On 06/30/2016 08:06 PM, Brian Gerst wrote:
>> > It's not like anybody will ever care about 32-bit page tables on
>> > Knights Landing anyway.
> Could this affect a 32-bit guest VM?
This isn't about 32-bit *mode*. It's about using the the 32-bit 2-level
_paging_ mode that supports only 4GB virtual
Dave Hansen writes:
> On 07/01/2016 07:25 AM, Eric W. Biederman wrote:
>> Linus Torvalds writes:
>>> > On Thu, Jun 30, 2016 at 9:39 PM, Dave Hansen wrote:
>>
>> I think what you suggest will work if we don't consider A/D in
>> pte_none(). I think there are a bunch of code path w
On Fri, Jul 1, 2016 at 9:14 AM, Dave Hansen wrote:
>
> As I understand it, the erratum only affects a thread which is about to
> page fault. The write associated with the dirty bit being set never
> actually gets executed. So, the bit really *is* stray and isn't
> something we need to preserve.
On 07/01/2016 09:07 AM, Linus Torvalds wrote:
> But I also started worrying about us just losing sight of the dirty
> bit in particular. It's not enough that we ignore the dirty bit - we'd
> still want to make sure that the underlying backing page gets marked
> dirty, even if the CPU is buggy and e
On Thu, Jun 30, 2016 at 9:39 PM, Dave Hansen wrote:
>
> I think what you suggest will work if we don't consider A/D in
> pte_none(). I think there are a bunch of code path where assume that
> !pte_present() && !pte_none() means swap.
Hmm.
Thinking about it some more, I still think it's a good i
On 07/01/2016 07:25 AM, Eric W. Biederman wrote:
> Linus Torvalds writes:
>> > On Thu, Jun 30, 2016 at 9:39 PM, Dave Hansen wrote:
>>> >>
>>> >> I think what you suggest will work if we don't consider A/D in
>>> >> pte_none(). I think there are a bunch of code path where assume that
>>> >> !pte_
Linus Torvalds writes:
> On Thu, Jun 30, 2016 at 9:39 PM, Dave Hansen wrote:
>>
>> I think what you suggest will work if we don't consider A/D in
>> pte_none(). I think there are a bunch of code path where assume that
>> !pte_present() && !pte_none() means swap.
>
> Yeah, we would need to chang
On Thu, Jun 30, 2016 at 9:39 PM, Dave Hansen wrote:
>
> I think what you suggest will work if we don't consider A/D in
> pte_none(). I think there are a bunch of code path where assume that
> !pte_present() && !pte_none() means swap.
Yeah, we would need to change pte_none() to mask off D/A, but
On 06/30/2016 07:55 PM, Linus Torvalds wrote:
> On Thu, Jun 30, 2016 at 5:12 PM, Dave Hansen wrote:
>> From: Dave Hansen
>> The Intel(R) Xeon Phi(TM) Processor x200 Family (codename: Knights
>> Landing) has an erratum where a processor thread setting the Accessed
>> or Dirty bits may not do so at
On Thu, Jun 30, 2016 at 10:55 PM, Linus Torvalds
wrote:
> On Thu, Jun 30, 2016 at 5:12 PM, Dave Hansen wrote:
>>
>> From: Dave Hansen
>>
>> The Intel(R) Xeon Phi(TM) Processor x200 Family (codename: Knights
>> Landing) has an erratum where a processor thread setting the Accessed
>> or Dirty bits
On Thu, Jun 30, 2016 at 5:12 PM, Dave Hansen wrote:
>
> From: Dave Hansen
>
> The Intel(R) Xeon Phi(TM) Processor x200 Family (codename: Knights
> Landing) has an erratum where a processor thread setting the Accessed
> or Dirty bits may not do so atomically against its checks for the
> Present bi
Dave Hansen wrote:
> +pte_t ptep_clear_flush(struct vm_area_struct *vma, unsigned long address,
> +pte_t *ptep)
> +{
> + struct mm_struct *mm = vma->vm_mm;
> + pte_t pte;
> +
> + pte = ptep_get_and_clear(mm, address, ptep);
> + if (pte_accessible(mm, pte)) {
>
On 06/30/2016 06:50 PM, Nadav Amit wrote:
> Dave Hansen wrote:
>> +pte_t ptep_clear_flush(struct vm_area_struct *vma, unsigned long address,
>> + pte_t *ptep)
>> +{
>> +struct mm_struct *mm = vma->vm_mm;
>> +pte_t pte;
>> +
>> +pte = ptep_get_and_clear(mm, address, pt
From: Dave Hansen
The Intel(R) Xeon Phi(TM) Processor x200 Family (codename: Knights
Landing) has an erratum where a processor thread setting the Accessed
or Dirty bits may not do so atomically against its checks for the
Present bit. This may cause a thread (which is about to page fault)
to set
14 matches
Mail list logo
