[PATCH] drm/ttm: set TTM allocated pages as reserved

[Paolo Bonzini]

On 3/29/23 16:28, Paolo Bonzini wrote:
> On 3/29/23 15:54, Christian König wrote:
>> KVM tries to grab references to pages in VMAs marked with VM_PFNMAP.
>> This is illegal and can cause data corruption with TTM pages because
>> only some of them are actually reference counted.

After some other offlist discussions, I also would like to understand 
what you mean by corruption.

First, is it a _host_ corruption or a guest corruption/crash?  A guest 
crash would be KVM doing exactly what it's meant to do: it detects the 
non-reserved, non-refcounted page and refuses to map it into the guest.

On the other hand, if it is a host crash, my understanding is that an 
order>0 allocation leaves the tail pages with a zero reference count 
(and without a compound_head if, as in this case, __GFP_COMP is unset). 
If that's correct, more analysis is needed to understand why 
get_page_unless_zero() isn't rejecting the tail pages.

Paolo


>> Mark all pages allocated by TTM as reserved, this way KVM handles the
>> PFNs like they would point to MMIO space.
>>
>> This still results in a warning, but at least no other problem.
> 
> What warning is it?
> 
> Paolo
> 
>> Signed-off-by: Christian König<christian.koenig at amd.com>
>