Re: [PATCHv4 3/3] x86/tdx: Handle load_unaligned_zeropad() page-cross to a shared page

["Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>]

On Wed, Jun 15, 2022 at 08:27:52AM -0700, Dave Hansen wrote:
> On 6/14/22 05:01, Kirill A. Shutemov wrote:
> > load_unaligned_zeropad() can lead to unwanted loads across page boundaries.
> > The unwanted loads are typically harmless. But, they might be made to
> > totally unrelated or even unmapped memory. load_unaligned_zeropad()
> > relies on exception fixup (#PF, #GP and now #VE) to recover from these
> > unwanted loads.
> > 
> > In TDX guests, the second page can be shared page and VMM may configure
> > it to trigger #VE.
> > 
> > Kernel assumes that #VE on a shared page is MMIO access and tries to
> > decode instruction to handle it. In case of load_unaligned_zeropad() it
> > may result in confusion as it is not MMIO access.
> > 
> > Fix it by detecting split page MMIO accesses and fail them.
> > load_unaligned_zeropad() will recover using exception fixups.
> > 
> > The issue was discovered by analysis. It was not triggered during the
> > testing.
> 
> I thought this whole exercise was kicked off by hitting this in testing.
>  Am I remembering this wrong?
> 
> > https://lore.kernel.org/all/20220517153444.11195-10-kirill.shutemov@linux.intel.com/
> 
> Says:
> 
> > This is an actual, real-world problem which was discovered during TDX
> > testing.
> 
> Or were you considering this a different problem somehow?

They are different.

The patch by the link addresses issue of load_unaligned_zeropad() stepping
onto unaccepted memory. This was triggered in practice.

This patch address stepping onto MMIO shared memory. I had to force the
situation manually as MMIO memory mapped with ioremap() and it is not next
to normally allocated memory used by load_unaligned_zeropad() (such as
dentry cache).

Although any shared memory (SWIOTLB buffer for instance) can generate EPT
violation #VE if the VMM is malicious enough.

-- 
 Kirill A. Shutemov