tag:old.nabble.com,2006:forum-3784Nabble - OProfile2009-11-30T11:20:42ZOProfile is a low-overhead, transparent profiler for Linux, capable of instruction-grain profiling of all processes, shared libraries, the kernel and device drivers, via the hardware performance counters. OProfile home is <a href="http://sourceforge.net/projects/oprofile/" target="_top" rel="nofollow">here</a>.tag:old.nabble.com,2006:post-26580383Re: question about oprofile on andriod mobile phone2009-11-30T11:20:42Z2009-11-30T11:20:42ZMaynard Johnsonhotstuff93 wrote:
<div class='shrinkable-quote'><div class='shrinkable-quote'><br>> hi, I'm new to android world and now I need to use oprofile on android mobile
<br>> phone. However, I got some problems.
<br>> Followed is what I did:
<br>> 1)use adb shell to run this command on the phone
<br>> opcontrol --setup
<br>> opcontrol --event=CPU_CYCLES
<br>> opcontrol --start
<br>> opcontrol --stop
<br>>
<br>> 2)after doing this, I can find the log file in /data/oprofile, and then I
<br>> need to analyse it on Linux(REHL) PC
<br>> I used "adb pull /data/oprofile [dest folder]" to pull the sample files
<br>> from the android phone
<br>>
<br>> 3) then I use opreport
<br>> opreport --session-dir=[dest folder] -p [symbols file]
<br>> But I get errors like this
<br>> opreport error: Invalid sample file, bad magic number:
</div>Probably a different ABI format for the sample file on your PC compared to
</div>android. The opimport program can be used to convert it. Attached is a script
<br>I've used to convert all sample files in a profile collection.
<br><br>-Maynard
<br>> android/external/oprofile/result_3/samples/current/{kern}/no-vmlinux/{dep}/{kern}/no-vmlinux/CPU_CYCLES.150000.0.all.all.all
<br>>
<br>> I found several posts here which is quite similar with it. But still don't
<br>> know exactly how to solve it. Can anyone kindly help me? Thanks in advance.
<br>>
<br>>
<br>>
<br>>
<br><br />#!/bin/bash
<br><br># NOTE: Start from the "samples" directory containing the "current" directory
<br># to be imported
<br><br>mkdir current-imported
<br>cd current-imported; (cd ../current; find . -type d ! -name .) |xargs mkdir
<br>cd ../current; find . -type f | while read line; do opimport -a /var/lib/oprofile/abi -o ../current-imported/$line $line; done
<br><br><br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26580383&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26577312RE: Got 'unknown' in oprofile/xenoprofile report2009-11-30T08:12:41Z2009-11-30T08:12:41ZSantos, Jose Renato Gxen-unknown corresponds to samples for code in xen which is not mapped to a symbol by the Xen image.
<br>This is usually associated with dynamic code executed from the stack which xen uses on hypercall entry and exit.
<br>That is normal behavior
<br><br>Renato
<br><div class='shrinkable-quote'><br>> -----Original Message-----
<br>> From: Joe Hewitt [mailto:<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26577312&i=0" target="_top" rel="nofollow">joe.z.hewitt@...</a>]
<br>> Sent: Sunday, November 29, 2009 7:24 PM
<br>> To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26577312&i=1" target="_top" rel="nofollow">oprofile-list@...</a>
<br>> Subject: Got 'unknown' in oprofile/xenoprofile report
<br>>
<br>> Hi,
<br>> I'm using oprofile 0.9.5 with xenoprofile patch. I did some
<br>> passive profiling and noticed 'xen-unknown' in my report but
<br>> don't have a clue how to explain it.
<br>>
<br>> Host OpenSuse 11.1, kernel 2.6.27.37, xen 3.3.1 Guest
<br>> OpenSuse 11.1, kernel 2.6.27.29
<br>>
<br>> Commands issued:
<br>> Dom0> opcontrol --reset
<br>> Dom0> opcontrol --start-daemon
<br>> --xen=/boot/xen-syms-3.3.1_18546_20-0.1.1
<br>> --vmlinux=/boot/vmlinux-2.6.27.37-0.1-xen --passive-domains=1
<br>> --passive-images=/boot/vmlinux-2.6.27.29-0.1-xen-Dom1
<br>> Dom0> opcontrol --start
<br>> Then do something
<br>> Dom0> opcontrol --shutdown
<br>> Dom0> opreport > myreport
<br>>
<br>> part of myreport:
<br>>
<br>> CPU: Core 2, speed 2660.08 MHz (estimated) Counted
<br>> CPU_CLK_UNHALTED events (Clock cycles when not halted) with a
<br>> unit mask of 0x00 (Unhalted core cycles) count 1000000
<br>> CPU_CLK_UNHALT...|
<br>> samples| %|
<br>> ------------------
<br>> 6609 35.0053 domain1-kernel
<br>> 4795 25.3972 domain1-xen
<br>> 4229 22.3994 domain1-apps
<br>> 458 2.4258 vmlinux-2.6.27.37-0.1-xen
<br>> 358 1.8962 domain1-xen-unknown
<br>> 208 1.1017 domain1-modules
<br>> 193 1.0222 xen-syms-3.3.1_18546_20-0.1.1
<br>> 24 0.1271 xen-unknown
<br>> ......
<br>>
<br>> How can these 'XX-unknown' come?
<br>> 358 1.8962 domain1-xen-unknown
<br>> and
<br>> 24 0.1271 xen-unknown
<br>>
<br>> Do i miss some image or symbol names which I should pass to opcontrol?
<br>>
<br>> cat /var/lib/oprofile/sampels/oprofiled.log
<br>> ----
<br>> oprofiled started Mon Nov 30 10:58:34 2009 kernel pointer size: 8
<br>>
<br>> Mon Nov 30 10:58:59 2009
<br>>
<br>>
<br>> -- OProfile Statistics --
<br>> Nr. sample dumps: 4
<br>> Nr. non-backtrace samples: 18788
<br>> Nr. kernel samples: 4729
<br>> Nr. lost samples (no kernel/user): 0
<br>> Nr. lost kernel samples: 0
<br>> Nr. incomplete code structs: 0
<br>> Nr. samples lost due to sample file open failure: 0 Nr.
<br>> samples lost due to no permanent mapping: 0 Nr. event lost
<br>> due to buffer overflow: 0 Nr. samples lost due to no mapping:
<br>> 1 Nr. backtraces skipped due to no file mapping: 0 Nr.
<br>> samples lost due to no mm: 0
<br>>
<br>> ---- Statistics for cpu : 1
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 18882
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost invalid pc: 0
<br>>
<br>> ---- Statistics for cpu : 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost invalid pc: 0
<br>> oprofiled stopped Mon Nov 30 10:58:59 2009
<br>>
<br>> ----
<br>>
<br>> Thanks in advance.
<br>>
<br>> --
<br>> Joe H
<br>>
<br>> --------------------------------------------------------------
<br>> ----------------
<br>> Let Crystal Reports handle the reporting - Free Crystal
<br>> Reports 2008 30-Day trial. Simplify your report design,
<br>> integration and deployment - and focus on what you do best,
<br>> core application coding. Discover what's new with Crystal
<br>> Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a>
<br>> _______________________________________________
<br>> oprofile-list mailing list
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26577312&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br>> <a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br>>
</div>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26577312&i=3" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26569093Got 'unknown' in oprofile/xenoprofile report2009-11-29T19:23:50Z2009-11-29T19:23:50ZJoe HewittHi,
<br>I'm using oprofile 0.9.5 with xenoprofile patch. I did some passive
<br>profiling and noticed 'xen-unknown' in my report but don't have a clue
<br>how to explain it.
<br><br>Host OpenSuse 11.1, kernel 2.6.27.37, xen 3.3.1
<br>Guest OpenSuse 11.1, kernel 2.6.27.29
<br><br>Commands issued:
<br>Dom0> opcontrol --reset
<br>Dom0> opcontrol --start-daemon
<br>--xen=/boot/xen-syms-3.3.1_18546_20-0.1.1
<br>--vmlinux=/boot/vmlinux-2.6.27.37-0.1-xen --passive-domains=1
<br>--passive-images=/boot/vmlinux-2.6.27.29-0.1-xen-Dom1
<br>Dom0> opcontrol --start
<br>Then do something
<br>Dom0> opcontrol --shutdown
<br>Dom0> opreport > myreport
<br><br>part of myreport:
<br><br>CPU: Core 2, speed 2660.08 MHz (estimated)
<br>Counted CPU_CLK_UNHALTED events (Clock cycles when not halted) with a
<br>unit mask of 0x00 (Unhalted core cycles) count 1000000
<br>CPU_CLK_UNHALT...|
<br> samples| %|
<br>------------------
<br> 6609 35.0053 domain1-kernel
<br> 4795 25.3972 domain1-xen
<br> 4229 22.3994 domain1-apps
<br> 458 2.4258 vmlinux-2.6.27.37-0.1-xen
<br> 358 1.8962 domain1-xen-unknown
<br> 208 1.1017 domain1-modules
<br> 193 1.0222 xen-syms-3.3.1_18546_20-0.1.1
<br> 24 0.1271 xen-unknown
<br>......
<br><br>How can these 'XX-unknown' come?
<br>358 1.8962 domain1-xen-unknown
<br>and
<br>24 0.1271 xen-unknown
<br><br>Do i miss some image or symbol names which I should pass to opcontrol?
<br><br>cat /var/lib/oprofile/sampels/oprofiled.log
<br>----
<br>oprofiled started Mon Nov 30 10:58:34 2009
<br>kernel pointer size: 8
<br><br>Mon Nov 30 10:58:59 2009
<br><br><br>-- OProfile Statistics --
<br>Nr. sample dumps: 4
<br>Nr. non-backtrace samples: 18788
<br>Nr. kernel samples: 4729
<br>Nr. lost samples (no kernel/user): 0
<br>Nr. lost kernel samples: 0
<br>Nr. incomplete code structs: 0
<br>Nr. samples lost due to sample file open failure: 0
<br>Nr. samples lost due to no permanent mapping: 0
<br>Nr. event lost due to buffer overflow: 0
<br>Nr. samples lost due to no mapping: 1
<br>Nr. backtraces skipped due to no file mapping: 0
<br>Nr. samples lost due to no mm: 0
<br><br>---- Statistics for cpu : 1
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 18882
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><br>---- Statistics for cpu : 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br>oprofiled stopped Mon Nov 30 10:58:59 2009
<br><br>----
<br><br>Thanks in advance.
<br><br>--
<br>Joe H
<br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26569093&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26535858question about oprofile on andriod mobile phone2009-11-26T19:16:12Z2009-11-27T00:04:25Zhotstuff93hi, I'm new to android world and now I need to use oprofile on android mobile phone. However, I got some problems.
<br>Followed is what I did:
<br>1)use adb shell to run this command on the phone
<br>opcontrol --setup
<br>opcontrol --event=CPU_CYCLES
<br>opcontrol --start
<br>opcontrol --stop
<br><br>During this step, I can check CPU type by
<br># cat /dev/oprofile/cpu_type
<br> -----
<br> arm/armv7
<br> -----
<br><br><br>2)And then, I tried to use opimport_pull to get the result, but following error shows:
<br><br>oprofile: could not open unit mask description file android/prebuilt/linux-x86/oprofile/<b>ppc64</b>/power5++/unit_masks
<br><br>It's quite strange why it's trying to find ppc64 unit_masks because the log data was generated on Android phone. And since no such folder in PC, this step will fail.
