linux-kernel - Re: [PATCH 186/190] Revert "virt: vbox: Only copy_from

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]

Message-ID: <YIDOUetRRiizeOBX@zeniv-ca.linux.org.uk>
Date:   Thu, 22 Apr 2021 01:16:01 +0000
From:   Al Viro <viro@...iv.linux.org.uk>
To:     Tavis Ormandy <taviso@...il.com>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH 186/190] Revert "virt: vbox: Only copy_from_user the
 request-header once"

On Wed, Apr 21, 2021 at 03:14:29PM -0000, Tavis Ormandy wrote:
> On 2021-04-21, Greg Kroah-Hartman wrote:
> > This reverts commit bd23a7269834dc7c1f93e83535d16ebc44b75eba.
> >
> > -	*((struct vbg_ioctl_hdr *)buf) = hdr;
> > -	if (copy_from_user(buf + sizeof(hdr), (void *)arg + sizeof(hdr),
> > -			   hdr.size_in - sizeof(hdr))) {
> > +	if (copy_from_user(buf, (void *)arg, hdr.size_in)) {
> >  		ret = -EFAULT;
> >  		goto out;
> >  	}
> 
> This one seems like a real bugfix, otherwise there's a double-fetch from
> userspace, and a TOCTOU with the hdr fields that could cause a OOB read.

	ACK, except that typecasts in there are messy as hell.  But that's,
alas, consistent with the rest of the function...

	Patch itself is correct, and AFAICS Wenwen Wang <wang6495@....edu>
might be an innocent collateral damage from that mess - commits from that
source appear to be fairly well-written.