[ oprofile-Bugs-617264 ] oprofpp: start ... is more than end ...

[<no...@so...>]

Bugs item #617264, was opened at 2002-10-01 21:24
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=116191&aid=617264&group_id=16191

Category: None
Group: None
>Status: Closed
Resolution: Fixed
Priority: 5
Submitted By: Greg Ames (gregames)
Assigned to: Nobody/Anonymous (nobody)
Summary: oprofpp: start ... is more than end ...

Initial Comment:
 I see the following running oprofpp on kernel samples      
on RedHat AS 2.1:               
               
 [gregames@edge17 grep2]$ oprofpp -Vl              
samples/\}boot\}vmlinux-2.4.9-e.3smp#0              
.text filepos 1000              
number of symbols before excluding 5424              
number of symbols now 5424              
symbol signature, value 0              
in section VSYSCALL, filepos 12c000              
start 12c000, end 12c000              
[...several symbols deleted here...]                
symbol vsyscall_data_end, value 9000              
in section VSYSCALL, filepos 12c000              
start 135000, end 1000              
start 135000 is more than end 1000              
     
nm says that vsyscall_data_end is a "?" symbol.  I put 
some debug code into op_bfd.cpp::symbol_size.  It 
complains about start > end when next->name is 
startup_32, the first "t" symbol.  next->filepos appears to 
go down to 1000 at that point.    
              
I'm running with the RH AS 2.1 hack mentioned in bug           
#605842, if that matters.  oprofpp works fine with 
non-kernel executables/libraries on this system.        
           
Thanks much,           
Greg             

----------------------------------------------------------------------

Comment By: Greg Ames (gregames)
Date: 2002-10-02 14:15

Message:
Logged In: YES 
user_id=607249

The cvs version solves the start > end problem. 
 
Thanks much, Phil and John. 
 
Greg 

----------------------------------------------------------------------

Comment By: Philippe Elie (phil_e)
Date: 2002-10-01 22:17

Message:
Logged In: YES 
user_id=318973

we fix this bug in cvs a few adys ago, can you retry
with the current cvs and report if ti is working for
you. We forget to open a bug entry for this once...

Note than you fix for #605842 is not incorporated in
cvs (basically we can't to work-around against
all existing patched kernel by various distributions)
so you must reapply after cvs update.

John, I open a bug entry for this, I let it as unfixed
until Greg report sucess/failure

regards,
Phil

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=116191&aid=617264&group_id=16191