<br><br>I've also tried to use opreport directly, the result is
<br>$ ./opreport --session-dir LINUX/android/external/oprofile/result -p LINUX/android/out/target/product/afon/symbols/
<br><br>Could not locate event 255
<br>Could not locate event 255
<br>CPU: <b>ppc64 POWER5++, speed 0 MHz (estimated)</b><br> CPU_CYCLES:15000|
<br> samples| %|
<br>------------------
<br> 1 100.000 no-vmlinux
<br><br><br>Can anyone kindly help me? Thanks in advance.
<br><br><br><br> <p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26530850回复： RTC mode with a 2.6 kernel?2009-11-26T04:48:12Z2009-11-26T04:48:12Zzhouye<div style="font-family:Verdana;font-size:14px;color:#000;"><DIV>I am sorry that I can not help you, as I am also a freshman with XenOprf. I just use it in doing some test and still have lots of questions!So sorry!</DIV><DIV>Maybe you can send your problem to the maillist, So that J. Renato Santos who designed XenOprf also can help you!</DIV><div> </div><div><sign signid="1"><div style="color:#909090;font-family:Arial Narrow;font-size:12px">------------------</div><div style='font-size:14px;font-family:Verdana;color:#000;'><div>
<div>Best Regarding！</div>
<div> Ye Zhou</div>
<div>--------------------------------------------------------</div>
<div>Cluster and Grid Computing Lab</div>
<div>Services Computing Technology and System Lab</div>
<div>Department of Computer Science</div>
<div>Huazhong University of Science and Technology Wuhan, 430074, China</div>
<div>Mobile:+86-13437170320</div><br /></div></div></sign></div><div><qzone signid=""></qzone></div><div><taotao signid=""></taotao></div><div> </div><div style="font:Verdana normal 14px;color:#000;"><div> </div><div> </div><div style="FONT-SIZE: 12px;FONT-FAMILY: Arial Narrow;padding:2px 0 2px 0;">------------------ 原始邮件 ------------------</div><div style="FONT-SIZE: 12px;"><div><b>发件人:</b> "jobhunts02"<<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=0" target="_top" rel="nofollow">jobhunts02@...</a>>;</div><div><b>发送时间:</b> 2009年11月25日(星期三) 凌晨5:16</div><div><b>收件人:</b> "<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=1" target="_top" rel="nofollow">jobhunts02@...</a>"<<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=2" target="_top" rel="nofollow">jobhunts02@...</a>>;</div><div><b>主题:</b> Re: RTC mode with a 2.6 kernel?</div></div><div> </div><DIV>Can RTC mode be used with the 2.6 kernel?<BR></DIV>
<DIV><BR>On Nov 18, 2009, at 6:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=3" target="_top" rel="nofollow">jobhunts02@...</a>wrote:<BR><BR></DIV>
<DIV></DIV>
<BLOCKQUOTE type="cite">
<DIV>
<DIV></DIV>
<div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I am running Linux 2.6.26 on an MPC8248 (which does not provide support to use the hardware performance counters ). HZ=250. Thus, I am getting samples every 4 ms. <o:p /></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p /> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I would like finer resolution. Can I use RTC mode with my 2.6 kernel? If so, how? When I try to force it, I get:<o:p /></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p /> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --event=RTC_INTERRUPTS:4096<o:p /></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">You cannot specify any performance counter events<o:p /></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">because OProfile is in timer mode.</span></p></span><br /></div>= </DIV></Q>
<BLOCKQUOTE type="cite">
<DIV><SPAN>------------------------------------------------------------------------------</SPAN>
<br /><SPAN>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day </SPAN>
<br /><SPAN>trial. Simplify your report design, integration and deployment - and focus on </SPAN>
<br /><SPAN>what you do best, core application coding. Discover what's new with</SPAN>
<br /><SPAN>Crystal Reports now. <A href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow"><A href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</A></A></SPAN>
<br /></DIV></Q>
<BLOCKQUOTE type="cite">
<DIV><SPAN>_______________________________________________</SPAN>
<br /><SPAN>oprofile-list mailing list</SPAN>
<br /><SPAN><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=4" target="_top" rel="nofollow">oprofile-list@...</a></SPAN>
<br /><SPAN><A href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</A></SPAN>
<br /></DIV></Q>= </BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></div></div><br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26530850&i=5" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26521706Issue with running oprofile on PB ARM 11 MPcore2009-11-25T14:51:48Z2009-11-25T14:51:48ZVikas Kumar-4Hi,<div><br></div><div>am getting following error:</div><div><br></div><div><div> ./opcontrol --vmlinux=$PWD/vmlinux</div><div>./ophelp: /root/install/lib/libpopt.so.0: no version information available (required by ./ophelp)</div>
<div>cpu_type 'unset' is not valid</div><div>you should upgrade oprofile or force the use of timer mode</div><div><br></div><div>I am using oprofile-0.9.6 and Linux AEL 2.6.28-arm1 #5 SMP Tue Nov 24 16:49:38 EST 2009 armv6l unknown</div>
<div><br></div><div>Please help resolve this issue.</div><div><br></div><div>Thanks</div><div>Vikas</div>
</div>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26521706&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26519906Re: [PATCH-V3] adds support for loongson22009-11-25T12:36:10Z2009-11-25T12:36:10ZMaynard JohnsonChen Jie wrote:
<div class='shrinkable-quote'><br>> Adds oprofile userland support for ICT loongson2 processor.
<br>> Signed-off-by Chen Jie <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26519906&i=0" target="_top" rel="nofollow">chenj@...</a>>
<br>>
<br>> This is the third version, changes since last version:
<br>> * unify the cpu's identifier, rename from "godson2" to "loongson2" (This
<br>> needs kernel support:
<br>> <a href="http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=55f4e1d4fe88f16f773ba5bb31ce8c1ae63724af" target="_top" rel="nofollow">http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=55f4e1d4fe88f16f773ba5bb31ce8c1ae63724af</a><br>> )
<br>>
<br>> The 'make distcheck' test is ok.
</div>Patch committed. Thanks.
<br><br>-Maynard
<br>>
<br>>
<br>> Regards,
<br>> - Chenj
<br>>
<br>>
<br><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26519906&i=1" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26503488Re: RTC mode with a 2.6 kernel?2009-11-24T13:16:52Z2009-11-24T13:16:52ZWashington Ratso<html><body bgcolor="#FFFFFF"><div>Can RTC mode be used with the 2.6 kernel?<br></div><div><br>On Nov 18, 2009, at 6:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26503488&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:<br><br></div><div></div><blockquote type="cite"><div><div></div><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I am running Linux 2.6.26 on an MPC8248 (which does not provide support to use the hardware performance counters ). HZ=250. Thus, I am getting samples every 4 ms. <o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I would like finer resolution. Can I use RTC mode with my 2.6 kernel? If so, how? When I try to force it, I get:<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --event=RTC_INTERRUPTS:4096<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">You cannot specify any performance counter events<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">because OProfile is in timer mode.</span></p></span><br></div>=
</div></blockquote><blockquote type="cite"><div><span>------------------------------------------------------------------------------</span><br><span>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day </span><br><span>trial. Simplify your report design, integration and deployment - and focus on </span><br><span>what you do best, core application coding. Discover what's new with</span><br><span>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow"><a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a></a></span><br></div></blockquote><blockquote type="cite"><div><span>_______________________________________________</span><br><span>oprofile-list mailing list</span><br><span><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26503488&i=1" target="_top" rel="nofollow">oprofile-list@...</a></span><br><span><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a></span><br></div></blockquote></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26503488&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26501086OProfile 0.9.6 GA release announcement2009-11-24T10:26:29Z2009-11-24T10:26:29ZMaynard JohnsonWe are pleased to announce OProfile 0.9.6. This is mainly a bug-fix release to
<br>correct some regressions that occurred in 0.9.5. It is highly recommended that
<br>all OProfile 0.9.5 users upgrade to 0.9.6. You can download this release at:
<br> <a href="http://sourceforge.net/projects/oprofile/files/" target="_top" rel="nofollow">http://sourceforge.net/projects/oprofile/files/</a><br><br>See the release notes for more details.
<br><br>Thanks.
<br><br>-Maynard Johnson
<br><br>-----------------------------------------------------------------
<br><br>Release Notes
<br>===============
<br>OProfile is a powerful system-wide profiler for Linux. Read
<br>more at <a href="http://oprofile.sf.net" target="_top" rel="nofollow">http://oprofile.sf.net</a><br><br>OProfile 0.9.6 has been released. OProfile is still in alpha,
<br>but has been proven stable for many users.
<br><br><br>New features
<br>------------
<br><br>None -- this is a bug-fix release.
<br><br><br>Bug fixes
<br>---------
<br><br>The main reason for this bug-fix release is to fix three regression bugs that
<br>occurred in 0.9.5:
<br><br> 1. bug #2845063: opcontrol cannot start daemon in timer mode
<br> 2. Improper handling of separate debuginfo files
<br> 3. XML callgraph output has no symbol-level sample counts
<br><br>Other fixes made in this release are:
<br><br>-Fix ophelp -X output for IBS events
<br><br>-Fix ophelp output for "counter:cpuid" and "ext:xxx" case
<br><br>-Fix opcontrol to prevent "Stopping profiling" message when profiler isn't running
<br><br>-Fix opcontrol to kill daemon only if it's running
<br><br>-Fix buffer overflows in xml generator, which can result in segfault
<br><br>-Add extra #includes in agents/jvmpi code since newer libstdc++ header files
<br> do not include standard C header files
<br><br>-Update events and unitmasks for AMD Family10 from publication 31116 revision
<br> 3.34 and fixes
<br><br>-Update events and groups for IBM POWER7
<br><br>-Change element definition for callers and callees in opreport.xsd and bumped
<br> schema version to "3.0"
<br><br>-Fix start-daemon problem on ia64
<br><br><br>Known problems
<br>--------------
<br><br>If NMI watchdog is not disabled (x86), opcontrol may fail to
<br>allocate the hardware performance counter(s) it needs (bug #1468916).
<br><br>configure script fails to handle --without-java correctly (bug #2153620).
<br><br>When profiling stripped libraries where OProfile needs to access a debuginfo
<br>file, it's possible that src filename/line # information may be incorrect
<br>(bug #2573008).
<br><br>Many Alpha ev67 events do not work (bug #931875).
<br><br>For 2.2 kernels, the module must be compiled as the same user
<br>that owns the kernel source tree.
<br><br><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26501086&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26499616[ oprofile-Bugs-2845063 ] opcontrol cannot start daemon in timer mode2009-11-24T09:00:56Z2009-11-24T09:00:56ZSourceForge.netBugs item #2845063, was opened at 2009-08-26 12:11
<br>Message generated for change (Settings changed) made by maynardj
<br>You can respond by visiting:
<br><a href="https://sourceforge.net/tracker/?func=detail&atid=116191&aid=2845063&group_id=16191" target="_top" rel="nofollow">https://sourceforge.net/tracker/?func=detail&atid=116191&aid=2845063&group_id=16191</a><br><br>Please note that this message will contain a full copy of the comment thread,
<br>including the initial issue submission, for this request,
<br>not just the latest update.
<br>Category: None
<br>Group: None
<br>>Status: Closed
<br>Resolution: Fixed
<br>Priority: 5
<br>Private: No
<br>Submitted By: Tom Donovan (tomdonovan)
<br>Assigned to: Maynard Johnson (maynardj)
<br>Summary: opcontrol cannot start daemon in timer mode
<br><br>Initial Comment:
<br>opcontrol --start in timer mode shows "oprofiled: no events specified." error.
