|
From: xiaoming g. <xia...@gm...> - 2007-03-24 19:53:07
|
Hi, all.
I get the following failed assertion problem when I'm trying to profile some
program
using a new tool based on valgrind.
===================
valgrind: ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63
(vgPlain_get_ThreadState): Assertion 'VG_(threads)[tid].tid == tid' failed.
In get_ThreadState.
===================
I aslo get the following information from debugging it.
===================
#0 vgPlain_get_ThreadState (tid=1)
at ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63
#1 0x38010aa4 in sync_signalhandler (sigNo=11, info=0x8907ab6c,
uc=0x8907abec)
at ../../valgrind-3.2.2-reda/coregrind/m_signals.c:1731
#2 0x3800e500 in vgPlain_redir_notify_new_SegInfo ()
Previous frame inner to this frame (corrupt stack?)
===================
After try my best, I still don't know where is the bug. I use system calls
provided by
valgrind itself. And when I try other programs with this new tool there is
no such a
failed assertion. Please give me some directions if you have the same
problem before.
Xiaoming Gu
|
|
From: Nicholas N. <nj...@cs...> - 2007-03-24 23:48:29
|
On Sat, 24 Mar 2007, xiaoming gu wrote: > I get the following failed assertion problem when I'm trying to profile some > program > using a new tool based on valgrind. > =================== > valgrind: ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63 > (vgPlain_get_ThreadState): Assertion 'VG_(threads)[tid].tid == tid' failed. > In get_ThreadState. > =================== > > I aslo get the following information from debugging it. > =================== > #0 vgPlain_get_ThreadState (tid=1) > at ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63 > #1 0x38010aa4 in sync_signalhandler (sigNo=11, info=0x8907ab6c, > uc=0x8907abec) > at ../../valgrind-3.2.2-reda/coregrind/m_signals.c:1731 > #2 0x3800e500 in vgPlain_redir_notify_new_SegInfo () > Previous frame inner to this frame (corrupt stack?) > =================== > > After try my best, I still don't know where is the bug. I use system calls > provided by > valgrind itself. And when I try other programs with this new tool there is > no such a > failed assertion. Please give me some directions if you have the same > problem before. Have you tried using debugging printfs to find out the incorrect value? Have you tried running other Valgrind tools (--tool=none, memcheck, cachegrind) to see if the same thing happens? Nick |
|
From: Julian S. <js...@ac...> - 2007-03-25 00:00:50
|
Did you make any changes in coregrind/ ? If yes, what are they? J On Saturday 24 March 2007 19:53, xiaoming gu wrote: > Hi, all. > > I get the following failed assertion problem when I'm trying to profile > some program > using a new tool based on valgrind. > =================== > valgrind: ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63 > (vgPlain_get_ThreadState): Assertion 'VG_(threads)[tid].tid == tid' failed. > In get_ThreadState. > =================== > > I aslo get the following information from debugging it. > =================== > #0 vgPlain_get_ThreadState (tid=1) > at ../../valgrind-3.2.2-reda/coregrind/m_threadstate.c:63 > #1 0x38010aa4 in sync_signalhandler (sigNo=11, info=0x8907ab6c, > uc=0x8907abec) > at ../../valgrind-3.2.2-reda/coregrind/m_signals.c:1731 > #2 0x3800e500 in vgPlain_redir_notify_new_SegInfo () > Previous frame inner to this frame (corrupt stack?) > =================== > > After try my best, I still don't know where is the bug. I use system calls > provided by > valgrind itself. And when I try other programs with this new tool there is > no such a > failed assertion. Please give me some directions if you have the same > problem before. > > Xiaoming Gu |