On 2/17/2022 8:04 PM, Aaron Lewis wrote:
> On Thu, Feb 17, 2022 at 3:55 AM Shukla, Manali <mashukla@xxxxxxx> wrote:
>>
>>
>>
>> On 2/15/2022 1:00 AM, Sean Christopherson wrote:
>>> On Mon, Feb 07, 2022, Manali Shukla wrote:
>>>> Add following 2 routines :
>>>> 1) set_user_mask_all() - set PT_USER_MASK for all the levels of page tables
>>>> 2) clear_user_mask_all - clear PT_USER_MASK for all the levels of page tables
>>>>
>>>> commit 916635a813e975600335c6c47250881b7a328971
>>>> (nSVM: Add test for NPT reserved bit and #NPF error code behavior)
>>>> clears PT_USER_MASK for all svm testcases. Any tests that requires
>>>> usermode access will fail after this commit.
>>>
>>> Gah, I took the easy route and it burned us. I would rather we start breaking up
>>> the nSVM and nVMX monoliths, e.g. add a separate NPT test and clear the USER flag
>>> only in that test, not the "common" nSVM test.
>>
>> Yeah. I agree. I will try to set/clear User flag in svm_npt_rsvd_bits_test() and
>> set User flag by default for all the test cases by calling setup_vm()
>> and use walk_pte() to set/clear User flag in svm_npt_rsvd_bits_test().
>>
>> Walk_pte() routine uses start address and length to walk over the memory
>> region where flag needs to be set/clear. So start address and length need to be
>> figured out to use this routine.
>
> For the start address and length you can use 'stext' and 'etext' to
> compute those. There's an example in access.c set_cr4_smep(),
> however, it uses walk_ptes() as opposed to walk_pte() (similar, but
> different).
>
Yeah PT_USER_MASK is only set for text segment in set_cr4_smep().
In #AC exception test, PT_USER_MASK needs to be set for user stack pages.
So, setting PT_USER_MASK for length('etext' - 'stext') of
memory will not resolve the problem.
Please let me know if I am missing something.
>>
>> I will work on this.
>>
>>>
>>> If we don't do that, this should really use walk_pte() instead of adding yet
>>> another PTE walker.
>>
>> -Manali