<br><br>OProfile 0.9.5 - ARM/Linux Debian 5 "Lenny". Boot arg oprofile.timer=1
<br>Running in timer mode due to Cortex-a8 hw bug with perf counters. OMAP3 cpu.
<br><br>Proposed patch:
<br>--- opcontrol 2009-08-25 17:28:23.000000000 -0400
<br>+++ /usr/local/bin/opcontrol 2009-07-10 16:03:40.000000000 -0400
<br>@@ -1300,7 +1300,6 @@
<br> fi
<br>
<br> if test "$IS_TIMER" = 1; then
<br>+ OPROFILED_EVENTS=TIMER_INT
<br> return
<br> fi
<br>
<br><br><br>----------------------------------------------------------------------
<br><br>Comment By: Maynard Johnson (maynardj)
<br>Date: 2009-10-19 09:23
<br><br>Message:
<br>Sorry for the delay in responding to this bugzilla. Hopefully, you've
<br>noticed that this bug was fixed with a patch posted on Sept 10, subject
<br>"Re: [PATCH1/1] Fix Timer mode".
<br><br>----------------------------------------------------------------------
<br><br>You can respond by visiting:
<br><a href="https://sourceforge.net/tracker/?func=detail&atid=116191&aid=2845063&group_id=16191" target="_top" rel="nofollow">https://sourceforge.net/tracker/?func=detail&atid=116191&aid=2845063&group_id=16191</a><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26499616&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26486170Re: Understanding oprofiled.log2009-11-23T13:23:35Z2009-11-23T13:23:35ZMaynard JohnsonRobert Peter wrote:
<br>> Hello!
<br>>
<br>> I have some questions concerning oprofile's logfile "oprofiled.log".
<br>> Looking at the log above, I don't understand why the "samples lost cpu
<br>I guess you mean "the log *below*" . . . :-)
<br>> buffer overflow" is issued several times.
<br>> Where can I see how many samples were actually recorded, and how many
<br>> samples were actually lost?
<br>>
<br>> In that cas, is "Lost samples = 1514776 + 1391176"? That would be nearly
<br>> as many as the "Nr. non-backtrace samples" (is that the number of
<br>> overall recorded samples?)
<br>The oprofile kernel driver uses a double-buffering scheme . . . a set of per-cpu buffers where samples are initially collected, and a main event buffer to which the per-cpu samples are periodically transferred. The the oprofile userspace daemon reads the sample data from the main event buffer. The first group of statistics below (ending with "Nr. samples lost due to no mm") is for the main event buffer; then you have a group of statistics printed for each active cpu on your system. You've got a *lot* of overflows. As I said to luludede in an earlier posting to this list (see subject "Re: Nr. lost kernel samples"), you need to lower your sampling frequency.
<br><br>-Maynard
<div class='shrinkable-quote'><br>>
<br>> Thanks in advance!
<br>> Robert
<br>>
<br>> Nr. sample dumps: 1770
<br>> Nr. non-backtrace samples: 2891740
<br>> Nr. kernel samples: 1455610
<br>> Nr. lost samples (no kernel/user): 0
<br>> Nr. lost kernel samples: 0
<br>> Nr. incomplete code structs: 18
<br>> Nr. samples lost due to sample file open failure: 0
<br>> Nr. samples lost due to no permanent mapping: 34
<br>> Nr. event lost due to buffer overflow: 6243466
<br>> Nr. samples lost due to no mapping: 1241
<br>> Nr. backtraces skipped due to no file mapping: 2201
<br>> Nr. samples lost due to no mm: 960
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 0
<br>> Nr. samples received: 0
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost cpu buffer overflow: 1514776
<br>> Nr. samples received: 2825879
<br>> Nr. backtrace aborted: 39
<br>> Nr. samples lost cpu buffer overflow: 1391176
<br>> Nr. samples received: 2702601
<br>> Nr. backtrace aborted: 37
<br>>
<br>> ------------------------------------------------------------------------------
<br>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>> trial. Simplify your report design, integration and deployment - and focus on
<br>> what you do best, core application coding. Discover what's new with
<br>> Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>> _______________________________________________
<br>> oprofile-list mailing list
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26486170&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br>> <a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a></div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26486170&i=1" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26480981Understanding oprofiled.log2009-11-23T08:08:39Z2009-11-23T08:08:39ZBugzilla from robert1peter@web.deHello!
<br><br>I have some questions concerning oprofile's logfile "oprofiled.log".
<br>Looking at the log above, I don't understand why the "samples lost cpu
<br>buffer overflow" is issued several times.
<br>Where can I see how many samples were actually recorded, and how many
<br>samples were actually lost?
<br><br>In that cas, is "Lost samples = 1514776 + 1391176"? That would be nearly
<br>as many as the "Nr. non-backtrace samples" (is that the number of
<br>overall recorded samples?)
<br><br>Thanks in advance!
<br>Robert
<br><br>Nr. sample dumps: 1770
<br>Nr. non-backtrace samples: 2891740
<br>Nr. kernel samples: 1455610
<br>Nr. lost samples (no kernel/user): 0
<br>Nr. lost kernel samples: 0
<br>Nr. incomplete code structs: 18
<br>Nr. samples lost due to sample file open failure: 0
<br>Nr. samples lost due to no permanent mapping: 34
<br>Nr. event lost due to buffer overflow: 6243466
<br>Nr. samples lost due to no mapping: 1241
<br>Nr. backtraces skipped due to no file mapping: 2201
<br>Nr. samples lost due to no mm: 960
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 0
<br>Nr. samples received: 0
<br>Nr. backtrace aborted: 0
<br>Nr. samples lost cpu buffer overflow: 1514776
<br>Nr. samples received: 2825879
<br>Nr. backtrace aborted: 39
<br>Nr. samples lost cpu buffer overflow: 1391176
<br>Nr. samples received: 2702601
<br>Nr. backtrace aborted: 37
<br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26480981&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26479726Re: Nr. lost kernel samples2009-11-23T07:06:10Z2009-11-23T07:06:10ZMaynard Johnsonluludede wrote:
<br>> Hi,
<br>>
<br>Is there a reason you double-posted this problem? One posting per problem, please.
<br><br>> I kept getting "Nr. samples lost cpu buffer overflow" when I use oprofile
<br>> 0.9.6-rc1 to profile the kernel.
<br>I presume you're referring to the message "WARNING! The OProfile kernel driver reports sample buffer overflows." This message was added in 0.9.5 -- it's not new in 0.9.6.
<br>>
<br>> Although I increased the "CPU buffer size" to be far greater than the (Nr.
<br>> samples lost cpu buffer overflow) plus (Nr. samples received).
<br>>
<br>> Is it a bug with oprofile or I did something wrong? Can anyone give me some
<br>> suggestion? Thans a lot !
<br>No, it's not a bug. As the rest of the warning message recommends, you should lower your sampling frequency. You're taking a sample every 100000 GLOBAL_POWER_EVENTS, which is a fairly high rate by itself. But when you throw in the fact that you're also asking oprofile to collect callgraph information on each sample, it simply can't keep up. A bigger buffer can help to a certain extent, but I suggest going back to your original buffer size (rm /root/.oprofile/daemonrc to remove the cached profiling values), and then double or triple the number of events per sample to lower your sampling frequency. If your ratio of overflows to non-backtrace samples is less than, say, 1 or 2%, you probably don't need to be concerned about it. Otherwise, you can adjust sampling frequency and/or buffer sizes.
<br><br>NOTE: The buffer watershed value is for the kernel-to-oprofile daemon event buffer, not the per-cpu buffers (which are internal to the oprofile kernel driver. And the value you've specified below is not a good one. As the man page states, it should be set between 0.25 and 0.5 of buffer size. If you're interested to know more about oprofile's double-buffering scheme, see <a href="http://oprofile.sourceforge.net/doc/internals/core-structure.html#id4685015" target="_top" rel="nofollow">http://oprofile.sourceforge.net/doc/internals/core-structure.html#id4685015</a><br><br>-Maynard
<div class='shrinkable-quote'><br>>
<br>> --------------------------------
<br>> steps I took for profiling:
<br>> # opcontrol --reset
<br>> # opcontrol --start
<br>> # my bench mark
<br>> # opcontrol --dump
<br>> # opcontrol --stop
<br>> --------------------------------
<br>> the profiler setting:
<br>> # opcontrol --status
<br>> Daemon running: pid 11302
<br>> Event 0: GLOBAL_POWER_EVENTS:100000:1:1:1
<br>> Separate options: none
<br>> vmlinux file:
<br>> /usr/lib/debug/lib/modules/2.6.29.4-167.fc11.i686.PAEdebug/vmlinux
<br>> Image filter: none
<br>> Call-graph depth: 10
<br>> Buffer size: 80000000
<br>> CPU buffer watershed: 40000000
<br>> CPU buffer size: 400000000
<br>> ----------------------------------
<br>> -- OProfile Statistics --
<br>> Nr. sample dumps: 128
<br>> Nr. non-backtrace samples: 3153835
<br>> Nr. kernel samples: 3066290
<br>> Nr. lost samples (no kernel/user): 0
<br>> Nr. lost kernel samples: 0
<br>> Nr. incomplete code structs: 0
<br>> Nr. samples lost due to sample file open failure: 0
<br>> Nr. samples lost due to no permanent mapping: 120
<br>> Nr. event lost due to buffer overflow: 0
<br>> Nr. samples lost due to no mapping: 83
<br>> Nr. backtraces skipped due to no file mapping: 0
<br>> Nr. samples lost due to no mm: 330
<br>>
<br>> ---- Statistics for cpu : 1
<br>> Nr. samples lost cpu buffer overflow: 44322
<br>> Nr. samples received: 73275
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost invalid pc: 0
<br>>
<br>> ---- Statistics for cpu : 0
<br>> Nr. samples lost cpu buffer overflow: 72740
<br>> Nr. samples received: 142325
<br>> Nr. backtrace aborted: 0
<br>> Nr. samples lost invalid pc: 0
<br>>
</div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26479726&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26473260Re: Help: Increasing CPU buffer size does not solve "samples lost cpu buffer overflow"2009-11-22T22:15:37Z2009-11-22T22:15:37ZluludedeI think I found the reason:
<br>There are upper-limits on the size of the cpu buffer and event buffer.
<br>When seting the number which is more than the upper-limits, it just be deemed as invalid value by Oprofile.
<br><br>I don't know the exact <b>upper-limt</b> on <b>cpu buffer</b> and <b>event buffer</b>.
<br><b>Can anyone answer it</b>? Thanks !
<br>----------------------------------------
<br>[root@crash-29 w_seq]# opcontrol --reset; opcontrol --start; fio w_seq.ini ; opcontrol --dump; opcontrol --stop; opcontrol --save=w_seq_7
<br>Using 2.6+ OProfile kernel interface.
<br><br><b>/usr/local/bin/oprofiled: out of memory allocating 3200000000 bytes after a total of 135168 bytes</b><br>Couldn't start oprofiled.
<br><blockquote class="quote light-black dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">luludede wrote:</div>
<div class="quote-message shrinkable-quote">Dear All,
<br><br>I kept getting "Nr. samples lost cpu buffer overflow" error when I used oprofile 0.9.6-rc1 to profile the kernel. Although I increased the "CPU buffer size" to be far greater than the ((Nr. samples lost cpu buffer overflow) + (Nr. samples received)).
<br><br>Is it a bug with oprofile or I did something wrong? Can anyone give me some suggestion? Thans a lot !
