From: Joerg Roedel <joro@8bytes.org>
To: Rajat Jain <rajatja@google.com>
Cc: Bjorn Helgaas <helgaas@kernel.org>,
ashok.raj@intel.com, lalithambika.krishnakumar@intel.com,
Bjorn Helgaas <bhelgaas@google.com>,
linux-pci <linux-pci@vger.kernel.org>,
Mika Westerberg <mika.westerberg@linux.intel.com>,
Jean-Philippe Brucker <jean-philippe@linaro.org>,
Prashant Malani <pmalani@google.com>,
Benson Leung <bleung@google.com>, Todd Broch <tbroch@google.com>,
Alex Levin <levinale@google.com>,
Mattias Nissler <mnissler@google.com>,
Zubin Mithra <zsm@google.com>, Rajat Jain <rajatxjain@gmail.com>,
Bernie Keany <bernie.keany@intel.com>,
Aaron Durbin <adurbin@google.com>,
Diego Rivas <diegorivas@google.com>,
Duncan Laurie <dlaurie@google.com>,
Furquan Shaikh <furquan@google.com>,
Jesse Barnes <jsbarnes@google.com>,
Christian Kellner <christian@kellner.me>,
Alex Williamson <alex.williamson@redhat.com>
Subject: Re: [RFC] Restrict the untrusted devices, to bind to only a set of "whitelisted" drivers
Date: Fri, 15 May 2020 14:44:48 +0200 [thread overview]
Message-ID: <20200515124448.GW18353@8bytes.org> (raw)
In-Reply-To: <CACK8Z6F8ncByr92+PUHPAGudZBM4fqKiau+t=JE6P1963et3fQ@mail.gmail.com>
On Wed, May 13, 2020 at 02:26:18PM -0700, Rajat Jain wrote:
> * A driver could be double fetching the memory, causing it to do
> different things than intended. E.g.
> * A driver could be (negligently) passing some kernel addresses to the device.
> * A driver could be using (for memory dereferencing, for e.g.) the
> address/indices, given by the device, without enough validation.
> * A driver may negligently be sharing the DMA memory with some other
> driver data in the same PAGE. Since the IOMMU restrictions are PAGE
> granular, this might give device access to that driver data.
The Intel IOMMU driver has a solution for that problem as it has iommu
based bounce-buffer dma ops. This means that a driver can't
accidentially share sensitive information on the same page with a
device.
This idea should be generalized and made available for all iommu-drivers
in the form of integrating it into the dma-iommu code, or have a
separate generic dma-ops implementation, which does:
1) Give the device direct access to DMA buffers if they are
IOMMU-page aligned (both start and size).
2) Use bounce buffering for DMA buffers that don't align with
iommu page-size.
This would at least eliminate this type of attack made possible by
uncautious drivers.
Regards,
Joerg
prev parent reply other threads:[~2020-05-15 12:44 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-01 23:07 [RFC] Restrict the untrusted devices, to bind to only a set of "whitelisted" drivers Rajat Jain
2020-05-04 11:47 ` Jean-Philippe Brucker
2020-05-04 11:59 ` Jean-Philippe Brucker
2020-05-04 19:17 ` Rajat Jain
2020-05-05 12:33 ` Mika Westerberg
2020-05-06 18:51 ` Rajat Jain
2020-05-11 20:31 ` Rajat Jain
2020-05-13 15:19 ` Bjorn Helgaas
2020-05-13 21:26 ` Rajat Jain
2020-05-14 13:42 ` Mika Westerberg
2020-05-14 19:12 ` Raj, Ashok
2020-05-15 2:18 ` Rajat Jain
2020-05-26 16:30 ` Rajat Jain
2020-06-01 23:25 ` Bjorn Helgaas
2020-06-02 5:06 ` Greg Kroah-Hartman
2020-06-03 2:27 ` Rajat Jain
2020-06-03 6:07 ` Greg Kroah-Hartman
2020-06-03 11:51 ` Rajat Jain
2020-06-03 12:16 ` Greg Kroah-Hartman
2020-06-03 12:57 ` Rajat Jain
2020-06-03 13:29 ` Greg Kroah-Hartman
2020-06-04 19:38 ` Rajat Jain
2020-06-05 8:02 ` Greg Kroah-Hartman
2020-06-06 1:08 ` Rajat Jain
2020-06-07 11:36 ` Greg Kroah-Hartman
2020-06-08 17:03 ` Jesse Barnes
2020-06-08 17:50 ` Greg Kroah-Hartman
2020-06-08 18:29 ` Jesse Barnes
2020-06-08 18:41 ` Rajat Jain
2020-06-09 9:54 ` Greg Kroah-Hartman
2020-06-30 21:46 ` Pavel Machek
2020-06-09 5:57 ` Greg Kroah-Hartman
2020-06-30 21:45 ` Pavel Machek
2020-07-01 6:54 ` Greg Kroah-Hartman
2020-07-01 8:47 ` Pavel Machek
2020-07-01 10:57 ` Greg Kroah-Hartman
2020-07-01 11:08 ` Pavel Machek
2020-06-09 21:04 ` Bjorn Helgaas
2020-06-09 23:23 ` Rajat Jain
2020-06-10 0:04 ` Bjorn Helgaas
2020-06-10 0:30 ` Rajat Jain
2020-06-10 20:17 ` Rajat Jain
2020-06-10 23:09 ` Bjorn Helgaas
2020-06-10 23:01 ` Bjorn Helgaas
2020-06-10 23:46 ` Rajat Jain
2020-06-10 7:13 ` Greg Kroah-Hartman
2020-06-10 1:34 ` Oliver O'Halloran
2020-06-10 19:57 ` Rajat Jain
2020-06-16 1:24 ` Rajat Jain
2020-06-10 7:12 ` Greg Kroah-Hartman
2020-05-15 12:44 ` Joerg Roedel [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200515124448.GW18353@8bytes.org \
--to=joro@8bytes.org \
--cc=adurbin@google.com \
--cc=alex.williamson@redhat.com \
--cc=ashok.raj@intel.com \
--cc=bernie.keany@intel.com \
--cc=bhelgaas@google.com \
--cc=bleung@google.com \
--cc=christian@kellner.me \
--cc=diegorivas@google.com \
--cc=dlaurie@google.com \
--cc=furquan@google.com \
--cc=helgaas@kernel.org \
--cc=jean-philippe@linaro.org \
--cc=jsbarnes@google.com \
--cc=lalithambika.krishnakumar@intel.com \
--cc=levinale@google.com \
--cc=linux-pci@vger.kernel.org \
--cc=mika.westerberg@linux.intel.com \
--cc=mnissler@google.com \
--cc=pmalani@google.com \
--cc=rajatja@google.com \
--cc=rajatxjain@gmail.com \
--cc=tbroch@google.com \
--cc=zsm@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).