Message-ID: <CAA93t1r9Kg0OFwoOhNUKYeSbZrFgu5k_eexer34msjsLBNohcg@mail.gmail.com>
Date:   Wed, 3 Jun 2020 00:23:48 +0000
From:   Rajat Jain <rajatxjain@...il.com>
To:     Prashant Malani <pmalani@...gle.com>
Cc:     Rajat Jain <rajatja@...gle.com>,
        David Woodhouse <dwmw2@...radead.org>,
        Lu Baolu <baolu.lu@...ux.intel.com>,
        Joerg Roedel <joro@...tes.org>,
        iommu@...ts.linux-foundation.org,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Mika Westerberg <mika.westerberg@...el.com>,
        Ashok Raj <ashok.raj@...el.com>,
        lalithambika.krishnakumar@...el.com, bleung@...gle.com,
        levinale@...gle.com, zsm@...gle.com, mnissler@...gle.com,
        tbroch@...gle.com
Subject: Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

On Tue, Jun 2, 2020 at 4:49 PM Prashant Malani <pmalani@...gle.com> wrote:
>
> Hi Rajat,

Hi Prashant, thanks for taking a look.

>
> On Tue, Jun 02, 2020 at 04:26:02PM -0700, Rajat Jain wrote:
> > Currently, an external malicious PCI device can masquerade the VID:PID
> > of faulty gfx devices, and thus apply iommu quirks to effectively
> > disable the IOMMU restrictions for itself.
> >
> > Thus we need to ensure that the device we are applying quirks to, is
> > indeed an internal trusted device.
> >
> > Signed-off-by: Rajat Jain <rajatja@...gle.com>
> > Acked-by: Lu Baolu <baolu.lu@...ux.intel.com>
> > ---
> > v3: - Separate out the warning mesage in a function to be called from
> >       other places. Change the warning string as suggested.
> > v2: - Change the warning print strings.
> >     - Add Lu Baolu's acknowledgement.
> >
> >  drivers/iommu/intel-iommu.c | 37 +++++++++++++++++++++++++++++++++++++
> >  1 file changed, 37 insertions(+)
> >
> > diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
> > index ef0a5246700e5..dc859f02985a0 100644
> > --- a/drivers/iommu/intel-iommu.c
> > +++ b/drivers/iommu/intel-iommu.c
> > @@ -6185,6 +6185,23 @@ intel_iommu_domain_set_attr(struct iommu_domain *domain,
> >       return ret;
> >  }
> >
> > +/*
> > + * Check that the device does not live on an external facing PCI port that is
> > + * marked as untrusted. Such devices should not be able to apply quirks and
> > + * thus not be able to bypass the IOMMU restrictions.
> > + */
> > +static bool risky_device(struct pci_dev *pdev)
> > +{
> > +     if (pdev->untrusted) {
> > +             pci_warn(pdev,
> > +                      "Skipping IOMMU quirk for dev (%04X:%04X) on untrusted"
> > +                      " PCI link. Please check with your BIOS/Platform"
> > +                      " vendor about this\n", pdev->vendor, pdev->device);
> > +             return true;
> > +     }
> > +     return false;
> minor suggestion: Perhaps you could use a guard clause here? It would save you
> a level of indentation, and possibly allow better string splitting
> (e.g keeping "untrusted PCI" together). So something like:
>
> if (!pdev->untrusted)
>         return false;

I personally have found double negation expressions always confusing,
even if negation is part of the variable. (For e.g. I have found I
need to be always stop and convince myself that:

"if (!pdev->untrusted)"
    <do something>

conceptually implies

"if (pdev->trusted)".
    <do something>

So I tend to keep negations to minimum. In this case, it doesn't buy
us much either, so I'd prefer to keep it the same unless there are
more opinions on this. OTOH I don't mind changing it too if you feel
strongly about this.

Thanks,

Rajat


>
> pci_warn(...);
>
> I also hear the column limit warning is now for 100 chars [1], though
> I'm not sure how it's being handled in this file.
>
> Best regards,
>
> -Prashant
>
> [1]:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/Documentation/process/coding-style.rst?id=bdc48fa11e46f867ea4d75fa59ee87a7f48be144
>
> > +}
> > +
> >  const struct iommu_ops intel_iommu_ops = {
> >       .capable                = intel_iommu_capable,
> >       .domain_alloc           = intel_iommu_domain_alloc,
> > @@ -6214,6 +6231,9 @@ const struct iommu_ops intel_iommu_ops = {
> >
> >  static void quirk_iommu_igfx(struct pci_dev *dev)
> >  {
> > +     if (risky_device(dev))
> > +             return;
> > +
> >       pci_info(dev, "Disabling IOMMU for graphics on this chipset\n");
> >       dmar_map_gfx = 0;
> >  }
> > @@ -6255,6 +6275,9 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x163D, quirk_iommu_igfx);
> >
> >  static void quirk_iommu_rwbf(struct pci_dev *dev)
> >  {
> > +     if (risky_device(dev))
> > +             return;
> > +
> >       /*
> >        * Mobile 4 Series Chipset neglects to set RWBF capability,
> >        * but needs it. Same seems to hold for the desktop versions.
> > @@ -6285,6 +6308,9 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev)
> >  {
> >       unsigned short ggc;
> >
> > +     if (risky_device(dev))
> > +             return;
> > +
> >       if (pci_read_config_word(dev, GGC, &ggc))
> >               return;
> >
> > @@ -6318,6 +6344,12 @@ static void __init check_tylersburg_isoch(void)
> >       pdev = pci_get_device(PCI_VENDOR_ID_INTEL, 0x3a3e, NULL);
> >       if (!pdev)
> >               return;
> > +
> > +     if (risky_device(pdev)) {
> > +             pci_dev_put(pdev);
> > +             return;
> > +     }
> > +
> >       pci_dev_put(pdev);
> >
> >       /* System Management Registers. Might be hidden, in which case
> > @@ -6327,6 +6359,11 @@ static void __init check_tylersburg_isoch(void)
> >       if (!pdev)
> >               return;
> >
> > +     if (risky_device(pdev)) {
> > +             pci_dev_put(pdev);
> > +             return;
> > +     }
> > +
> >       if (pci_read_config_dword(pdev, 0x188, &vtisochctrl)) {
> >               pci_dev_put(pdev);
> >               return;
> > --
> > 2.27.0.rc2.251.g90737beb825-goog
> >

Powered by blists - more mailing lists