|
From: <Joh...@3c...> - 2004-02-04 19:27:33
|
Hi, Jeremy
These errors occur at startup and only on specific programs. Most programs
don't evidence this problem. We use the
ACE framework in most of our programs . The ACE regression tests don't
produce these same errors; it seems that something
specific in our code is provoking the behavior.
Thanks,
John
Jeremy Fitzhardinge <je...@go...> on 02/04/2004 11:58:54 AM
Sent by: Jeremy Fitzhardinge <je...@go...>
To: John Poplett/MW/US/3Com
cc: val...@li...
Subject: Re: [Valgrind-users] Unexpected result running valgrind 2.1 on
RH9
On Tue, 2004-02-03 at 08:11, Joh...@3c... wrote:
> I am new to valgrind and am running 2.1, which I built from sources for
> Redhat 9. I get these error reports from valgrind, which surprisingly
> references the valgrind source itself:
>
> ==27188== Conditional jump or move depends on uninitialised value(s)
> ==27188== at 0x4067A49B: thread_wrapper (vg_libpthread.c:641)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
> ==27188== Thread 2:
> ==27188== Conditional jump or move depends on uninitialised value(s)
> ==27188== at 0x4067A4AF: thread_wrapper (vg_libpthread.c:644)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
> ==27188== Thread 4:
> ==27188== thread_wrapper: invalid attr->__detachstate
> ==27188== at 0x40679C89: pthread_error (vg_libpthread.c:280)
> ==27188== by 0x4067A4A8: thread_wrapper (vg_libpthread.c:643)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
>
> I've had a casual look through the archives and don't see other reports
of
> a similar problem. Please advise.
That looks like it might be related to the libc_freeres stuff. Do these
happen when your program exits, or during its run?
J
|
|
From: <Joh...@3c...> - 2004-02-18 16:35:55
|
Hi, all
For what it's worth, I discovered by accident that I wasn't specify
-lpthread during the link step. Adding this
directive cleared up the problem I reported earlier.
John
---------------------- Forwarded by John Poplett/MW/US/3Com on 02/18/2004
10:29 AM ---------------------------
Jeremy Fitzhardinge <je...@go...> on 02/04/2004 11:58:54 AM
Sent by: Jeremy Fitzhardinge <je...@go...>
To: John Poplett/MW/US/3Com
cc: val...@li...
Subject: Re: [Valgrind-users] Unexpected result running valgrind 2.1 on
RH9
On Tue, 2004-02-03 at 08:11, Joh...@3c... wrote:
> I am new to valgrind and am running 2.1, which I built from sources for
> Redhat 9. I get these error reports from valgrind, which surprisingly
> references the valgrind source itself:
>
> ==27188== Conditional jump or move depends on uninitialised value(s)
> ==27188== at 0x4067A49B: thread_wrapper (vg_libpthread.c:641)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
> ==27188== Thread 2:
> ==27188== Conditional jump or move depends on uninitialised value(s)
> ==27188== at 0x4067A4AF: thread_wrapper (vg_libpthread.c:644)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
> ==27188== Thread 4:
> ==27188== thread_wrapper: invalid attr->__detachstate
> ==27188== at 0x40679C89: pthread_error (vg_libpthread.c:280)
> ==27188== by 0x4067A4A8: thread_wrapper (vg_libpthread.c:643)
> ==27188== by 0x40153E87: do__quit (vg_scheduler.c:1787)
> ==27188==
>
> I've had a casual look through the archives and don't see other reports
of
> a similar problem. Please advise.
That looks like it might be related to the libc_freeres stuff. Do these
happen when your program exits, or during its run?
J
|
|
From: Nicholas N. <nj...@ca...> - 2004-02-19 13:41:26
|
On Wed, 18 Feb 2004 Joh...@3c... wrote: > For what it's worth, I discovered by accident that I wasn't specify > -lpthread during the link step. Adding this > directive cleared up the problem I reported earlier. If your program uses pthreads, how did it link if you didn't specify -lpthread? N |
|
From: Jeremy F. <je...@go...> - 2004-02-19 16:43:47
|
On Thu, 2004-02-19 at 05:35, Nicholas Nethercote wrote: > On Wed, 18 Feb 2004 Joh...@3c... wrote: > > > For what it's worth, I discovered by accident that I wasn't specify > > -lpthread during the link step. Adding this > > directive cleared up the problem I reported earlier. > > If your program uses pthreads, how did it link if you didn't specify > -lpthread? Some other library may have linked against -lpthread, which brought it in late. That's something which even plain glibc pthreads doesn't really support, so its understandable something strange might happen. J |