linux-kernel - Re: [PATCH v2 2/7] Drivers: hv: vmbus: Avoid double fetch of msgtype in vmbus_on_msg

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:   Wed, 2 Dec 2020 15:14:33 +0100
From:   Andrea Parri <parri.andrea@...il.com>
To:     Wei Liu <wei.liu@...nel.org>
Cc:     linux-kernel@...r.kernel.org, linux-hyperv@...r.kernel.org,
        "K . Y . Srinivasan" <kys@...rosoft.com>,
        Haiyang Zhang <haiyangz@...rosoft.com>,
        Stephen Hemminger <sthemmin@...rosoft.com>,
        Michael Kelley <mikelley@...rosoft.com>,
        Juan Vazquez <juvazq@...rosoft.com>,
        Saruhan Karademir <skarade@...rosoft.com>
Subject: Re: [PATCH v2 2/7] Drivers: hv: vmbus: Avoid double fetch of msgtype
 in vmbus_on_msg_dpc()

On Wed, Dec 02, 2020 at 01:40:04PM +0000, Wei Liu wrote:
> On Wed, Dec 02, 2020 at 02:37:16PM +0100, Andrea Parri wrote:
> > > > @@ -1072,12 +1073,19 @@ void vmbus_on_msg_dpc(unsigned long data)
> > > >  		/* no msg */
> > > >  		return;
> > > >  
> > > > +	/*
> > > > +	 * The hv_message object is in memory shared with the host.  The host
> > > > +	 * could erroneously or maliciously modify such object.  Make sure to
> > > > +	 * validate its fields and avoid double fetches whenever feasible.
> > > > +	 */
> > > > +
> > > >  	hdr = (struct vmbus_channel_message_header *)msg->u.payload;
> > > > +	msgtype = hdr->msgtype;
> > > 
> > > Should READ_ONCE be used here?
> > 
> > I think it should.  Thank you for pointing this out.
> 
> Glad I can help.
> 
> The same comment applies to other patches as well, of course.

(As discussed offline/for reference:) I can spot a similar case in
patch #3; however, #4 is supposed to make that access 'non-shared'.

I should probably just squash patches #3 and #4; I'll try to do so
in v3...

Thanks,
  Andrea