[PATCH/RFC] KVM: s390: vsie/gmap: reduce gmap_rmap overhead

[PATCH/RFC] KVM: s390: vsie/gmap: reduce gmap_rmap overhead

[Christian Borntraeger]

there are cases that trigger a 2nd shadow event for the same
vmaddr/raddr combination. (prefix changes, reboots, some known races)
This will increase memory usages and it will result in long latencies
when cleaning up, e.g. on shutdown. To avoid cases with a list that has
hundreds of identical raddrs we check existing entries at insert time.
As this measurably reduces the list length this will be faster than
traversing the list at shutdown time.

In the long run several places will be optimized to create less entries
and a shrinker might be necessary.

Fixes: 4be130a08420 ("s390/mm: add shadow gmap support")
Signed-off-by: Christian Borntraeger <borntraeger@linux.ibm.com>
---
 arch/s390/mm/gmap.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c
index 69c08d966fda..0fc0c26a71f2 100644
--- a/arch/s390/mm/gmap.c
+++ b/arch/s390/mm/gmap.c
@@ -1185,12 +1185,19 @@ static inline void gmap_insert_rmap(struct gmap *sg, unsigned long vmaddr,
 				    struct gmap_rmap *rmap)
 {
 	void __rcu **slot;
+	struct gmap_rmap *temp;
 
 	BUG_ON(!gmap_is_shadow(sg));
 	slot = radix_tree_lookup_slot(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT);
 	if (slot) {
 		rmap->next = radix_tree_deref_slot_protected(slot,
 							&sg->guest_table_lock);
+		for (temp = rmap->next; temp; temp = temp->next) {
+			if (temp->raddr == rmap->raddr) {
+				kfree(rmap);
+				return;
+			}
+		}
 		radix_tree_replace_slot(&sg->host_to_rmap, slot, rmap);
 	} else {
 		rmap->next = NULL;
-- 
2.35.1

Re: [PATCH/RFC] KVM: s390: vsie/gmap: reduce gmap_rmap overhead

[David Hildenbrand]

On 29.04.22 17:15, Christian Borntraeger wrote:
> there are cases that trigger a 2nd shadow event for the same
> vmaddr/raddr combination. (prefix changes, reboots, some known races)
> This will increase memory usages and it will result in long latencies
> when cleaning up, e.g. on shutdown. To avoid cases with a list that has
> hundreds of identical raddrs we check existing entries at insert time.
> As this measurably reduces the list length this will be faster than
> traversing the list at shutdown time.
> 
> In the long run several places will be optimized to create less entries
> and a shrinker might be necessary.
> 
> Fixes: 4be130a08420 ("s390/mm: add shadow gmap support")
> Signed-off-by: Christian Borntraeger <borntraeger@linux.ibm.com>
> ---
>  arch/s390/mm/gmap.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c
> index 69c08d966fda..0fc0c26a71f2 100644
> --- a/arch/s390/mm/gmap.c
> +++ b/arch/s390/mm/gmap.c
> @@ -1185,12 +1185,19 @@ static inline void gmap_insert_rmap(struct gmap *sg, unsigned long vmaddr,
>  				    struct gmap_rmap *rmap)
>  {
>  	void __rcu **slot;
> +	struct gmap_rmap *temp;
>  
>  	BUG_ON(!gmap_is_shadow(sg));
>  	slot = radix_tree_lookup_slot(&sg->host_to_rmap, vmaddr >> PAGE_SHIFT);
>  	if (slot) {
>  		rmap->next = radix_tree_deref_slot_protected(slot,
>  							&sg->guest_table_lock);
> +		for (temp = rmap->next; temp; temp = temp->next) {
> +			if (temp->raddr == rmap->raddr) {
> +				kfree(rmap);
> +				return;
> +			}
> +		}
>  		radix_tree_replace_slot(&sg->host_to_rmap, slot, rmap);
>  	} else {
>  		rmap->next = NULL;

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Thanks,

David / dhildenb