Re: oprofile callgraph problem

[Maynard J. <may...@us...>]

Joshua Verburg-Sachs wrote:
> *I wanted to query the mailing list before I filed a bug, because I wanted
> to make sure I wasn't misunderstanding something. I'm running oprofile to
> profile userspace code, and I start it with:*
> 
> opcontrol --no-vmlinux --callgraph=10 --start
> 
> [root@oprofile:Active] shared # opcontrol --status
> Daemon running: pid 21737
> Separate options: none
> vmlinux file: none
> Image filter: none
> Call-graph depth: 10
> 
> *But the backtrace_depth is zero*
> [root@oprofile:Active] shared # cat /dev/oprofile/backtrace_depth
> 0
hmmm . . . that is odd.  This sequence of commands works fine for me to set up callgraph profiling (using current git pull on ppc64 arch).  What architecture/processor type are you on?  After starting oprofile, what do you see in 'dmesg' output?  What version of oprofile are you using, and is it provided as an rpm by your distro or did you build it from scratch?  Please add the "--verbose" option to the above command where you're starting oprofile, and then paste in all the output from that command in your reply.

-Maynard

> 
> *and*
> opreport --callgraph --symbols /shared/xyz
> 
> *gives me results that look like this:*
> 
> 1        5.1e-06  mpi_mem_recv_pkt
>   1        100.000  mpi_mem_recv_pkt [self]
> -------------------------------------------------------------------------------
> 1        5.1e-06  mpi_mem_send_iov
>   1        100.000  mpi_mem_send_iov [self]
> -------------------------------------------------------------------------------
> 1        5.1e-06  neighbor_update_entry
>   1        100.000  neighbor_update_entry [self]
> -------------------------------------------------------------------------------
> 1        5.1e-06  nexthop_flow_accept
>   1        100.000  nexthop_flow_accept [self]
> -------------------------------------------------------------------------------
> 1        5.1e-06  ppsratecheck
>   1        100.000  ppsratecheck [self]
> 
> *How do I get the actual call stack?
> *--
> Comparing yourself to other people is like comparing strings in c++. When
> you use ==, you're really just comparing their pointers. Since no two people
> are the same person, they're never equal, and what you thought was a valid
> comparison was actually bad code.
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Create and publish websites with WebMatrix
> Use the most popular FREE web apps or write code yourself; 
> WebMatrix provides all the features you need to develop and 
> publish your website. http://p.sf.net/sfu/ms-webmatrix-sf
> 
> 
> 
> _______________________________________________
> oprofile-list mailing list
> opr...@li...
> https://lists.sourceforge.net/lists/listinfo/oprofile-list