Re: [PATCH] x86: add phys addr validity check for /dev/mem mmap

[Frantisek Hrbata]

On Wed, May 01, 2013 at 11:19:15AM -0700, Dave Hansen wrote:
> On 04/27/2013 08:17 PM, Will Huck wrote:
> >>     PTE bits 51 - M are reserved, where M is physical address width
> >> found 2)
> >>     Note: step 2) is actually not needed, we can always set just the
> >> 51th bit
> >>     (0x8000000000000)
> > 
> > What's the meaning here? You trigger oops since the address is beyond
> > max address cpu supported or access to a reserved page? If the answer is
> > the latter, I'm think it's not right. For example, the kernel code/data
> > section is reserved in memory, kernel access it will trigger oops? I
> > don't think so.
> 
> I think you're confusing the original problem here with how we would
> implement the solution.
> 
> /dev/mem essentially lets you create ptes with as large of a value as
> you like.  You just mmap() it, and the kernel will build you a pte to
> access whatever crazy offset you choose.
> 
> The problem is that on _some_ systems, you won't just get a bus error,
> the kernel actually sets up some ptes which the hardware objects to (the
> reserved bits in the pte), and we'll panic when the hardware sees the
> ptes.  We're trying to avoid these panics by ensuring that we never
> create these nasty ptes.
> 
> Those "nasty" PTEs point to memory which can not even possibly be
> *addressed* on the CPUs where they upset the hardware.  In other words,
> if we limit /dev/mem to *possible* memory on the system (which is sane
> all by itself) we will also fix this particular problem.

Hi Dave,

thank you for jumping in. This is exactly what the patch is doing. Note that the
same check is already done in ioremap. This patch just uses the same approach
and adds the same check to mmap_mem. I would like to ask what do you think about
the fix and if it looks ok to you.

Many thanks

-- 
Frantisek Hrbata
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/