Re: possible race condition in vfs code

[Mika Penttilä]

I mean busy inodes after umount doing bogus write_inode_now. Busy inodes 
don't pin the superblock (vfsmnt does but it's gone, otherwise we 
wouldn't be in kill_sb in the first place).

Dave Peterson's fix solves the double free issue, but does potentially 
introduce another busy inode after sb has gone.

--Mika

[e-mail] wrote:

>On Wed, Apr 30, 2003 at 01:51:42AM +0300, Mika Penttilä wrote:
>  
>
>>That piece of code looks wrong in other ways also..if we have unmounted 
>>an active fs (which shouldn't be done but happens!) we shouldn't be at 
>>least writing back to it anything! The !sb test isn't useful (we never 
>>clear it in live inodes) and MS_ACTIVE handling is racy as hell wrt 
>>umount...
>>    
>>
>
>Would you mind actually _reading_ kill_super() and stuff called by it?
>-
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to [e-mail]
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/
>
>  
>