p (91,3),0,32;_M_offset:(0,17),32,32;; remains) @@ don't know what type 'o' is @@ parsing om_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;;gave NULL type (m_access_iterator_tag,bool,ptrdiff_t,bool*,bool&>:(91,4),0,64;_M_p (91,3),0,32;_M_offset:(0,17),32,32;; remains) @@ parsing dom_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;; gave NULL type (m_access_iterator_tag,bool,ptrdiff_t,bool*,bool&>:(91,4),0,64;_M_p (91,3),0,32;_M_offset:(0,17),32,32;; remains) @@ expected ';' at range TYPE (remains="_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;;") @@ expected ':' at struct method MANGLE-ARGS (remains="andom_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;;") @@ parsing (91,5)=s8iterator<std::random_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;; gave NULL type (s8iterator<std::random_access_iterator_tag,bool,ptrdiff_t,bool*,bool&> (91,4),0,64;_M_p:(91,3),0,32;_M_offset:(0,17),32,32;; remains) --26895-- INTERNAL ERROR: Valgrind received a signal 11 (SIGSEGV) - exiting --26895-- si_code=1; Faulting address: 0x0; sp: 0xB0B09B7C valgrind: the 'impossible' happened: Killed by fatal signal ==26895== at 0xB002A314: vgModuleLocal_st_basetype (symtypes.c:433) ==26895== by 0xB002EE10: initSym (stabs.c:1101) ==26895== by 0xB002F94C: vgModuleLocal_read_debuginfo_stabs (stabs.c:1677) ==26895== by 0xB0028758: read_lib_symbols (symtab.c:1615) ==26895== by 0xB00288A4: vgPlain_read_seg_symbols (symtab.c:1674) ==26895== by 0xB003B5A4: vgPlain_map_file_segment (aspacemgr.c:816) ==26895== by 0xB003B6D5: vgPlain_map_fd_segment (aspacemgr.c:862) ==26895== by 0xB003C7E2: vgModuleLocal_mmap_segment (syswrap-generic.c:148) ==26895== by 0xB004E812: vgSysWrap_x86_linux_old_mmap_before (syswrap-x86-linux.c:1538) ==26895== by 0xB004FA7A: vgPlain_client_syscall (syswrap-main.c:622) ==26895== by 0xB0038DC1: handle_syscall (scheduler.c:589) ==26895== by 0xB0039056: vgPlain_scheduler (scheduler.c:688) ==26895== by 0xB00503E7: vgModuleLocal_thread_wrapper (syswrap-linux.c:80) ==26895== by 0xB004C7FB: run_a_thread_NORETURN (syswrap-x86-linux.c:150) sched status: running_tid=1 Thread 1: status = VgTs_Runnable ==26895== at 0x3A9764CD: (within /lib/ld-2.3.2.so) ==26895== by 0x3A96B5A5: _dl_map_object (in /lib/ld-2.3.2.so) ==26895== by 0x3A96FB08: (within /lib/ld-2.3.2.so) ==26895== by 0x3A971015: _dl_catch_error (in /lib/ld-2.3.2.so) ==26895== by 0x3A970580: _dl_map_object_deps (in /lib/ld-2.3.2.so) ==26895== by 0x3A966BB1: (within /lib/ld-2.3.2.so) ==26895== by 0x3A9741F6: (within /lib/ld-2.3.2.so) ==26895== by 0x3A965F3A: (within /lib/ld-2.3.2.so) ==26895== by 0x3A965C26: (within /lib/ld-2.3.2.so) -- Cheers, Harry Harry J Mangalam hj...@ta... <<plain text preferred>> |
|
From: Tom H. <to...@co...> - 2005-08-10 06:16:31
|
In message <ddbea5$ckp$1...@se...>
Harry Mangalam <hj...@ta...> wrote:
> For some reason, the set of apps I'm currently dealing with (the netcdf
> operators - aka nco) require valgrind & friends to be tweaked more than
> other apps that I've used it on.
What do you mean by 'tweaked' exactly? Have you got patches that you
have to apply to valgrind to make it work with your code?
> V3 is no different, unfortunately: on a PIII coppermine laptop (IBM a22p),
> 512M, Debian unstable, app compiled with gcc 3.4.2. ANy ideas?
It's a bug. Please raise a bug in the bug tracker for it.
Thanks,
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: Harry M. <hj...@ta...> - 2005-08-11 02:01:38
|
Tom Hughes wrote: > In message <ddbea5$ckp$1...@se...> > Harry Mangalam <hj...@ta...> wrote: > >> For some reason, the set of apps I'm currently dealing with (the netcdf >> operators - aka nco) require valgrind & friends to be tweaked more than >> other apps that I've used it on. > > What do you mean by 'tweaked' exactly? Have you got patches that you > have to apply to valgrind to make it work with your code? I got this working (version 2.4 w/ callgrind ) via some patches supplied by Josef. I doubt that they would work with V3. >> V3 is no different, unfortunately: on a PIII coppermine laptop (IBM >> a22p), >> 512M, Debian unstable, app compiled with gcc 3.4.2. ANy ideas? > > It's a bug. Please raise a bug in the bug tracker for it. Will do, thanks! hjm -- Cheers, Harry Harry J Mangalam hj...@ta... <<plain text preferred>> |