<br><br>--------------------------------
<br><b>steps I took for profiling:</b><br># opcontrol --reset
<br># opcontrol --start
<br># my bench mark
<br># opcontrol --dump
<br># opcontrol --stop
<br>--------------------------------
<br>the profiler setting:
<br># opcontrol --status
<br>Daemon running: pid 11302
<br>Event 0: GLOBAL_POWER_EVENTS:100000:1:1:1
<br>Separate options: none
<br>vmlinux file: /usr/lib/debug/lib/modules/2.6.29.4-167.fc11.i686.PAEdebug/vmlinux
<br>Image filter: none
<br>Call-graph depth: 10
<br><b>Buffer size: 80000000
<br>CPU buffer watershed: 40000000
<br>CPU buffer size: 400000000</b><br>----------------------------------
<br>-- OProfile Statistics --
<br>Nr. sample dumps: 128
<br>Nr. non-backtrace samples: 3153835
<br>Nr. kernel samples: 3066290
<br>Nr. lost samples (no kernel/user): 0
<br>Nr. lost kernel samples: 0
<br>Nr. incomplete code structs: 0
<br>Nr. samples lost due to sample file open failure: 0
<br>Nr. samples lost due to no permanent mapping: 120
<br>Nr. event lost due to buffer overflow: 0
<br>Nr. samples lost due to no mapping: 83
<br>Nr. backtraces skipped due to no file mapping: 0
<br>Nr. samples lost due to no mm: 330
<br><br>---- Statistics for cpu : 1
<br><b>Nr. samples lost cpu buffer overflow: 44322
<br>Nr. samples received: 73275</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><br>---- Statistics for cpu : 0
<br><b>Nr. samples lost cpu buffer overflow: 72740
<br>Nr. samples received: 142325</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br></div>
</div></blockquote>
<p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26472870Help: Increasing CPU buffer size does not solve "samples lost cpu buffer overflow"2009-11-22T21:04:07Z2009-11-22T21:04:07ZluludedeDear All,
<br><br>I kept getting "Nr. samples lost cpu buffer overflow" error when I used oprofile 0.9.6-rc1 to profile the kernel. Although I increased the "CPU buffer size" to be far greater than the ((Nr. samples lost cpu buffer overflow) + (Nr. samples received)).
<br><br>Is it a bug with oprofile or I did something wrong? Can anyone give me some suggestion? Thans a lot !
<br><br>--------------------------------
<br><b>steps I took for profiling:</b><br># opcontrol --reset
<br># opcontrol --start
<br># my bench mark
<br># opcontrol --dump
<br># opcontrol --stop
<br>--------------------------------
<br>the profiler setting:
<br># opcontrol --status
<br>Daemon running: pid 11302
<br>Event 0: GLOBAL_POWER_EVENTS:100000:1:1:1
<br>Separate options: none
<br>vmlinux file: /usr/lib/debug/lib/modules/2.6.29.4-167.fc11.i686.PAEdebug/vmlinux
<br>Image filter: none
<br>Call-graph depth: 10
<br><b>Buffer size: 80000000
<br>CPU buffer watershed: 40000000
<br>CPU buffer size: 400000000</b><br>----------------------------------
<br>-- OProfile Statistics --
<br>Nr. sample dumps: 128
<br>Nr. non-backtrace samples: 3153835
<br>Nr. kernel samples: 3066290
<br>Nr. lost samples (no kernel/user): 0
<br>Nr. lost kernel samples: 0
<br>Nr. incomplete code structs: 0
<br>Nr. samples lost due to sample file open failure: 0
<br>Nr. samples lost due to no permanent mapping: 120
<br>Nr. event lost due to buffer overflow: 0
<br>Nr. samples lost due to no mapping: 83
<br>Nr. backtraces skipped due to no file mapping: 0
<br>Nr. samples lost due to no mm: 330
<br><br>---- Statistics for cpu : 1
<br><b>Nr. samples lost cpu buffer overflow: 44322
<br>Nr. samples received: 73275</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><br>---- Statistics for cpu : 0
<br><b>Nr. samples lost cpu buffer overflow: 72740
<br>Nr. samples received: 142325</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26472849Re: Nr. lost kernel samples2009-11-22T20:59:12Z2009-11-22T20:59:12ZluludedeHi,
<br><br>I kept getting "Nr. samples lost cpu buffer overflow" when I use oprofile 0.9.6-rc1 to profile the kernel.
<br><br>Although I increased the "CPU buffer size" to be far greater than the (Nr. samples lost cpu buffer overflow) plus (Nr. samples received).
<br><br>Is it a bug with oprofile or I did something wrong? Can anyone give me some suggestion? Thans a lot !
<br><br>--------------------------------
<br>steps I took for profiling:
<br># opcontrol --reset
<br># opcontrol --start
<br># my bench mark
<br># opcontrol --dump
<br># opcontrol --stop
<br>--------------------------------
<br>the profiler setting:
<br># opcontrol --status
<br>Daemon running: pid 11302
<br>Event 0: GLOBAL_POWER_EVENTS:100000:1:1:1
<br>Separate options: none
<br>vmlinux file: /usr/lib/debug/lib/modules/2.6.29.4-167.fc11.i686.PAEdebug/vmlinux
<br>Image filter: none
<br>Call-graph depth: 10
<br><b>Buffer size: 80000000
<br>CPU buffer watershed: 40000000
<br>CPU buffer size: 400000000</b><br>----------------------------------
<br>-- OProfile Statistics --
<br>Nr. sample dumps: 128
<br>Nr. non-backtrace samples: 3153835
<br>Nr. kernel samples: 3066290
<br>Nr. lost samples (no kernel/user): 0
<br>Nr. lost kernel samples: 0
<br>Nr. incomplete code structs: 0
<br>Nr. samples lost due to sample file open failure: 0
<br>Nr. samples lost due to no permanent mapping: 120
<br>Nr. event lost due to buffer overflow: 0
<br>Nr. samples lost due to no mapping: 83
<br>Nr. backtraces skipped due to no file mapping: 0
<br>Nr. samples lost due to no mm: 330
<br><br>---- Statistics for cpu : 1
<br><b>Nr. samples lost cpu buffer overflow: 44322
<br>Nr. samples received: 73275</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><br>---- Statistics for cpu : 0
<br><b>Nr. samples lost cpu buffer overflow: 72740
<br>Nr. samples received: 142325</b><br>Nr. backtrace aborted: 0
<br>Nr. samples lost invalid pc: 0
<br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26471738Fwd: Stable configuration for oprofile + Xen2009-11-22T17:37:36Z2009-11-22T17:37:36ZNishikant Deshmukh<br><br>
<div class="gmail_quote">---------- Forwarded message ----------<br>From: <b class="gmail_sendername">Nishikant Deshmukh</b> <span dir="ltr"><<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26471738&i=0" target="_top" rel="nofollow">nishikant.deshmukh@...</a>></span><br>
Date: Wed, Nov 18, 2009 at 7:50 PM<br>Subject: Stable configuration for oprofile + Xen<br>To: <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26471738&i=1" target="_top" rel="nofollow">oprofile-list@...</a>, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26471738&i=2" target="_top" rel="nofollow">oprofile-list@...</a><br>
<br><br>Hi,<br><br> I have 2 Xeon processors (Dell T7400) and I want to know a stable Xen distribution on which oprofile (xenoprof) can be run.<br clear="all"><font color="#888888"><br>-- <br>Best Regards,<br>Nishikant Deshmukh<br>
Graduate Student<br>Johns Hopkins University Homewood Campus<br><br>Sent from Baltimore, MD, United States<br></font></div><br><br clear="all"><br>-- <br>Best Regards,<br>Nishikant Deshmukh<br>Graduate Student<br>Johns Hopkins University Homewood Campus<br>
<br><br>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26471738&i=3" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26445416RE: Event Multiplexing2009-11-20T07:40:59Z2009-11-20T07:40:59ZSuthikulpanit, Suravee<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 11 (filtered medium)">
<!--[if !mso]>
<style>
v\:* {behavior:url(#default#VML);}
o\:* {behavior:url(#default#VML);}
w\:* {behavior:url(#default#VML);}
.shape {behavior:url(#default#VML);}
</style>
<![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>With out the patch, you would need to do multiple
profile runs without doing “opcontrol –reset” in between each
run.<o:p></o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>Suravee<o:p></o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<div>
<div class=MsoNormal align=center style='text-align:center'><font size=3 face="Times New Roman"><span style='font-size:12.0pt'>
<hr size=2 width="100%" align=center tabindex=-1>
</span></font></div>
<p class=MsoNormal><b><font size=2 face=Tahoma><span style='font-size:10.0pt;
font-family:Tahoma;font-weight:bold'>From:</span></font></b><font size=2 face=Tahoma><span style='font-size:10.0pt;font-family:Tahoma'> Thomas Marmaduke
[mailto:<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26445416&i=0" target="_top" rel="nofollow">tkm@...</a>] <br>
<b><span style='font-weight:bold'>Sent:</span></b> Thursday, November 19, 2009
11:45 PM<br>
<b><span style='font-weight:bold'>To:</span></b> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26445416&i=1" target="_top" rel="nofollow">oprofile-list@...</a><br>
<b><span style='font-weight:bold'>Subject:</span></b> Event Multiplexing</span></font><o:p></o:p></p>
</div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'><o:p> </o:p></span></font></p>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>I am currently working on a project where we are trying to profile<br>
around 20 different events, and needless to say event multiplexing would<br>
be very useful. I found a source on line which gives a patch (which<br>
doesn't have a patch file in it. . .) that will allow this.<br>
<br>
Are there any easier ways of achieving event multiplexing than replacing<br>
the oprofile associated files in my kernel source with those from the<br>
patch and recompiling? Any other suggestions for this?<br>
<br>
Thanks,<br>
<font color="#888888"><span style='color:#888888'>Tom</span></font> <o:p></o:p></span></font></p>
</div>
</body>
</html>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26445416&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26444701Re: Support for hardware counters on Cortext A92009-11-20T06:55:31Z2009-11-20T06:55:31ZJean PihetHi,
<br><br>On Friday 20 November 2009 10:10:03 vrukesh panse wrote:
<br>> Hi,
<br>>
<br>> We are use Oprofile with ARM CortexTM A9 dual core SMP based Linux kernel.
<br>The kernel has support for Cortex-A8 performance counters, cf.
<br>arch/arm/oprofile/op_model_v7.[ch]. I do not know if Cortex-A9 is backward
<br>compatible with it, if so you can use the A8 code.
<br><br>> We are using kernel 2.6.29 with Oprofile 0.9.5.
<br>>
<br>> Right now, by default, Oprofile works in timer interrupt mode.
<br>>
<br>> Kindly let us know how to enable hardware counters.
<br>To enable it you need to compile you need to enable CONFIG_OPROFILE_ARMV7 in
<br>the kernel config, and remove the 'oprofile.timer=1' argument from the kernel
<br>command line.
