Re: [PATCH v4 06/10] KVM: MMU: fast path of handling guest page fault

[Marcelo Tosatti <mtosatti@redhat.com>]

On Sun, Apr 29, 2012 at 05:50:04PM +0900, Takuya Yoshikawa wrote:
> On Fri, 27 Apr 2012 11:52:13 -0300
> Marcelo Tosatti <mtosatti@redhat.com> wrote:
> 
> > Yes but the objective you are aiming for is to read and write sptes
> > without mmu_lock. That is, i am not talking about this patch. 
> > Please read carefully the two examples i gave (separated by "example)").
> 
> The real objective is not still clear.
> 
> The ~10% improvement reported before was on macro benchmarks during live
> migration.  At least, that optimization was the initial objective.
> 
> But at some point, the objective suddenly changed to "lock-less" without
> understanding what introduced the original improvement.
> 
> Was the problem really mmu_lock contention?
> 
> If the path being introduced by this patch is really fast, isn't it
> possible to achieve the same improvement still using mmu_lock?

Right. Supposedly, mmu_lock cacheline bouncing is the problem. Hum:

$ pahole -C "kvm" /tmp/kvm.ko 
struct kvm {
	spinlock_t                 mmu_lock;             /*     0     2
*/

	/* XXX 6 bytes hole, try to pack */

	struct mutex               slots_lock;           /*     8    32
*/
	struct mm_struct *         mm;                   /*    40     8
*/
	struct kvm_memslots *      memslots;             /*    48     8
*/
	struct srcu_struct         srcu;                 /*    56    48
*/
	/* --- cacheline 1 boundary (64 bytes) was 40 bytes ago --- */
	u32                        bsp_vcpu_id;          /*   104     4
*/

	/* XXX 4 bytes hole, try to pack */

Oops. False sharing?

> Note: During live migration, the fact that the guest gets faulted is
> itself a limitation.  We could easily see noticeable slowdown of a
> program even if it runs only between two GET_DIRTY_LOGs.
> 
> 
> > The rules for code under mmu_lock should be:
> > 
> > 1) Spte updates under mmu lock must always be atomic and 
> > with locked instructions.
> > 2) Spte values must be read once, and appropriate action
> > must be taken when writing them back in case their value
> > has changed (remote TLB flush might be required).
> 
> Although I am not certain about what will be really needed in the
> final form, if this kind of maybe-needed-overhead is going to be
> added little by little, I worry about possible regression.
> 
> Thanks,
> 	Takuya

Yes, that is a possibility.