Re: [PATCH RFC v7 01/64] KVM: Fix memslot boundary condition for large page

[Nikunj A. Dadhania]

On 05/01/23 09:04, Chao Peng wrote:
> On Wed, Jan 04, 2023 at 12:01:05PM +0000, Jarkko Sakkinen wrote:
>> On Wed, Dec 14, 2022 at 01:39:53PM -0600, Michael Roth wrote:
>>> From: Nikunj A Dadhania <nikunj@xxxxxxx>
>>>
>>> Aligned end boundary causes a kvm crash, handle the case.
>>>
>>
>> Link: https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flore.kernel.org%2Fkvm%2F20221202061347.1070246-8-chao.p.peng%40linux.intel.com%2F&data=05%7C01%7Cnikunj.dadhania%40amd.com%7C7a95933fac1b433e339c08daeece6c2c%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C638084867591405299%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=vDEu9Uxs0QRdzbUkJbE2LsJnMHJJHBdQijkePbE2woc%3D&reserved=0
>>
>> Chao, are you aware of this issue already?
> 
> Thanks Jarkko adding me. I'm not aware of there is a fix.

It was discussed here: https://lore.kernel.org/all/e234d307-0b05-6548-5882-c24fc32c8e77@xxxxxxx/

I was hitting this with one of the selftests case.

> 
>>
>>> Signed-off-by: Nikunj A Dadhania <nikunj@xxxxxxx>
>>> Signed-off-by: Michael Roth <michael.roth@xxxxxxx>
>>> ---
>>>  arch/x86/kvm/mmu/mmu.c | 3 +++
>>>  1 file changed, 3 insertions(+)
>>>
>>> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
>>> index b1953ebc012e..b3ffc61c668c 100644
>>> --- a/arch/x86/kvm/mmu/mmu.c
>>> +++ b/arch/x86/kvm/mmu/mmu.c
>>> @@ -7159,6 +7159,9 @@ static void kvm_update_lpage_private_shared_mixed(struct kvm *kvm,
>>>  		for (gfn = first + pages; gfn < last; gfn += pages)
>>>  			linfo_set_mixed(gfn, slot, level, false);
>>>  
>>> +		if (gfn == last)
>>> +			goto out;
>>> +
> 
> Nikunj or Michael, could you help me understand in which case it causes
> a KVM crash? To me, even the end is aligned to huge page boundary, but:
>     last = (end - 1) & mask;
> so 'last' is the base address for the last effective huage page. Even
> when gfn == last, it should still a valid page and needs to be updated
> for mem_attrs, correct?

Yes, that is correct with: last = (end - 1) & mask;

We can drop this patch from SNP series.

Regards
Nikunj