[Varnish] #566: varnishncsa segfault during (very) high load

[Varnish]

#566: varnishncsa segfault during (very) high load
-------------------------+--------------------------------------------------
 Reporter:  kristian     |       Owner:       
     Type:  defect       |      Status:  new  
 Priority:  normal       |   Milestone:       
Component:  varnishncsa  |     Version:  trunk
 Severity:  normal       |    Keywords:       
-------------------------+--------------------------------------------------

Old description:

> {{{
> [New process 14993]
> #0  0x00002b0fc9fdb397 in VSL_NextLog (vd=0x17729010, pp=0x7fffe0ef8990)
> at shmlog.c:315
> 315                     if (vd->c_opt && !(vd->map[u] & M_CLIENT))
> (gdb) bt
> #0  0x00002b0fc9fdb397 in VSL_NextLog (vd=0x17729010, pp=0x7fffe0ef8990)
> at shmlog.c:315
> #1  0x00002b0fc9fdb47b in VSL_Dispatch (vd=0x17729010, func=0x4016f0
> <h_ncsa>, priv=0x33dff50780) at shmlog.c:349
> #2  0x00000000004012f3 in main (argc=2, argv=<value optimized out>) at
> varnishncsa.c:589
> (gdb) print vd
> $1 = (struct VSL_data *) 0x17729010
> (gdb) print vd->c_opt
> $2 = 1
> (gdb) print vd->map
> $3 = '\0' <repeats 12 times>, "\001\000", '\001' <repeats 14 times>, '\0'
> <repeats 65507 times>
> (gdb) print u
> $4 = 393216
> (gdb) print vd->map[u]
> Cannot access memory at address 0x17789154
> (gdb)
> }}}
>
> This typically happens in a matter of a few seconds when I launch my
> tests. This is during complete CPU utilization. Core file etc on request.

New description:

 {{{
 [New process 14993]
 #0  0x00002b0fc9fdb397 in VSL_NextLog (vd=0x17729010, pp=0x7fffe0ef8990)
 at shmlog.c:315
 315                     if (vd->c_opt && !(vd->map[u] & M_CLIENT))
 (gdb) bt
 #0  0x00002b0fc9fdb397 in VSL_NextLog (vd=0x17729010, pp=0x7fffe0ef8990)
 at shmlog.c:315
 #1  0x00002b0fc9fdb47b in VSL_Dispatch (vd=0x17729010, func=0x4016f0
 <h_ncsa>, priv=0x33dff50780) at shmlog.c:349
 #2  0x00000000004012f3 in main (argc=2, argv=<value optimized out>) at
 varnishncsa.c:589
 (gdb) print vd
 $1 = (struct VSL_data *) 0x17729010
 (gdb) print vd->c_opt
 $2 = 1
 (gdb) print vd->map
 $3 = '\0' <repeats 12 times>, "\001\000", '\001' <repeats 14 times>, '\0'
 <repeats 65507 times>
 (gdb) print u
 $4 = 393216
 (gdb) print vd->map[u]
 Cannot access memory at address 0x17789154
 (gdb)
 }}}

 This typically happens in a matter of a few seconds when I launch my
 tests. This is during complete CPU utilization. Core file etc on request.

--

Comment(by tfheen):

 Can you see if this is still reproducible?  I believe it has been fixed.

-- 
Ticket URL: <http://www.varnish-cache.org/trac/ticket/566#comment:1>
Varnish <http://varnish-cache.org/>
The Varnish HTTP Accelerator