Re: [PATCH] vfs: Fix possible chmod/truncate race condition.

[Dmitri Monakhov <dmonakhov@openvz.org>]

Andrew Morton <akpm@linux-foundation.org> writes:

> On Wed, 15 Oct 2008 23:35:27 +0400 Dmitri Monakhov <dmonakhov@openvz.org> wrote:
>
>> 
>> Signed-off-by: Dmitri Monakhov <dmonakhov@openvz.org>
>> ---
>>  fs/open.c |    3 +--
>>  1 files changed, 1 insertions(+), 2 deletions(-)
>> 
>> diff --git a/fs/open.c b/fs/open.c
>> index 07da935..3423b94 100644
>> --- a/fs/open.c
>> +++ b/fs/open.c
>> @@ -214,10 +214,9 @@ int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
>>  		newattrs.ia_valid |= ATTR_FILE;
>>  	}
>>  
>> +	mutex_lock(&dentry->d_inode->i_mutex);
>>  	/* Remove suid/sgid on truncate too */
>>  	newattrs.ia_valid |= should_remove_suid(dentry);
>> -
>> -	mutex_lock(&dentry->d_inode->i_mutex);
>>  	err = notify_change(dentry, &newattrs);
>>  	mutex_unlock(&dentry->d_inode->i_mutex);
>>  	return err;
>
> OK, I give up.  What race?
i_mode read not protected by i_mutex in should_remove_suid() and we
may have race condition with chown(). Sorry, seems I've been too crazy
about this. The truth is:
1: No one will call truncate() and chown() concurrently.
2: This race is still possible regardless to this fix.