linux-kernel - Re: linux-next: manual merge of the security tree with the vfs tree

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 31 Dec 2015 12:45:35 +0200
From:	Petko Manolov <petkan@...-labs.com>
To:	Al Viro <viro@...IV.linux.org.uk>
Cc:	Stephen Rothwell <sfr@...b.auug.org.au>,
	James Morris <jmorris@...ei.org>, linux-next@...r.kernel.org,
	linux-kernel@...r.kernel.org, Mimi Zohar <zohar@...ux.vnet.ibm.com>
Subject: Re: linux-next: manual merge of the security tree with the vfs tree

On 15-12-31 04:30:19, Al Viro wrote:
> On Thu, Dec 31, 2015 at 03:24:53PM +1100, Stephen Rothwell wrote:
> > Hi James,
> > 
> > Today's linux-next merge of the security tree got a conflict in:
> > 
> >   security/integrity/ima/ima_fs.c
> > 
> > between commit:
> > 
> >   3bc8f29b149e ("new helper: memdup_user_nul()")
> > 
> > from the vfs tree and commit:
> > 
> >   38d859f991f3 ("IMA: policy can now be updated multiple times")
> > 
> > from the security tree.
> > 
> > I fixed it up (hopefully, see below) and can carry the fix as necessary
> > (no action is required).
>  
> > + 	res = mutex_lock_interruptible(&ima_write_mutex);
> > + 	if (res)
> > + 		return res;
> >   
> >   	if (datalen >= PAGE_SIZE)
> >   		datalen = PAGE_SIZE - 1;
> >   
> >   	/* No partial writes. */
> > + 	result = -EINVAL;
> >   	if (*ppos != 0)
> > - 		return -EINVAL;
> > + 		goto out;
> >   
> >  -	result = -ENOMEM;
> >  -	data = kmalloc(datalen + 1, GFP_KERNEL);
> >  -	if (!data)
> >  -		goto out;
> >  -
> >  -	*(data + datalen) = '\0';
> >  -
> >  -	result = -EFAULT;
> >  -	if (copy_from_user(data, buf, datalen))
> >  +	data = memdup_user_nul(buf, datalen);
> > - 	if (IS_ERR(data))
> > - 		return PTR_ERR(data);
> > ++	if (IS_ERR(data)) {
> > ++		result = PTR_ERR(data);
> > + 		goto out;
> > ++	}
> 
> Why do it in this order?  With or without opencoding memdup_user_nul(),
> what's the point of taking the mutex before copying the data from
> userland?  All it achieves is holding it longer, over the area that
> needs no exclusion whatsoever.

I introduced the write mutex when ima_write_policy() stopped being serialized by 
other means.  Come to think about it the semaphore could be taken right before 
copy_from_user() so it is my fault, not Stephen's.

The patch, however, leaves out a bug where free without allocation can occur.  
Look at *ppos evaluation.  Instead of "goto out" it should be "return -EINVAL;".  
This requires the mutex lock to be moved down, though.


cheers,
Petko
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/