This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug translator/12121] Apparent non-determinism in DWARF symbol resolution


http://sourceware.org/bugzilla/show_bug.cgi?id=12121

Josh Stone <jistone at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|                            |jistone at redhat dot com

--- Comment #8 from Josh Stone <jistone at redhat dot com> 2010-10-14 18:12:59 UTC ---
I managed to reproduce the non-determinism with a much smaller test:

$ cat xulcast.stp 
probe begin {
  println(@cast(0, "JSContext", @1)->regs)
  println(@cast(0, "JSObject", @1)->fslots)
}

$ stap -vvv -p2 xulcast.stp $PWD/libxul.so

The success of this is extremely variable.  The only consistency is that the
two @casts always succeed or fail together, never one without the other.

The difference I found with the extra verbosity is in these lines:

literal_stmt_for_pointer: finding value for JSContext(SRCFILE)
literal_stmt_for_pointer: finding value for JSObject(SRCFILE)

The SRCFILE varies from run to run, with only some apparently having good type
information.  I still don't know why our choice of SRCFILE would vary, as I
would expect the iteration order to be consistent, but this seems to be the
key.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]