dtrace discuss - Aug 2006 - plockstat with Segmentation Fault

Hi:

I am on T2000 with S10 U3 bits of KernelID = Generic_118833-18 

# plockstat -x aggsize=500m -x dynvarsize=200m -x bufresize=auto -p 654
^C
Mutex block

Count     nsec Lock                         Caller
-------------------------------------------------------------------------------
Segmentation Fault(coredump)



Thanks and pointers are appreciated
 
 
This message posted from opensolaris.org

The core file is attached.

-----------------  lwp# 1 / thread# 1  --------------------
 ffffffff7ee39a50 strlen (100003faa, ffffffff7ffff6e8, ffffffff7eea0ed4,
ffffffff7fffed99, 0, 100003fa9) + 50
 ffffffff7eea4fdc snprintf (ffffffff7ffffa10, 0, 100003fa8, 0, ffffffff7ffff740,
100003000) + 88
 0000000100001cd8 ???????? (1003bff20, ff2b9bec, ffffffff7ffffa10, 28, 1, 1d)
 0000000100001ff4 ???????? (10040e4d8, 10040e4f8, 2, 10040e4d8, 10040e4c8,
100106990)
 ffffffff7f225664 dt_aggregate_walk_sorted (10010b500, 100001de8, 0,
ffffffff7f224f08, 0, 1004395c0) + a4
 0000000100002e9c main (100001, 100106000, 100000, 100001000, 100107228, 100106)
+ ad0
 000000010000159c _start (0, 0, 0, 0, 0, 0) + 17c
-----------------  lwp# 2 / thread# 2  --------------------
 ffffffff7eece820 _write (102, ffffffff725fbf48, 8, 0, ffffffff7e300000, 0) + c
 ffffffff7f252768 dt_proc_control (0, 1fc000, 8, ffffffff725fbf48, 1, 1) + 1f0
 ffffffff7eecd2d8 _lwp_start (0, 0, 0, 0, 0, 0)
 
 
This message posted from opensolaris.org

Sounds like you''re hitting this bug:

  6296612 plockstat dumps core on second call

This has probably been fixed by the rewrite of plockstat in the fix for
these bugs:

  6217718 plockstat(1m) should sort by total
  6384324 plockstat output can be very confusing

That''s being confirmed by a Solaris engineer.

Adam

On Mon, Aug 07, 2006 at 02:04:48PM -0700, ttoulliu2002
wrote:
> The core file is attached.
> 
> -----------------  lwp# 1 / thread# 1  --------------------
>  ffffffff7ee39a50 strlen (100003faa, ffffffff7ffff6e8, ffffffff7eea0ed4,
ffffffff7fffed99, 0, 100003fa9) + 50
>  ffffffff7eea4fdc snprintf (ffffffff7ffffa10, 0, 100003fa8, 0,
ffffffff7ffff740, 100003000) + 88
>  0000000100001cd8 ???????? (1003bff20, ff2b9bec, ffffffff7ffffa10, 28, 1,
1d)
>  0000000100001ff4 ???????? (10040e4d8, 10040e4f8, 2, 10040e4d8, 10040e4c8,
100106990)
>  ffffffff7f225664 dt_aggregate_walk_sorted (10010b500, 100001de8, 0,
ffffffff7f224f08, 0, 1004395c0) + a4
>  0000000100002e9c main (100001, 100106000, 100000, 100001000, 100107228,
100106) + ad0
>  000000010000159c _start (0, 0, 0, 0, 0, 0) + 17c
> -----------------  lwp# 2 / thread# 2  --------------------
>  ffffffff7eece820 _write (102, ffffffff725fbf48, 8, 0, ffffffff7e300000, 0)
+ c
>  ffffffff7f252768 dt_proc_control (0, 1fc000, 8, ffffffff725fbf48, 1, 1) +
1f0
>  ffffffff7eecd2d8 _lwp_start (0, 0, 0, 0, 0, 0)
>  
>  
> This message posted from opensolaris.org
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
-- 
Adam Leventhal, Solaris Kernel Development       http://blogs.sun.com/ahl