linux-kernel - Re: [PATCH v2 4/7] Drivers: hv: vmbus: Copy the hv_message object 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] [day] [month] [year] [list]

Message-ID: <20201206213308.GA11423@andrea>
Date:   Sun, 6 Dec 2020 22:33:08 +0100
From:   Andrea Parri <parri.andrea@...il.com>
To:     Michael Kelley <mikelley@...rosoft.com>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-hyperv@...r.kernel.org" <linux-hyperv@...r.kernel.org>,
        KY Srinivasan <kys@...rosoft.com>,
        Haiyang Zhang <haiyangz@...rosoft.com>,
        Stephen Hemminger <sthemmin@...rosoft.com>,
        Wei Liu <wei.liu@...nel.org>,
        Juan Vazquez <juvazq@...rosoft.com>,
        Saruhan Karademir <skarade@...rosoft.com>
Subject: Re: [PATCH v2 4/7] Drivers: hv: vmbus: Copy the hv_message object in
 vmbus_on_msg_dpc()

On Sun, Dec 06, 2020 at 06:39:39PM +0000, Michael Kelley wrote:
> From: Andrea Parri (Microsoft) <parri.andrea@...il.com> Sent: Wednesday, December 2, 2020 1:22 AM
> > 
> > The hv_message object is in memory shared with the host.  To prevent
> > an erroneous or a malicious host from 'corrupting' such object, copy
> > the object into private memory.

[...]

> But if we're going to just make a copy at the start and use the copy for
> everything, then the motivation for the changes in Patches 2 and 3 goes
> away.  The double-fetch problem is solved entirely by Patch 4.  The
> changes in Patches 2 and 3 *are* nice for simplifying the code, but
> that's all.  The code simplification is still useful as prep to reduce the
> number of references to "msg" that have to be changed to "msg_copy",
> but the commit message should be changed to reflect that, rather than
> to eliminate double fetches.

I'm okay with this: I will revisit the commit messages of 2/3, maybe
squash those 'simplifications' into a single patch, and then 'solve'
/guard against the double fetch problem with 4.

Thanks,
  Andrea