<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=us-ascii">
<META content="MSHTML 6.00.5730.11" name=GENERATOR></HEAD>
<BODY text=#000000 bgColor=#ffffff>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>Xavier:</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2></FONT></SPAN> </DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>I suspect that if rpc_execute is duplicated, there are
other duplicated symbols</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>and so we matched some other symbol, got another offset,
and at that </FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>point we're headed down the wrong path. It hadn't
occurred to me that</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>there could be conflicts, but in retrospect it makes
sense.</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2></FONT></SPAN> </DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>Not quite sure how to resolve this. I'd hate to trust
that we could reliably</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=189591818-02022007><FONT face=Arial
color=#0000ff size=2>duplicate Linux'es module loading
algorithm.</FONT></SPAN></DIV><BR>
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma size=2><B>From:</B> crash-utility-bounces@redhat.com
[mailto:crash-utility-bounces@redhat.com] <B>On Behalf Of </B>xb<BR><B>Sent:</B>
Friday, February 02, 2007 1:32 AM<BR><B>To:</B>
anderson@redhat.com<BR><B>Cc:</B> Discussion list for crash utility usage,
maintenance and development<BR><B>Subject:</B> Re: [Crash-utility] crash version
4.0-3.18 is available<BR></FONT><BR></DIV>
<DIV></DIV>Hi Dave and all<BR><BR>I still have some problems with module symbols
on ia64. (I built crash 4.0-3.17 + 4.0-3.17-to-4.0-3.18.patch).<BR><BR>It seems
that when I run the commans:<BR><BR>mod -s <module name> <BR><BR>the text
symbol addresses are not the ones provided by kallsyms.<BR>Any idea ?<BR>Thanks
in advance.<BR>Xavier<BR><BR><TT><I>Traces
-------------------------------------------------------------------------------------------------------------------------------------<BR>#
./crash /boot/vmlinux-2.6.18-B64k.1.7<BR><BR>crash 4.0-3.17<BR>Copyright (C)
2002, 2003, 2004, 2005, 2006, 2007 Red Hat, Inc.<BR>...<BR>GNU gdb
6.1<BR>...<BR>This GDB was configured as
"ia64-unknown-linux-gnu"...<BR><BR> KERNEL:
/boot/vmlinux-2.6.18-B64k.1.7<BR> DUMPFILE:
/dev/mem<BR> CPUS:
8<BR> DATE: Fri Feb 2 10:25:27
2007<BR> UPTIME: 13 days, 16:26:51<BR>LOAD
AVERAGE: 0.75, 0.66, 0.58<BR> TASKS:
286<BR> NODENAME: xxx<BR> RELEASE:
2.6.18-B64k.1.7<BR> VERSION: #1 SMP Fri Jan 5 14:05:06
CET 2007<BR> MACHINE: ia64 (1599
Mhz)<BR> MEMORY: 63.9
GB<BR> PID:
17276<BR> COMMAND:
"crash"<BR> TASK:
e000002025600000 [THREAD_INFO:
e000002025600f40]<BR> CPU:
7<BR> STATE: TASK_RUNNING
(ACTIVE)<BR><BR>crash> sym -l |grep rpc_execute<BR>a000000207d34a80 (t)
