|
From: Christoph B. <bar...@or...> - 2005-10-04 16:12:14
|
Hi, After setting M_PROCMAP_BUF to 500000 the programms runs further than before. But now I get the following error from valgrind. Is there anything I can do? ==VG_(ctime) HACK! 31115== Warning: set address range perms: large range 268439552, a 1, v 1 --VG_(ctime) HACK! 31115-- INTERNAL ERROR: Valgrind received a signal 7 (SIGBUS) - exiting --VG_(ctime) HACK! 31115-- si_code=2; Faulting address: 0x30BCA114; sp: 0x4029BB870 valgrind: the 'impossible' happened: Killed by fatal signal ==VG_(ctime) HACK! 31115== at 0x700C520E: disInstr_AMD64_WRK (toIR.c:7824) ==VG_(ctime) HACK! 31115== by 0x700D4CF4: disInstr_AMD64 (toIR.c:13412) ==VG_(ctime) HACK! 31115== by 0x7008F052: bb_to_IR (bb_to_IR.c:187) ==VG_(ctime) HACK! 31115== by 0x700656F3: LibVEX_Translate (vex_main.c:405) ==VG_(ctime) HACK! 31115== by 0x70023C9D: vgPlain_translate (libvex_basictypes.h:154) ==VG_(ctime) HACK! 31115== by 0x70032FDF: handle_tt_miss (scheduler.c:591) ==VG_(ctime) HACK! 31115== by 0x70033405: vgPlain_scheduler (scheduler.c:712) ==VG_(ctime) HACK! 31115== by 0x700531D2: vgModuleLocal_thread_wrapper (syswrap-linux.c:82) ==VG_(ctime) HACK! 31115== by 0x70044C2D: run_a_thread_NORETURN (syswrap-amd64-linux.c:117) sched status: running_tid=1 Thread 1: status = VgTs_Runnable ==VG_(ctime) HACK! 31115== at 0x30BCA114: BN_INV::Blockages::modify_placement_blockages(List_List*&, xi_shape_t&, BN_INV::Blockages::Modus) (inv_blockages.C:296) ==VG_(ctime) HACK! 31115== by 0x30BCB0D8: BN_INV::Blockages::pingpong() (inv_blockages.C:469) ==VG_(ctime) HACK! 31115== by 0x30BCBCEB: BN_INV::Blockages::Blockages(BN::Mem&, BN::Logger&, BN_OPT::OptBase&, int, int) (inv_blockages.C:689) ==VG_(ctime) HACK! 31115== by 0x30C06F3C: (anonymous namespace)::cmd_blockages(CiCommand*) (invtree_ci.C:816) ==VG_(ctime) HACK! 31115== by 0x491564: CiCommand::operator()() (in /data/bin/nut) ==VG_(ctime) HACK! 31115== by 0x49EA15: ci__call_binding(CiBinding*&, CiLanguage, void**, int, void**) (in /data/bin/nut) ... ==VG_(ctime) HACK! 31115== by 0x468D08: main (in /data/bin/nut) Thread 2: status = VgTs_WaitSys ==VG_(ctime) HACK! 31115== at 0x378BA089AA: pthread_cond_wait@@GLIBC_2.3.2 (in /lib64/tls/libpthread-2.3.4.so) ==VG_(ctime) HACK! 31115== by 0x129E77A4: LongLock (in /data/lib/libd.so) ==VG_(ctime) HACK! 31115== by 0x129EE105: CycleThread (in /data/lib/libd.so) ==VG_(ctime) HACK! 31115== by 0x378BA06139: start_thread (in /lib64/tls/libpthread-2.3.4.so) ==VG_(ctime) HACK! 31115== by 0x378ADC52B2: clone (in /lib64/tls/libc-2.3.4.so) Thread 3: status = VgTs_WaitSys ==VG_(ctime) HACK! 31115== at 0x378BA089AA: pthread_cond_wait@@GLIBC_2.3.2 (in /lib64/tls/libpthread-2.3.4.so) ==VG_(ctime) HACK! 31115== by 0x129E77A4: LongLock (in /data/lib/libd.so) ==VG_(ctime) HACK! 31115== by 0x129EE105: CycleThread (in /data/lib/libd.so) ==VG_(ctime) HACK! 31115== by 0x378BA06139: start_thread (in /lib64/tls/libpthread-2.3.4.so) ==VG_(ctime) HACK! 31115== by 0x378ADC52B2: clone (in /lib64/tls/libc-2.3.4.so) Christoph Bartoschek |
|
From: Julian S. <js...@ac...> - 2005-10-04 20:41:27
|
The usual way to get sigbus on an x86 machine is to have a file mapping where the mapping is longer than the file, and access memory in that overhang area at the end. I wonder if this is somehow related to the shm bugs that Tom just fixed today. Can you svn up and see if the behaviour has changed? J |
|
From: Christoph B. <bar...@or...> - 2005-10-10 07:51:04
|
Am Dienstag, 4. Oktober 2005 22:47 schrieb Julian Seward: > The usual way to get sigbus on an x86 machine is to have a file > mapping where the mapping is longer than the file, and access > memory in that overhang area at the end. > > I wonder if this is somehow related to the shm bugs that Tom > just fixed today. Can you svn up and see if the behaviour > has changed? Yes, the latest version does not have this problem anymore. Christoph |