<br>>
<br>> Thanks and Regards,
<br>> Vrukesh
<br><br>Regards,
<br>Jean
<br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26444701&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26444129Support for hardware counters on Cortext A92009-11-20T01:10:03Z2009-11-20T01:10:03Zvrukesh panse<div>Hi,</div>
<div> </div>
<div>We are use Oprofile with ARM CortexTM A9 dual core SMP based Linux kernel.</div>
<div> </div>
<div>We are using kernel 2.6.29 with Oprofile 0.9.5.</div>
<div> </div>
<div>Right now, by default, Oprofile works in timer interrupt mode.</div>
<div> </div>
<div>Kindly let us know how to enable hardware counters.</div>
<div> </div>
<div>Thanks and Regards,</div>
<div>Vrukesh</div>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26444129&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26438877Event Multiplexing2009-11-19T21:44:37Z2009-11-19T21:44:37ZThomas MarmadukeI am currently working on a project where we are trying to profile<br>
around 20 different events, and needless to say event multiplexing would<br>
be very useful. I found a source on line which gives a patch (which<br>
doesn't have a patch file in it. . .) that will allow this.<br>
<br>
Are there any easier ways of achieving event multiplexing than replacing<br>
the oprofile associated files in my kernel source with those from the<br>
patch and recompiling? Any other suggestions for this?<br>
<br>
Thanks,<br>
<font color="#888888">Tom</font>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26438877&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26433706Re: Release Candidate 1 for OProfile 0.9.62009-11-19T12:35:21Z2009-11-19T12:35:21ZMaynard JohnsonMaynard Johnson wrote:
<div class='shrinkable-quote'><br>> Maynard Johnson wrote:
<br>>> We are pleased to announce OProfile 0.9.6 Release Candidate 1. This is primarily a bug-fix release. You can download this release at:
<br>>> <a href="http://sourceforge.net/projects/oprofile/files/" target="_top" rel="nofollow">http://sourceforge.net/projects/oprofile/files/</a><br>>>
<br>>> Please download and test this release candidate, and send your feedback by replying to this message. Please include your hardware platform and Linux distribution information in your reply.
<br>>
<br>> So far, I've run my oprofile testsuite on a POWER5 system, which includes JVMTI
<br>> testing. Also manually verified handling of separate debuginfo files.
<br>>
<br>> *OProfile Community members*, please test this release candidate in your
<br>> environments and give us your feedback.
</div>I've gotten feedback from a couple community members on their testing of
<br>oprofile 0.9.6-rc1. Thanks those who have spent a little time to test this out.
<br> So far, there have been no problems found. I will give a few more days for
<br>further comments from the community. If no issues arise, I'll probably GA the
<br>0.9.6 release sometime next week. Please try to give it a test run before then
<br>and let me know your results. Better to find any problems now than after GA.
<br><br>Thanks.
<br>-Maynard
<div class='shrinkable-quote'><br>>
<br>> Thank you.
<br>> -Maynard
<br>>
<br>>> Thanks.
<br>>> -Maynard Johnson
<br>>>
<br>>> -----------------------------------------------------------------
<br>>>
<br>>> Release Notes
<br>>> ===============
<br>>> OProfile is a powerful system-wide profiler for Linux. Read
<br>>> more at <a href="http://oprofile.sf.net" target="_top" rel="nofollow">http://oprofile.sf.net</a><br>>>
<br>>> OProfile 0.9.6 has been released. OProfile is still in alpha,
<br>>> but has been proven stable for many users.
<br>>>
<br>>>
<br>>> New features
<br>>> ------------
<br>>>
<br>>> None -- this is strictly a bug-fix release.
<br>>>
<br>>>
<br>>> Bug fixes
<br>>> ---------
<br>>>
<br>>> The main reason for this bug-fix release is to fix three regression bugs that
<br>>> occurred in 0.9.5:
<br>>>
<br>>> 1. bug #2845063: opcontrol cannot start daemon in timer mode
<br>>> 2. Improper handling of separate debuginfo files
<br>>> 3. XML callgraph output has no symbol-level sample counts
<br>>>
<br>>> Other fixes made in this release are:
<br>>>
<br>>> -Fix ophelp -X output for IBS events
<br>>>
<br>>> -Fix ophelp output for "counter:cpuid" and "ext:xxx" case
<br>>>
<br>>> -Fix opcontrol to prevent "Stopping profiling" message when profiler isn't running
<br>>>
<br>>> -Fix opcontrol to kill daemon only if it's running
<br>>>
<br>>> -Fix buffer overflows in xml generator, which can result in segfault
<br>>>
<br>>> -Add extra #includes in agents/jvmpi code since newer libstdc++ header files
<br>>> do not include standard C header files
<br>>>
<br>>> -Update events and unitmasks for AMD Family10 from publication 31116 revision
<br>>> 3.34 and fixes
<br>>>
<br>>> -Update events and groups for IBM POWER7
<br>>>
<br>>> -Change element definition for callers and callees in opreport.xsd and bumped
<br>>> schema version to "3.0"
<br>>>
<br>>> -Fix start-daemon problem on ia64
<br>>>
<br>>>
<br>>> Known problems
<br>>> --------------
<br>>>
<br>>> If NMI watchdog is not disabled (x86), opcontrol may fail to
<br>>> allocate the hardware performance counter(s) it needs (bug #1468916).
<br>>>
<br>>> configure script fails to handle --without-java correctly (bug #2153620).
<br>>>
<br>>> When profiling stripped libraries where OProfile needs to access a debuginfo
<br>>> file, it's possible that src filename/line # information may be incorrect
<br>>> (bug #2573008).
<br>>>
<br>>> Many Alpha ev67 events do not work (bug #931875).
<br>>>
<br>>> For 2.2 kernels, the module must be compiled as the same user
<br>>> that owns the kernel source tree.
<br>>>
<br>>>
<br>>> ------------------------------------------------------------------------------
<br>>> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
<br>>> is the only developer event you need to attend this year. Jumpstart your
<br>>> developing skills, take BlackBerry mobile applications to market and stay
<br>>> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
<br>>> <a href="http://p.sf.net/sfu/devconference" target="_top" rel="nofollow">http://p.sf.net/sfu/devconference</a><br>>> _______________________________________________
<br>>> oprofile-list mailing list
<br>>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26433706&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br>>> <a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br>>
<br>>
<br>> ------------------------------------------------------------------------------
<br>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>> trial. Simplify your report design, integration and deployment - and focus on
<br>> what you do best, core application coding. Discover what's new with
<br>> Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>> _______________________________________________
<br>> oprofile-list mailing list
<br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26433706&i=1" target="_top" rel="nofollow">oprofile-list@...</a>
<br>> <a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a></div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26433706&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26419500RTC mode with a 2.6 kernel?2009-11-18T18:42:07Z2009-11-18T18:42:07ZWashington Ratso<html><body bgcolor="#FFFFFF"><div></div><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I am running Linux 2.6.26 on an MPC8248 (which does not provide support to use the hardware performance counters ). HZ=250. Thus, I am getting samples every 4 ms. <o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I would like finer resolution. Can I use RTC mode with my 2.6 kernel? If so, how? When I try to force it, I get:<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --event=RTC_INTERRUPTS:4096<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">You cannot specify any performance counter events<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">because OProfile is in timer mode.</span></p></span><br></div></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26419500&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26419405oprofile on AMD Opteron Processor 2354 (ProLiant BL465c G5 )2009-11-18T18:27:36Z2009-11-18T18:27:36ZMonchiero, Matteo<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal>I’m running oprofile 0.9.5 on a Quad-Core AMD
Opteron(tm) Processor 2354, Linux 2.6.29.6. My machine is a ProLiant BL465c G5.<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>I do <o:p></o:p></p>
<p class=MsoNormal>opcontrol --start --vmlinux
/home/monchier/linux-2.6.29.6/vmlinux -e CPU_CLK_UNHALTED:500000
--separate=lib,cpu<o:p></o:p></p>
<p class=MsoNormal>yes>/dev/null<o:p></o:p></p>
<p class=MsoNormal>opcontrol --dump<o:p></o:p></p>
<p class=MsoNormal>opreport<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>but what I get is <o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>opreport error: No sample file found: try running opcontrol
--dump<o:p></o:p></p>
<p class=MsoNormal>or specify a session containing sample files<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>I put a printk in /drivers/oprofile/cpu_buffer.c in
the log_sample function, and this is never triggered. Only if I generate an NMI
from my BMC, it gets triggered, but only once. I guess, this means that if I
generate other NMIs these are not detected.<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Anybody knows of any issue with this setup?<o:p></o:p></p>
<p class=MsoNormal>Any help is welcome,<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Thanks<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>--matteo<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
</div>
</body>
</html>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26419405&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26418575Stable configuration for oprofile + Xen2009-11-18T16:50:17Z2009-11-18T16:50:17ZNishikant DeshmukhHi,<br><br> I have 2 Xeon processors (Dell T7400) and I want to know a stable Xen distribution on which oprofile (xenoprof) can be run.<br clear="all"><br>-- <br>Best Regards,<br>Nishikant Deshmukh<br>Graduate Student<br>
Johns Hopkins University Homewood Campus<br><br>Sent from Baltimore, MD, United States<br>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418575&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26418228Re: Trouble profiling kernel module2009-11-18T16:10:05Z2009-11-18T16:10:05ZWashington RatsoSent from my iPhone 3G
<br><br>On Nov 18, 2009, at 4:00 PM, Maynard Johnson <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=0" target="_top" rel="nofollow">maynardj@...</a>>
<br>wrote:
<br><div class='shrinkable-quote'><br>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=1" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>> On Nov 18, 2009, at 10:08 AM, Maynard Johnson <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=2" target="_top" rel="nofollow">maynardj@...</a>>
<br>>> wrote:
<br>>>
<br>>>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=3" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>> On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=4" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>>
<br>>>>>> By looking at EIP values, I can see that samples are being
<br>>>>>> collected
<br>>>>>> for my test module. In the function get_file in daemon/
<br>>>>>> opd_sfile.c
<br>>>>>> which is called after issuing the opcontrol --shutdown command, I
<br>>>>>> can see trans->pc values that correspond to the EIP values of
<br>>>>>> samples collected for my test module but no new sample file is
<br>>>>>> created for these samples.
<br>>>>>>
<br>>>>>> Since I can see the samples are there for my test module, why
<br>>>>>> isn't
<br>>>>>> a sample file created for the test module samples?
<br>>>>> When get_file is called for the samples for my test module, the
<br>>>>> file
<br>>>>> it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/
<br>>>>> {dep}/
<br>>>>> {kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count.
<br>>>>> Thus, no sample file is created for my test module.
<br>>>>>
<br>>>>> Why would samples for my test module result in getting the sample
<br>>>>> file
<br>>>>> for no-vmlinux?
<br>>>> Back on Nov 10, Jose Santos correctly told you that using the --no-
<br>>>> vmlinux option will not work if you want to see samples for kernel
<br>>>> modules. When you use that option, oprofile throws all kernel
<br>>>> samples (whether from base kernel or kernel modules) into the no-
<br>>>> vmlinux sample file.
<br>>>>
<br>>>> -Maynard
<br>>>
<br>>> Thank you for the clarification. Here are the commands I am now
<br>>> using
<br>>> after the system boots up:
<br>>>
<br>>> Eclipse # rm -rf /var/lib/oprofile/
<br>>> Eclipse # mkdir /var/lib/oprofile/
<br>>> Eclipse # mount -t ramfs non /var/lib/oprofile/
<br>>> Eclipse # rm /root/.oprofile/daemonrc
<br>>> Eclipse # rm /var/log/messages
<br>>> Eclipse # insmod oprofile.ko
<br>>> Eclipse # insmod helloworld.ko
<br>>> Eclipse # /usr/bin/opcontrol --setup --kernel-
<br>>> range=0xC0000000,0xC9500000 --vml
<br>>> inux=/root/vmlinux.PRO_MOD
<br>>> Eclipse # /usr/bin/opcontrol --start
<br>>> Using 2.6+ OProfile kernel interface.
<br>>> Reading module info.
