Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-03 Thread Rajat Jain
On Tue, Jun 2, 2020 at 10:30 PM Mika Westerberg wrote: > > On Tue, Jun 02, 2020 at 04:26:02PM -0700, Rajat Jain wrote: > > +static bool risky_device(struct pci_dev *pdev) > > +{ > > + if (pdev->untrusted) { > > + pci_warn(pdev, > > + "Skipping IOMMU quirk for d

Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-02 Thread Mika Westerberg
On Tue, Jun 02, 2020 at 04:26:02PM -0700, Rajat Jain wrote: > +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. Plea

Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-02 Thread Prashant Malani
(Trimming text) On Wed, Jun 03, 2020 at 12:23:48AM +, Rajat Jain wrote: > On Tue, Jun 2, 2020 at 4:49 PM Prashant Malani wrote: > > > > Hi Rajat, > > Hi Prashant, thanks for taking a look. > > > > > On Tue, Jun 02, 2020 at 04:26:02PM -0700, Rajat Jain wrote: > > > +static bool risky_device(

Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-02 Thread Rajat Jain
On Tue, Jun 2, 2020 at 4:49 PM Prashant Malani 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 qu

Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-02 Thread Prashant Malani
Hi Rajat, 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 devic

Re: [PATCH v3] iommu/vt-d: Don't apply gfx quirks to untrusted devices

2020-06-02 Thread Raj, Ashok
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 ap