[asterisk-dev] [Code Review] 3704: Infinite loop possible in ast_careful_fwrite()

Corey Farrell reviewboard at asterisk.org
Wed Jul 2 11:22:03 CDT 2014


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/3704/#review12439
-----------------------------------------------------------



trunk/main/utils.c
<https://reviewboard.asterisk.org/r/3704/#comment22703>

    If fflush does not always reset errno then this is not paranoia, it's a reaction to reality.  Start the comment with: fflush() does not...
    
    Assuming you are correct about fflush not resetting errno, we probably want "errno = 0;" before the loop as well.  This way the first loop iteration cannot see the errno from fwrite above.


- Corey Farrell


On July 2, 2014, 12:04 p.m., one47 wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/3704/
> -----------------------------------------------------------
> 
> (Updated July 2, 2014, 12:04 p.m.)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Bugs: ASTERISK-23984
>     https://issues.asterisk.org/jira/browse/ASTERISK-23984
> 
> 
> Repository: Asterisk
> 
> 
> Description
> -------
> 
> Discovered this by playing with WebRTC - It occurs sometimes if the websocket client closes the connection suddenly.
> 
> fflush() does not always reset errno correctly, particularly at EOF, and can cause a 100% CPU infinite loop.
> 
> 
> Diffs
> -----
> 
>   trunk/main/utils.c 417704 
> 
> Diff: https://reviewboard.asterisk.org/r/3704/diff/
> 
> 
> Testing
> -------
> 
> Repeated websocket opens/closes to be sure the issue is gone.
> 
> 
> Thanks,
> 
> one47
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20140702/a0828ebd/attachment.html>


More information about the asterisk-dev mailing list