<br>>> Using log file /var/lib/oprofile/samples/oprofiled.log
<br>>> Daemon started.
<br>>> Profiler running.
<br>>> Eclipse # /usr/bin/opcontrol --shutdown
<br>>> Stopping profiling.
<br>>> Killing daemon.
<br>>> Eclipse # /usr/bin/opreport --image-path=/root
<br>>> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
<br>>> Profiling through timer interrupt
<br>>> TIMER:0|
<br>>> samples| %|
<br>>> ------------------
<br>>> 9373 93.0600 vmlinux.PRO_MOD
<br>>> 587 5.8280 libc-2.6.so
<br>>> 39 0.3872 oprofiled
<br>>> 27 0.2681EclipseMSW.elf.NOV8
<br>>> 23 0.2284 busybox-1.15.2.mtab
<br>>> 19 0.1886 ld-2.6.so
<br>>> 3 0.0298 snmpd
<br>>> 1 0.0099 gateway
<br>>>
<br>>>
<br>>>
<br>>> Using printf's, I can see in get_file that for the first sample (and
<br>>> all subsequent samples) with a trans->pc value that corresponds to
<br>>> helloworld (e.g., 0xC907C0DC, helloworld start=0xC907C000
<br>>> end=0xC907C5a0), odb_open_count(file) returns 1. Since
<br>>> odb_open_count
<br>>
<br>> hmmm . . . from the addresses you state above for where your kernel
<br>> module resides, it seems you are incorrectly specifying the --kernel-
<br>> range value. That value should *only* be for the base kernel. On
<br>> most platforms, you don't need to specify this value at all --
<br>> opcontrol will automagically figure it out.
<br>>
<br>> -Maynard
</div><br>You are correct. The end of the .text section should be 0xC02633B0.
<br>I now see my sample module in the report.
<br><br>Thank you.
<br><div class='shrinkable-quote'><br>>
<br>>> (file) returns 1, no new sample file is created for helloworld.ko.
<br>>> The value of odb->data->filename in odb_open_count is /var/lib/
<br>>> oprofile/samples//current/{root}//root/vmlinux.PRO_MOD/{dep}/{root}//
<br>>> root/vmlinux.PRO_MOD/TIMER.0.0.all.all.all. Thus, the samples for
<br>>> helloworld.ko are put into the sample file for vmlinux.PRO_MOD.
<br>>>
<br>>> Should odb_open_count(file) return 1 when odb_open_count(file) is
<br>>> called for the first time for a sample from helloworld.ko? Shouldn't
<br>>> odb->data be NULL the first time a sample is processed for
<br>>> helloworld.ko so that a sample file for helloworld.ko is created?
<br>>>
<br>>>
<br>>>>>> On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=5" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>>>
<br>>>>>>> Could the fact that I am following the instructions below be an
<br>>>>>>> issue as far as sampling inserted kernel modules? Of course,
<br>>>>>>> samples for other than the kernel modules seem to be stored just
<br>>>>>>> fine.
<br>>>>>>> If running from an image in the internal jffs2 file system, you
<br>>>>>>> will need to mount another file system to store the samples.
<br>>>>>>> Below
<br>>>>>>> are examples of the commands to set that up. You will need to
<br>>>>>>> modify for local conditions.
<br>>>>>>> mkdir /var/lib/oprofile
<br>>>>>>> mount -t ramfs none /var/lib/oprofile
<br>>>>>>>
<br>>>>>>> =
<br>>>>>> =
<br>>>
<br>>
</div><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418228&i=6" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26418147Re: Trouble profiling kernel module2009-11-18T16:00:03Z2009-11-18T16:00:03ZMaynard Johnson<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<div class='shrinkable-quote'><br>> On Nov 18, 2009, at 10:08 AM, Maynard Johnson <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=1" target="_top" rel="nofollow">maynardj@...</a>>
<br>> wrote:
<br>>
<br>>> <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=2" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>> On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=3" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>
<br>>>>> By looking at EIP values, I can see that samples are being collected
<br>>>>> for my test module. In the function get_file in daemon/opd_sfile.c
<br>>>>> which is called after issuing the opcontrol --shutdown command, I
<br>>>>> can see trans->pc values that correspond to the EIP values of
<br>>>>> samples collected for my test module but no new sample file is
<br>>>>> created for these samples.
<br>>>>>
<br>>>>> Since I can see the samples are there for my test module, why isn't
<br>>>>> a sample file created for the test module samples?
<br>>>> When get_file is called for the samples for my test module, the file
<br>>>> it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/
<br>>>> {dep}/
<br>>>> {kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count.
<br>>>> Thus, no sample file is created for my test module.
<br>>>>
<br>>>> Why would samples for my test module result in getting the sample
<br>>>> file
<br>>>> for no-vmlinux?
<br>>> Back on Nov 10, Jose Santos correctly told you that using the --no-
<br>>> vmlinux option will not work if you want to see samples for kernel
<br>>> modules. When you use that option, oprofile throws all kernel
<br>>> samples (whether from base kernel or kernel modules) into the no-
<br>>> vmlinux sample file.
<br>>>
<br>>> -Maynard
<br>>
<br>> Thank you for the clarification. Here are the commands I am now using
<br>> after the system boots up:
<br>>
<br>> Eclipse # rm -rf /var/lib/oprofile/
<br>> Eclipse # mkdir /var/lib/oprofile/
<br>> Eclipse # mount -t ramfs non /var/lib/oprofile/
<br>> Eclipse # rm /root/.oprofile/daemonrc
<br>> Eclipse # rm /var/log/messages
<br>> Eclipse # insmod oprofile.ko
<br>> Eclipse # insmod helloworld.ko
<br>> Eclipse # /usr/bin/opcontrol --setup --kernel-
<br>> range=0xC0000000,0xC9500000 --vml
<br>> inux=/root/vmlinux.PRO_MOD
<br>> Eclipse # /usr/bin/opcontrol --start
<br>> Using 2.6+ OProfile kernel interface.
<br>> Reading module info.
<br>> Using log file /var/lib/oprofile/samples/oprofiled.log
<br>> Daemon started.
<br>> Profiler running.
<br>> Eclipse # /usr/bin/opcontrol --shutdown
<br>> Stopping profiling.
<br>> Killing daemon.
<br>> Eclipse # /usr/bin/opreport --image-path=/root
<br>> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
<br>> Profiling through timer interrupt
<br>> TIMER:0|
<br>> samples| %|
<br>> ------------------
<br>> 9373 93.0600 vmlinux.PRO_MOD
<br>> 587 5.8280 libc-2.6.so
<br>> 39 0.3872 oprofiled
<br>> 27 0.2681EclipseMSW.elf.NOV8
<br>> 23 0.2284 busybox-1.15.2.mtab
<br>> 19 0.1886 ld-2.6.so
<br>> 3 0.0298 snmpd
<br>> 1 0.0099 gateway
<br>>
<br>>
<br>>
<br>> Using printf's, I can see in get_file that for the first sample (and
<br>> all subsequent samples) with a trans->pc value that corresponds to
<br>> helloworld (e.g., 0xC907C0DC, helloworld start=0xC907C000
<br>> end=0xC907C5a0), odb_open_count(file) returns 1. Since odb_open_count
</div><br>hmmm . . . from the addresses you state above for where your kernel module resides, it seems you are incorrectly specifying the --kernel-range value. That value should *only* be for the base kernel. On most platforms, you don't need to specify this value at all -- opcontrol will automagically figure it out.
<br><br>-Maynard
<br><div class='shrinkable-quote'><br>> (file) returns 1, no new sample file is created for helloworld.ko.
<br>> The value of odb->data->filename in odb_open_count is /var/lib/
<br>> oprofile/samples//current/{root}//root/vmlinux.PRO_MOD/{dep}/{root}//
<br>> root/vmlinux.PRO_MOD/TIMER.0.0.all.all.all. Thus, the samples for
<br>> helloworld.ko are put into the sample file for vmlinux.PRO_MOD.
<br>>
<br>> Should odb_open_count(file) return 1 when odb_open_count(file) is
<br>> called for the first time for a sample from helloworld.ko? Shouldn't
<br>> odb->data be NULL the first time a sample is processed for
<br>> helloworld.ko so that a sample file for helloworld.ko is created?
<br>>
<br>>
<br>>>>> On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=4" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>>
<br>>>>>> Could the fact that I am following the instructions below be an
<br>>>>>> issue as far as sampling inserted kernel modules? Of course,
<br>>>>>> samples for other than the kernel modules seem to be stored just
<br>>>>>> fine.
<br>>>>>> If running from an image in the internal jffs2 file system, you
<br>>>>>> will need to mount another file system to store the samples. Below
<br>>>>>> are examples of the commands to set that up. You will need to
<br>>>>>> modify for local conditions.
