Re: [FreeBSD 5.3-RC2] Processes STILL hanging in unkillable state

From: David Xu <davidxu_at_freebsd.org>
Date: Fri, 05 Nov 2004 06:16:51 +0800
Marc Ramirez wrote:

>Thanks for looking!
>
>On Thursday 04 November 2004 03:47 pm, Julian Elischer wrote:
>  
>
>>>  thread 0xc2081e10 ksegrp 0xc2154310 [SUSP]
>>>  thread 0xc187a960 ksegrp 0xc2154310 [SLPQ spltwt 0xc138fd98][SLP]
>>>  thread 0xc2156960 ksegrp 0xc2154310 [SUSP]
>>>  thread 0xc2157000 ksegrp 0xc2154310 [SUSP]
>>>  thread 0xc2157190 ksegrp 0xc2154310 [SLPQ select 0xc08e44c4][SLP][SUSP]
>>>  thread 0xc2156190 ksegrp 0xc2154310 [SLPQ accept 0xc216ae26][SLP][SUSP]
>>>  thread 0xc2156320 ksegrp 0xc187dc40 [SUSP]
>>>      
>>>
>>showing the output of
>>show thread 0xc2081e10
>>(etc)
>>for each thread
>>should show the backtrace of each thread (I think)
>>    
>>
>
>The output of this is at bottom.
>  
>
>>I suspect that all the threads are waiting for thread 0xc187a960 to wake
>>up and suspend
>>for some single-threading purpose.
>>but it is hard to tell.
>>    
>>

I have committed a fix, can you try ?

>davidxu     2004-11-04 22:13:16 UTC
>
>  FreeBSD src repository
>
>  Modified files:
>    sys/kern             kern_thread.c 
>  Log:
>  Don't forget to turn off P_SINGLE_BOUNDARY for thread_single(SINGLE_EXIT),
>  otherwise a threaded process which calls execv() will hang in kernel and
>  may can not be killed!
>  
>  Revision  Changes    Path
>  1.205     +1 -1      src/sys/kern/kern_thread.c
>
Received on Thu Nov 04 2004 - 21:16:53 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:21 UTC