__rpc_execute<BR>a000000207d351e0 (t) rpc_execute<BR>a000000207d5a590 (?)
rpc_execute<BR>a000000207d5bde0 (r) __ksymtab_rpc_execute<BR>a000000207d5c6d0
(r) __kstrtab_rpc_execute<BR>a000000207d5cab0 (r)
__kcrctab_rpc_execute<BR>crash> disas rpc_execute<BR>No symbol "rpc_execute"
in current context.<BR>crash> mod -s sunrpc<BR>
MODULE
NAME
SIZE OBJECT FILE<BR>a000000207d82280
sunrpc
414976
/lib/modules/2.6.18-B64k.1.7/kernel/net/sunrpc/sunrpc.ko<BR>crash> disas
rpc_execute<BR>Dump of assembler code for function
rpc_execute:<BR>0xa000000207d5ac10
<rpc_execute+0>:
[MII] break.m 0x0<BR>0xa000000207d5ac11
<rpc_execute+1>:
break.i 0x0<BR>0xa000000207d5ac12
<rpc_execute+2>:
break.i 0x0<BR>...<BR>End of assembler dump.<BR>crash> disas
0xa000000207d351e0
0xa000000207d35200
# This one is OK<BR>Dump of assembler code from 0xa000000207d351e0 to
0xa000000207d35200:<BR>0xa000000207d351e0:
[MMI] nop.m
0x0;;<BR>0xa000000207d351e1:
alloc
r34=ar.pfs,5,4,0<BR>0xa000000207d351e2:
mov r33=b0<BR>0xa000000207d351f0:
[MMI] nop.m
0x0<BR>0xa000000207d351f1:
mov
r35=r1<BR>0xa000000207d351f2:
adds r17=208,r32;;<BR>End of assembler
dump.<BR>crash>
</I></TT><BR><BR>Dave Anderson wrote:
<BLOCKQUOTE cite=mid45C214A7.FC47A4E2@redhat.com type="cite"><TT>- Enhancement
to the "mod" command to expand the number of section</TT> <BR><TT>
arguments to the internal "add-symbol-file" command issued to gdb to</TT>
<BR><TT> load the debug data for module objects. On most
architectures, this</TT> <BR><TT> allows the usage of the command
construct "p [module-symbol-name]" to</TT> <BR><TT> print out the module
data structure in the same way that is done for</TT> <BR><TT> kernel
proper data structure names. (<A class=moz-txt-link-abbreviated
href="mailto:castor.fu@3pardata.com">castor.fu@3pardata.com</A>)</TT>
<P><TT>- Two enhancements to significantly speed up the initialization of</TT>
<BR><TT> crash sessions when running against multi-gigabyte xen kernels
or</TT> <BR><TT> xendumps. The cache of
mfn-to-phys_to_machine_mapping page has been</TT> <BR><TT> changed from
a single-mfn-to-phys_to_machine_mapping page format to</TT> <BR><TT>
storing a contiguous-range-of-mfns-to-phys_to_machine_mapping format.</TT>
<BR><TT> This benefit is primarily seen during the "gathering module
symbol</TT> <BR><TT> data" phase. The second change simply
increases the size of the</TT> <BR><TT> pfn-to-xendump-page-offset
cache. (<A class=moz-txt-link-abbreviated
href="mailto:anderson@redhat.com">anderson@redhat.com</A>)</TT> </P>
<P><TT>- Fix for a segmentation violation during the "gathering task
table</TT> <BR><TT> data" phase of initialization if the thread_info
structure of the</TT> <BR><TT> runqueue-advertised active task has been
freed. This has only ever</TT> <BR><TT> been seen in a xendump
created by "xm dump-core -L [guest-domain]".</TT> <BR><TT> (<A
class=moz-txt-link-abbreviated
href="mailto:anderson@redhat.com">anderson@redhat.com</A>)</TT> </P>
<P><TT>- Cosmetic fix to prepend newlines to messages that happen to be</TT>
<BR><TT> generated during any of the "please wait" segments of
initialization.</TT> <BR><TT> (<A class=moz-txt-link-abbreviated
href="mailto:anderson@redhat.com">anderson@redhat.com</A>)</TT> </P>
<P><TT>- Addressed several compiler warnings when using
-D_FORTIFY_SOURCE=2.</TT> <BR><TT> Some are in gdb code that is never
exercised, others were legitimate</TT> <BR><TT> but would require
impossible code paths, but one of them could</TT> <BR><TT> result in
runaway "help -t" output if the kernel was built without</TT> <BR><TT>
IKCONFIG. (<A class=moz-txt-link-abbreviated
href="mailto:bwalle@suse.de">bwalle@suse.de</A>)</TT> </P>
<P><TT>- Fix for the s390x "bt -f" command option, which was displaying
the</TT> <BR><TT> stack as a sequence of 32-bit words which were dumped
"backwards",</TT> <BR><TT> i.e., at the wrong offset. (<A
class=moz-txt-link-abbreviated
href="mailto:krader@us.ibm.com">krader@us.ibm.com</A>)</TT> </P>
<P><TT>- <A
href="http://people.redhat.com/anderson/4.0-3.17-to-4.0-3.18.patch">http://people.redhat.com/anderson/4.0-3.17-to-4.0-3.18.patch</A></TT>
<BR> </P>
<P><TT>Download from: <A
href="http://people.redhat.com/anderson">http://people.redhat.com/anderson</A></TT>
<BR> <BR> <BR> <BR> </P><PRE wrap=""><HR width="90%" SIZE=4>
--
Crash-utility mailing list
<A class=moz-txt-link-abbreviated href="mailto:Crash-utility@redhat.com">Crash-utility@redhat.com</A>
<A class=moz-txt-link-freetext href="https://www.redhat.com/mailman/listinfo/crash-utility">https://www.redhat.com/mailman/listinfo/crash-utility</A>
</PRE></BLOCKQUOTE></BODY></HTML>