<br>>>>>> mkdir /var/lib/oprofile
<br>>>>>> mount -t ramfs none /var/lib/oprofile
<br>>>>>>
<br>>>>>> =
<br>>>>> =
<br>>
</div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26418147&i=5" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26416585Re: Trouble profiling kernel module2009-11-18T13:58:33Z2009-11-18T13:58:33ZWashington Ratso<html><body bgcolor="#FFFFFF"><div><span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); ">On Nov 18, 2009, at 10:08 AM, Maynard Johnson <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26416585&i=0" target="_top" rel="nofollow">maynardj@...</a>> wrote:</span></div><div><br></div><div></div><blockquote type="cite"><div><span><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26416585&i=1" target="_top" rel="nofollow">jobhunts02@...</a> wrote:</span><br><blockquote type="cite"><span>On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26416585&i=2" target="_top" rel="nofollow">jobhunts02@...</a> wrote:</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><blockquote type="cite"><span>By looking at EIP values, I can see that samples are being collected </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>for my test module. In the function get_file in daemon/opd_sfile.c </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>which is called after issuing the opcontrol --shutdown command, I </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>can see trans->pc values that correspond to the EIP values of </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>samples collected for my test module but no new sample file is </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>created for these samples.</span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span></span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>Since I can see the samples are there for my test module, why isn't </span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>a sample file created for the test module samples?</span><br></blockquote></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>When get_file is called for the samples for my test module, the file </span><br></blockquote><blockquote type="cite"><span>it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/{dep}/ </span><br></blockquote><blockquote type="cite"><span>{kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count. </span><br></blockquote><blockquote type="cite"><span>Thus, no sample file is created for my test module.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Why would samples for my test module result in getting the sample file </span><br></blockquote><blockquote type="cite"><span>for no-vmlinux?</span><br></blockquote><span>Back on Nov 10, Jose Santos correctly told you that using the --no-vmlinux option will not work if you want to see samples for kernel modules. When you use that option, oprofile throws all kernel samples (whether from base kernel or kernel modules) into the no-vmlinux sample file.</span><br><span></span><br><span>-Maynard</span><br></div></blockquote><div><br></div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Thank you for the clarification. Here are the commands I am now using after the system boots up:<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # rm -rf /var/lib/oprofile/<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # mkdir /var/lib/oprofile/<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # mount -t ramfs non /var/lib/oprofile/<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # rm /root/.oprofile/daemonrc<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # rm /var/log/messages<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # insmod oprofile.ko<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # insmod helloworld.ko<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --setup --kernel-range=0xC0000000,0xC9500000 --vml<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">inux=/root/vmlinux.PRO_MOD<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --start<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Using 2.6+ OProfile kernel interface.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Reading module info.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Using log file /var/lib/oprofile/samples/oprofiled.log<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Daemon started.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Profiler running.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opcontrol --shutdown<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Stopping profiling.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Killing daemon.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opreport --image-path=/root<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">CPU: CPU with timer interrupt, speed 0 MHz (estimated)<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Profiling through timer interrupt<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> TIMER:0|<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> samples| %|<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:9373%2093.0600" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">9373 93.0600</a> vmlinux.PRO_MOD<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:587%C2%A0%205.8280" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">587 5.8280</a> libc-2.6.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:39%C2%A0%200.3872" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">39 0.3872</a> oprofiled<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:27%C2%A0%200.2681" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">27 0.2681</a>EclipseMSW.elf.NOV8<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:23%C2%A0%200.2284" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">23 0.2284</a> busybox-1.15.2.mtab<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:19%C2%A0%200.1886" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">19 0.1886</a> ld-2.6.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 3 0.0298 snmpd<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 1 0.0099 gateway<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Using printf's, I can see in get_file that for the first sample (and all subsequent samples) with a trans->pc value that corresponds to helloworld (e.g., 0xC907C0DC, helloworld start=0xC907C000 end=0xC907C5a0), odb_open_count(file) returns 1. Since odb_open_count(file) returns 1, no new sample file is created for helloworld.ko. The value of odb->data->filename in odb_open_count is /var/lib/oprofile/samples//current/{root}//root/vmlinux.PRO_MOD/{dep}/{root}//root/vmlinux.PRO_MOD/TIMER.0.0.all.all.all. Thus, the samples for helloworld.ko are put into the sample file for vmlinux.PRO_MOD.<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Should odb_open_count(file) return 1 when odb_open_count(file) is called for the first time for a sample from helloworld.ko? Shouldn't odb->data be NULL the first time a sample is processed for helloworld.ko so that a sample file for helloworld.ko is created?</p></span><div><br></div><br><blockquote type="cite"><div><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><blockquote type="cite"><span></span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26416585&i=3" target="_top" rel="nofollow">jobhunts02@...</a> wrote:</span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span></span><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>Could the fact that I am following the instructions below be an </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>issue as far as sampling inserted kernel modules? Of course, </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>samples for other than the kernel modules seem to be stored just </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>fine.</span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>If running from an image in the internal jffs2 file system, you </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>will need to mount another file system to store the samples. Below </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>are examples of the commands to set that up. You will need to </span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>modify for local conditions.</span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>mkdir /var/lib/oprofile</span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>mount -t ramfs none /var/lib/oprofile</span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span></span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><span>=</span><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><span>=</span><br></blockquote></blockquote><blockquote type="cite"><span></span><br></blockquote><span></span><br></div></blockquote></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26416585&i=4" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26412829Re: Trouble profiling kernel module2009-11-18T10:08:37Z2009-11-18T10:08:37ZMaynard Johnson<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412829&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<div class='shrinkable-quote'><br>> On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412829&i=1" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>
<br>>> By looking at EIP values, I can see that samples are being collected
<br>>> for my test module. In the function get_file in daemon/opd_sfile.c
<br>>> which is called after issuing the opcontrol --shutdown command, I
<br>>> can see trans->pc values that correspond to the EIP values of
<br>>> samples collected for my test module but no new sample file is
<br>>> created for these samples.
<br>>>
<br>>> Since I can see the samples are there for my test module, why isn't
<br>>> a sample file created for the test module samples?
<br>>
<br>> When get_file is called for the samples for my test module, the file
<br>> it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/{dep}/
<br>> {kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count.
<br>> Thus, no sample file is created for my test module.
<br>>
<br>> Why would samples for my test module result in getting the sample file
<br>> for no-vmlinux?
</div>Back on Nov 10, Jose Santos correctly told you that using the --no-vmlinux option will not work if you want to see samples for kernel modules. When you use that option, oprofile throws all kernel samples (whether from base kernel or kernel modules) into the no-vmlinux sample file.
<br><br>-Maynard
<div class='shrinkable-quote'><br>>
<br>>>
<br>>> On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412829&i=2" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>
<br>>>> Could the fact that I am following the instructions below be an
<br>>>> issue as far as sampling inserted kernel modules? Of course,
<br>>>> samples for other than the kernel modules seem to be stored just
<br>>>> fine.
<br>>>> If running from an image in the internal jffs2 file system, you
<br>>>> will need to mount another file system to store the samples. Below
<br>>>> are examples of the commands to set that up. You will need to
<br>>>> modify for local conditions.
<br>>>> mkdir /var/lib/oprofile
<br>>>> mount -t ramfs none /var/lib/oprofile
<br>>>>
<br>>>> =
<br>>> =
<br>>
</div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412829&i=3" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26412645Re: Trouble profiling kernel module2009-11-18T09:55:38Z2009-11-18T09:55:38ZMaynard Johnson<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412645&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<div class='shrinkable-quote'><br>> On Nov 17, 2009, at 4:43 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412645&i=1" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>
<br>>> On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412645&i=2" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>
<br>>>> By looking at EIP values, I can see that samples are being
<br>>>> collected for my test module. In the function get_file in daemon/
<br>>>> opd_sfile.c which is called after issuing the opcontrol --shutdown
<br>>>> command, I can see trans->pc values that correspond to the EIP
<br>>>> values of samples collected for my test module but no new sample
<br>>>> file is created for these samples.
<br>>>>
<br>>>> Since I can see the samples are there for my test module, why isn't
<br>>>> a sample file created for the test module samples?
<br>>> When get_file is called for the samples for my test module, the file
<br>>> it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/
<br>>> {dep}/{kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero
<br>>> count. Thus, no sample file is created for my test module.
<br>>>
<br>>> Why would samples for my test module result in getting the sample
<br>>> file for no-vmlinux?
<br>>>
<br>>
<br>> I put vmlinux.PRO_MOD on my board. Looking at opreport output for the
<br>> same sampling, the results look inconsistent.
<br>>
<br>> Running /usr/bin/opreport gave:
<br>>
<br>> Eclipse # /usr/bin/opreport
<br>> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
<br>> Profiling through timer interrupt
<br>> TIMER:0|
<br>> samples| %|
<br>> ------------------
<br>> 7703 93.3923 vmlinux.PRO_MOD
<br>> 464 5.6256 libc-2.6.so
<br>> 26 0.3152 oprofiled
<br>> 19 0.2304EclipseMSW.elf.NOV8
<br>> 17 0.2061 busybox-1.15.2.mtab
<br>> 15 0.1819 ld-2.6.so
<br>> 3 0.0364 snmpd
<br>> 1 0.0121 libEclipsePai.so
<br>>
<br>>
<br>> If 7703 samples is 93.3923 %, then total samples = 8248.
<br>>
<br>>
<br>> My test module is in the /root directory. Running /usr/bin/opreport -
<br>> c --image-path=/root gave:
<br>>
<br>> /usr/bin/opreport -c --image-path=/root
<br>> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
<br>> Profiling through timer interrupt
<br>> samples % app name symbol name
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 1343 38.0993 vmlinux.PRO_MOD __delay
<br>> 1343 100.000 vmlinux.PRO_MOD __delay [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 464 13.1631 libc-2.6.so /lib/libc-2.6.so
<br>> 464 100.000 libc-2.6.so /lib/libc-2.6.so [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 267 7.5745 vmlinux.PRO_MOD memset
<br>> 267 100.000 vmlinux.PRO_MOD memset [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 100 2.8369 vmlinux.PRO_MOD memcpy
<br>> 100 100.000 vmlinux.PRO_MOD memcpy [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 83 2.3546 vmlinux.PRO_MOD cfi_amdstd_write_buffers
<br>> 83 100.000 vmlinux.PRO_MOD cfi_amdstd_write_buffers
<br>> [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>> 61 1.7305 vmlinux.PRO_MOD deflate_fast
<br>> 61 100.000 vmlinux.PRO_MOD deflate_fast [self]
<br>> ---
<br>> ---
<br>> ---
<br>> ----------------------------------------------------------------------
<br>>
<br>> I DELETED THE REST OF THE OUTPUT.
<br>>
<br>>
<br>> If 1343 samples is 38.0993 %, then total samples = 3525.
<br>>
<br>>
<br>> Are these results inconsistent or am I mistaken?
</div>You're mistaken. Please read section 4.2.3 in the oprofile manual on "Callgraph output". As it states there, "The samples and percentage values here (in entries above the non-indented line) refer to the number of times we took a sample where this call was found in the stack; the percentage is relative to all other callers of the function we're focussing on."
<br><br>-Maynard
<div class='shrinkable-quote'><br>>
<br>>
<br>>
<br>>
<br>>
<br>>
<br>>
<br>>>>
<br>>>> On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412645&i=3" target="_top" rel="nofollow">jobhunts02@...</a> wrote:
<br>>>>
<br>>>>> Could the fact that I am following the instructions below be an
<br>>>>> issue as far as sampling inserted kernel modules? Of course,
<br>>>>> samples for other than the kernel modules seem to be stored just
<br>>>>> fine.
<br>>>>> If running from an image in the internal jffs2 file system, you
<br>>>>> will need to mount another file system to store the samples. Below
<br>>>>> are examples of the commands to set that up. You will need to
<br>>>>> modify for local conditions.
<br>>>>> mkdir /var/lib/oprofile
<br>>>>> mount -t ramfs none /var/lib/oprofile
<br>>>>>
<br>>>>> =
<br>>>> =
<br>>> =
<br>>
</div><br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26412645&i=4" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26409140Oprofile on ARM: cpu_type 'unset' is not valid2009-11-18T06:37:23Z2009-11-18T06:37:23ZBugzilla from robert1peter@web.deHello!
<br><br>I am trying to use Oprofile on an ARM-RealView PB1176 board (ARM11
<br>processor). When I try to initialize the daemon, I get:
<br><br>opcontrol --init
<br>cpu_type 'unset' is not valid
<br>you should upgrade oprofile or force the use of timer mode
<br><br>I use current userspace tools (0.9.5). It seems that the CPU is not
<br>correctly recognized by the kernel:
<br><br>(dmesg)
<br>oprofile: cpu_architecture() returns 0x9, using arm/11mpcore model
<br><br>Kernel is 2.6.28. In the kernel-tree (arch/arm/include/asm/system.h),
<br>there is no CPU-Type defind for 0x9, so maybe it defaults to arm/11mpcore.
<br><br><br>Any help would be appreciated!
