Interrupt spawns Zombies and Vampires

Peter Renzland peter at dancing.org
Tue Jan 13 02:37:33 UTC 2009


OK.  Attaching the config_g

Shell is ash:

BusyBox v1.12.2 (2008-12-05 06:20:11 PST) built-in shell (ash)
also
BusyBox v1.12.3 (2008-12-14 02:54:58 PST) built-in shell (ash)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: config_g
Type: application/octet-stream
Size: 24260 bytes
Desc: not available
Url : http://lists.busybox.net/pipermail/busybox/attachments/20090112/f38feb04/attachment-0001.obj 
-------------- next part --------------




Thanks,

Peter



On 09  Jan 12, at 20:45 , Denys Vlasenko wrote:

> On Monday 12 January 2009 22:38, Peter Renzland wrote:
>> When I interrupt (^C) a looping process (ping, tail -f, or a script
>> with a while or for loop in it),
>
> busybox version?
> .config?
> which shell do you use? (busybox has three)
>
>> very strange things happen,
>> including ...
>>
>> * a phantom sh process is created which has my shell process as its
>> parent
>> * the next command line goes "nowhere"
>> * command history is partly deleted, and CWD may change to an earlier
>> one
>> * if the killed script reads from STDIO, it may grab some of my
>> command lines.
>> * STTY settings may be erratic -- e.g. no more newlines
>> * sometimes zombies result.
>>
>> Attempts to kill these processes may succeed, for some, but one  
>> always
>> remains, and each time it is killed, it reappears as a new process,
>> with the killing shell as its parent.
>>
>> If these loops are run in the BG, they can be killed ok.
>>
>> I can provide transcripts if necessary, but I thought I'd start by
>> asking if this is a known condition, perhaps caused by inconsistent
>> busybox shell feature configuration?
>
> It may be the symproms of standalone mode. I desperately
> need more bug reports to debug it.
>
> Please provide more info (see questions above).
> --
> vda
>



More information about the busybox mailing list