Re: [PATCH] kvm: don't call mmu_shrinker w/o used_mmu_pages

[Ying Han <yinghan@google.com>]

On Sat, Apr 14, 2012 at 4:44 AM, Hillf Danton <dhillf@gmail.com> wrote:
> On Sat, Apr 14, 2012 at 6:38 AM, Ying Han <yinghan@google.com> wrote:
>> The mmu_shrink() is heavy by itself by iterating all kvms and holding
>> the kvm_lock. spotted the code w/ Rik during LSF, and it turns out we
>> don't need to call the shrinker if nothing to shrink.
>>
>> Signed-off-by: Ying Han <yinghan@google.com>
>> ---
>>  arch/x86/kvm/mmu.c |   10 +++++++++-
>>  1 files changed, 9 insertions(+), 1 deletions(-)
>>
>> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
>> index 4cb1642..7025736 100644
>> --- a/arch/x86/kvm/mmu.c
>> +++ b/arch/x86/kvm/mmu.c
>> @@ -188,6 +188,11 @@ static u64 __read_mostly shadow_mmio_mask;
>>
>>  static void mmu_spte_set(u64 *sptep, u64 spte);
>>
>> +static inline int get_kvm_total_used_mmu_pages()
>> +{
>> +       return percpu_counter_read_positive(&kvm_total_used_mmu_pages);
>> +}
>> +
>>  void kvm_mmu_set_mmio_spte_mask(u64 mmio_mask)
>>  {
>>        shadow_mmio_mask = mmio_mask;
>> @@ -3900,6 +3905,9 @@ static int mmu_shrink(struct shrinker *shrink, struct shrink_control *sc)
>>        if (nr_to_scan == 0)
>>                goto out;
>>
>> +       if (!get_kvm_total_used_mmu_pages())
>> +               return 0;
>> +
>>        raw_spin_lock(&kvm_lock);
>>
>>        list_for_each_entry(kvm, &vm_list, vm_list) {
>> @@ -3926,7 +3934,7 @@ static int mmu_shrink(struct shrinker *shrink, struct shrink_control *sc)
>>        raw_spin_unlock(&kvm_lock);
>>
>>  out:
>> -       return percpu_counter_read_positive(&kvm_total_used_mmu_pages);
>> +       return get_kvm_total_used_mmu_pages();
>>  }
>>
> Just nitpick.
> If new helper not created, there is only one hunk needed.

Hmm, thought it looks nicer with the helpful function instead of long
percpu_counter_read_positive() in the if block.

>
> btw, make sense to check nr_to_scan while scanning vm_list, and bail out
> if it hits zero?

Not totally understand the nr_to_scan in that function, but we could
do a separate patch if that is needed.

--Ying
>
> Good Weekend
> -hd

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>