Re: Redoing eXclusive Page Frame Ownership (XPFO) with isolated CPUs in mind (for KVM to isolate its guests per CPU)

[Khalid Aziz <khalid.aziz@oracle.com>]

On 08/30/2018 10:00 AM, Julian Stecklina wrote:
> Hey everyone,
> 
> On Mon, 20 Aug 2018 15:27 Linus Torvalds <torvalds@linux-foundation.org> wrote:
>> On Mon, Aug 20, 2018 at 3:02 PM Woodhouse, David <dwmw@amazon.co.uk> wrote:
>>>
>>> It's the *kernel* we don't want being able to access those pages,
>>> because of the multitude of unfixable cache load gadgets.
>>
>> Ahh.
>>
>> I guess the proof is in the pudding. Did somebody try to forward-port
>> that patch set and see what the performance is like?
> 
> I've been spending some cycles on the XPFO patch set this week. For the
> patch set as it was posted for v4.13, the performance overhead of
> compiling a Linux kernel is ~40% on x86_64[1]. The overhead comes almost
> completely from TLB flushing. If we can live with stale TLB entries
> allowing temporary access (which I think is reasonable), we can remove
> all TLB flushing (on x86). This reduces the overhead to 2-3% for
> kernel compile.
> 
> There were no problems in forward-porting the patch set to master.
> You can find the result here, including a patch makes the TLB flushing
> configurable:
> http://git.infradead.org/users/jsteckli/linux-xpfo.git/shortlog/refs/heads/xpfo-master
> 
> It survived some casual stress-ng runs. I can rerun the benchmarks on
> this version, but I doubt there is any change.
> 
>> It used to be just 500 LOC. Was that because they took horrible
>> shortcuts?
> 
> The patch is still fairly small. As for the horrible shortcuts, I let
> others comment on that.


Looks like the performance impact can be whole lot worse. On my test 
system with 2 Xeon Platinum 8160 (HT enabled) CPUs and 768 GB of memory, 
I am seeing very high penalty with XPFO when building 4.18.6 kernel 
sources with "make -j60":

              No XPFO patch   XPFO patch(No TLB flush)  XPFO(TLB Flush)
sys time      52m 54.036s       55m 47.897s              434m 8.645s

That is ~8% worse with TLB flush disabled and ~720% worse with TLB flush 
enabled. This test was with kernel sources being compiled on an ext4 
filesystem. XPFO seems to affect ext2 even more. With ext2 filesystem, 
impact was ~18.6% and ~900%.

--
Khalid