<br><br>Robert
<br><br>------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26409140&i=0" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26401763Re: Trouble profiling kernel module2009-11-17T19:04:08Z2009-11-17T19:04:08ZWashington Ratso<html><body bgcolor="#FFFFFF"><div><span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); ">On Nov 17, 2009, at 4:43 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26401763&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:</span></div><div><br></div><div></div><blockquote type="cite"><div><div><span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); ">On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26401763&i=1" target="_top" rel="nofollow">jobhunts02@...</a></a> wrote:</span><br></div><div><br></div><div></div><blockquote type="cite"><div><div>By looking at EIP values, I can see that samples are being collected for my test module. In the function get_file in daemon/opd_sfile.c which is called after issuing the opcontrol --shutdown command, I can see trans->pc values that correspond to the EIP values of samples collected for my test module but no new sample file is created for these samples.</div><div><br></div><div>Since I can see the samples are there for my test module, why isn't a sample file created for the test module samples?</div></div></blockquote><br><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.289062); -webkit-composition-fill-color: rgba(175, 192, 227, 0.222656); -webkit-composition-frame-color: rgba(77, 128, 180, 0.222656); font-family: Calibri; ">When get_file is called for the samples for my test module, the file it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/{dep}/{kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count. Thus, no sample file is created for my test module.</span><br></div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Why would samples for my test module result in getting the sample file for no-vmlinux?</p></span><div><br></div></div></blockquote><div><br></div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">I put vmlinux.PRO_MOD on my board. Looking at opreport output for the same sampling, the results look inconsistent.<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Running /usr/bin/opreport gave:<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Eclipse # /usr/bin/opreport<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">CPU: CPU with timer interrupt, speed 0 MHz (estimated)<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Profiling through timer interrupt<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> TIMER:0|<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> samples| %|<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:7703%2093.3923" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">7703 93.3923</a> vmlinux.PRO_MOD<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:464%C2%A0%205.6256" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">464 5.6256</a> libc-2.6.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:26%C2%A0%200.3152" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">26 0.3152</a> oprofiled<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:19%C2%A0%200.2304" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">19 0.2304</a>EclipseMSW.elf.NOV8<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:17%C2%A0%200.2061" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">17 0.2061</a> busybox-1.15.2.mtab<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> <a href="tel:15%C2%A0%200.1819" x-apple-data-detectors="true" style="color: blue; text-decoration: underline; " target="_top" rel="nofollow">15 0.1819</a> ld-2.6.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 3 0.0364 snmpd<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 1 0.0121 libEclipsePai.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">If 7703 samples is 93.3923 %, then total samples = 8248.<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">My test module is in the /root directory. Running /usr/bin/opreport -c --image-path=/root gave:<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">/usr/bin/opreport -c --image-path=/root<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">CPU: CPU with timer interrupt, speed 0 MHz (estimated)<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">Profiling through timer interrupt<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">samples % app name symbol name<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">1343 38.0993 vmlinux.PRO_MOD __delay<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 1343 100.000 vmlinux.PRO_MOD __delay [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">464 13.1631 libc-2.6.so /lib/libc-2.6.so<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 464 100.000 libc-2.6.so /lib/libc-2.6.so [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">267 7.5745 vmlinux.PRO_MOD memset<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 267 100.000 vmlinux.PRO_MOD memset [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">100 2.8369 vmlinux.PRO_MOD memcpy<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 100 100.000 vmlinux.PRO_MOD memcpy [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">83 2.3546 vmlinux.PRO_MOD cfi_amdstd_write_buffers<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 83 100.000 vmlinux.PRO_MOD cfi_amdstd_write_buffers [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">61 1.7305 vmlinux.PRO_MOD deflate_fast<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "> 61 100.000 vmlinux.PRO_MOD deflate_fast [self]<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">-------------------------------------------------------------------------------<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; ">I DELETED THE REST OF THE OUTPUT.<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">If 1343 samples is 38.0993 %, then total samples = 3525.<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Are these results inconsistent or am I mistaken?<o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-family: 'Courier New'; "><o:p> </o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p></span><div><br></div><br><blockquote type="cite"><div><blockquote type="cite"><div><div><br></div><div><br></div><div>On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26401763&i=2" target="_top" rel="nofollow">jobhunts02@...</a><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26401763&i=3" target="_top" rel="nofollow">jobhunts02@...</a></a> wrote:<br><br></div><div></div><blockquote type="cite"><div><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Could the fact that I am following the instructions below be an issue as far as sampling inserted kernel modules? Of course, samples for other than the kernel modules seem to be stored just fine.<span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.289062); -webkit-composition-fill-color: rgba(175, 192, 227, 0.222656); -webkit-composition-frame-color: rgba(77, 128, 180, 0.222656); "> </span></p><ul type="disc" style="margin-bottom: 0in; "><li class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 12pt; font-family: 'Times New Roman', serif; ">If running from an image in the internal jffs2 file system, you will need to mount another file system to store the samples. Below are examples of the commands to set that up. You will need to modify for local conditions.</span></li></ul><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mkdir /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mount -t ramfs none /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p></span></div>=
</div></blockquote>=
</div></blockquote>=
</div></blockquote></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26401763&i=4" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26400572Re: Trouble profiling kernel module2009-11-17T16:43:39Z2009-11-17T16:43:39ZWashington Ratso<html><body bgcolor="#FFFFFF"><div><span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); ">On Nov 17, 2009, at 9:23 AM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26400572&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:</span><br></div><div><br></div><div></div><blockquote type="cite"><div><div>By looking at EIP values, I can see that samples are being collected for my test module. In the function get_file in daemon/opd_sfile.c which is called after issuing the opcontrol --shutdown command, I can see trans->pc values that correspond to the EIP values of samples collected for my test module but no new sample file is created for these samples.</div><div><br></div><div>Since I can see the samples are there for my test module, why isn't a sample file created for the test module samples?</div></div></blockquote><br><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.289062); -webkit-composition-fill-color: rgba(175, 192, 227, 0.222656); -webkit-composition-frame-color: rgba(77, 128, 180, 0.222656); font-family: Calibri; ">When get_file is called for the samples for my test module, the file it gets is /var/lib/oprofile/samples//current/{kern}/no-vmlinux/{dep}/{kern}/no-vmlinux/TIMER.0.0.all.all.all, which has a nonzero count. Thus, no sample file is created for my test module.</span><br></div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p></o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Why would samples for my test module result in getting the sample file for no-vmlinux?</p></span><div><br></div><blockquote type="cite"><div><div><br></div><div><br></div><div>On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26400572&i=1" target="_top" rel="nofollow">jobhunts02@...</a></a> wrote:<br><br></div><div></div><blockquote type="cite"><div><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Could the fact that I am following the instructions below be an issue as far as sampling inserted kernel modules? Of course, samples for other than the kernel modules seem to be stored just fine.<span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.289062); -webkit-composition-fill-color: rgba(175, 192, 227, 0.222656); -webkit-composition-frame-color: rgba(77, 128, 180, 0.222656); "> </span></p><ul type="disc" style="margin-bottom: 0in; "><li class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 12pt; font-family: 'Times New Roman', serif; ">If running from an image in the internal jffs2 file system, you will need to mount another file system to store the samples. Below are examples of the commands to set that up. You will need to modify for local conditions.</span></li></ul><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mkdir /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mount -t ramfs none /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p></span></div>=
</div></blockquote>=
</div></blockquote></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26400572&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26394010Re: Trouble profiling kernel module2009-11-17T09:23:05Z2009-11-17T09:23:05ZWashington Ratso<html><body bgcolor="#FFFFFF"><div>By looking at EIP values, I can see that samples are being collected for my test module. In the function get_file in daemon/opd_sfile.c which is called after issuing the opcontrol --shutdown command, I can see trans->pc values that correspond to the EIP values of samples collected for my test module but no new sample file is created for these samples.</div><div><br></div><div>Since I can see the samples are there for my test module, why isn't a sample file created for the test module samples?</div><div><br></div><div><br></div><div>On Nov 12, 2009, at 3:42 PM, <a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26394010&i=0" target="_top" rel="nofollow">jobhunts02@...</a> wrote:<br><br></div><div></div><blockquote type="cite"><div><div><span class="Apple-style-span" style="font-size: 15px; -webkit-tap-highlight-color: rgba(26, 26, 26, 0.296875); -webkit-composition-fill-color: rgba(175, 192, 227, 0.230469); -webkit-composition-frame-color: rgba(77, 128, 180, 0.230469); "><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; ">Could the fact that I am following the instructions below be an issue as far as sampling inserted kernel modules? Of course, samples for other than the kernel modules seem to be stored just fine.<span class="Apple-style-span" style="-webkit-tap-highlight-color: rgba(26, 26, 26, 0.289062); -webkit-composition-fill-color: rgba(175, 192, 227, 0.222656); -webkit-composition-frame-color: rgba(77, 128, 180, 0.222656); "> </span></p><ul type="disc" style="margin-bottom: 0in; "><li class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 12pt; font-family: 'Times New Roman', serif; ">If running from an image in the internal jffs2 file system, you will need to mount another file system to store the samples. Below are examples of the commands to set that up. You will need to modify for local conditions.</span></li></ul><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mkdir /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><span style="font-size: 10pt; font-family: 'Courier New'; ">mount -t ramfs none /var/lib/oprofile<o:p></o:p></span></p><p class="MsoNormal" style="margin-top: 0in; margin-right: 0in; margin-bottom: 0.0001pt; margin-left: 0in; font-size: 11pt; font-family: Calibri, sans-serif; "><o:p> </o:p></p></span></div>=
</div></blockquote></body></html>=
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26394010&i=1" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>tag:old.nabble.com,2006:post-26369540Re: Profiling network software2009-11-16T02:13:54Z2009-11-16T02:13:54ZSam Liao<h3 class="title"></h3><h3 style="font-weight: normal;" class="title">From the doc:</h3><a href="http://oprofile.sourceforge.net/doc/kernel-profiling.html" target="_top" rel="nofollow">http://oprofile.sourceforge.net/doc/kernel-profiling.html</a><br><br>
<h3 class="title">2.2. Idle time</h3>
<p>
Your kernel is likely to support halting the processor when a CPU is idle. As
the typical hardware events like <code class="constant">CPU_CLK_UNHALTED</code> do not
count when the CPU is halted, the kernel profile will not reflect the actual
amount of time spent idle. You can change this behaviour by booting with
the <code class="option">idle=poll</code> option, which uses a different idle routine. This
will appear as <code class="function">poll_idle()</code> in your kernel profile. <br></p><p><br></p>-Sam<br><br><div class="gmail_quote">2009/11/13 "Mgr. Šimon Tóth" <span dir="ltr"><<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26369540&i=0" target="_top" rel="nofollow">simon@...</a>></span><br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">I'm trying to profile a batch system (Torque). It consists of a server<br>
daemon, scheduler daemon and many node daemons.<br>
<br>
Now the problem is, that when I get the result from oprofile using<br>
CLK_UNHALTED the number of samples does not correspond to the time the<br>
specific daemon was running (0.5 sec, vs. 10 minutes). Also the profile<br>
looks weird, the server has only light load (5-10% cpu use) but oprofile<br>
shows that a majority of the samples were recorded in cpu heavy functions.<br>
<br>
I did try valgrind, but the slowdown makes the results unusable<br>
(although the profile did look much better, actually showing that there<br>
was a lot of waiting for system calls).<br>
<br>
What should I do to get better results?<br>
<br>
--<br>
Mgr. Simon Toth<br>
<br>
------------------------------------------------------------------------------<br>
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day<br>
trial. Simplify your report design, integration and deployment - and focus on<br>
what you do best, core application coding. Discover what's new with<br>
Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_blank" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br>
_______________________________________________<br>
oprofile-list mailing list<br>
<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26369540&i=1" target="_top" rel="nofollow">oprofile-list@...</a><br>
<a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_blank" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br>
</blockquote></div><br>
<br />------------------------------------------------------------------------------
<br>Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
<br>trial. Simplify your report design, integration and deployment - and focus on
<br>what you do best, core application coding. Discover what's new with
<br>Crystal Reports now. <a href="http://p.sf.net/sfu/bobj-july" target="_top" rel="nofollow">http://p.sf.net/sfu/bobj-july</a><br />_______________________________________________
<br>oprofile-list mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26369540&i=2" target="_top" rel="nofollow">oprofile-list@...</a>
<br><a href="https://lists.sourceforge.net/lists/listinfo/oprofile-list" target="_top" rel="nofollow">https://lists.sourceforge.net/lists/listinfo/oprofile-list</a><br><p>From forum: <a href="http://old.nabble.com/oprofile-list-f3785.html" embed="fixTarget[3785]" target="_top" >oprofile-list</a></p>