prysm-pulse/beacon-chain/cache/skip_slot_cache.go
terence tsao 0cbd8bc03d
Use root + slot for skip slot cache key (#7654)
* Update skip slot cache's key

* Add tests

* Refactor into its own function

Co-authored-by: Raul Jordan <raul@prysmaticlabs.com>
2020-10-28 03:09:05 +00:00

149 lines
3.5 KiB
Go

package cache
import (
"context"
"math"
"sync"
"time"
lru "github.com/hashicorp/golang-lru"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/promauto"
stateTrie "github.com/prysmaticlabs/prysm/beacon-chain/state"
"go.opencensus.io/trace"
)
var (
// Metrics
skipSlotCacheHit = promauto.NewCounter(prometheus.CounterOpts{
Name: "skip_slot_cache_hit",
Help: "The total number of cache hits on the skip slot cache.",
})
skipSlotCacheMiss = promauto.NewCounter(prometheus.CounterOpts{
Name: "skip_slot_cache_miss",
Help: "The total number of cache misses on the skip slot cache.",
})
)
// SkipSlotCache is used to store the cached results of processing skip slots in state.ProcessSlots.
type SkipSlotCache struct {
cache *lru.Cache
lock sync.RWMutex
disabled bool // Allow for programmatic toggling of the cache, useful during initial sync.
inProgress map[[32]byte]bool
}
// NewSkipSlotCache initializes the map and underlying cache.
func NewSkipSlotCache() *SkipSlotCache {
cache, err := lru.New(8)
if err != nil {
panic(err)
}
return &SkipSlotCache{
cache: cache,
inProgress: make(map[[32]byte]bool),
}
}
// Enable the skip slot cache.
func (c *SkipSlotCache) Enable() {
c.disabled = false
}
// Disable the skip slot cache.
func (c *SkipSlotCache) Disable() {
c.disabled = true
}
// Get waits for any in progress calculation to complete before returning a
// cached response, if any.
func (c *SkipSlotCache) Get(ctx context.Context, r [32]byte) (*stateTrie.BeaconState, error) {
ctx, span := trace.StartSpan(ctx, "skipSlotCache.Get")
defer span.End()
if c.disabled {
// Return a miss result if cache is not enabled.
skipSlotCacheMiss.Inc()
return nil, nil
}
delay := minDelay
// Another identical request may be in progress already. Let's wait until
// any in progress request resolves or our timeout is exceeded.
inProgress := false
for {
if ctx.Err() != nil {
return nil, ctx.Err()
}
c.lock.RLock()
if !c.inProgress[r] {
c.lock.RUnlock()
break
}
inProgress = true
c.lock.RUnlock()
// This increasing backoff is to decrease the CPU cycles while waiting
// for the in progress boolean to flip to false.
time.Sleep(time.Duration(delay) * time.Nanosecond)
delay *= delayFactor
delay = math.Min(delay, maxDelay)
}
span.AddAttributes(trace.BoolAttribute("inProgress", inProgress))
item, exists := c.cache.Get(r)
if exists && item != nil {
skipSlotCacheHit.Inc()
span.AddAttributes(trace.BoolAttribute("hit", true))
return item.(*stateTrie.BeaconState).Copy(), nil
}
skipSlotCacheMiss.Inc()
span.AddAttributes(trace.BoolAttribute("hit", false))
return nil, nil
}
// MarkInProgress a request so that any other similar requests will block on
// Get until MarkNotInProgress is called.
func (c *SkipSlotCache) MarkInProgress(r [32]byte) error {
if c.disabled {
return nil
}
c.lock.Lock()
defer c.lock.Unlock()
if c.inProgress[r] {
return ErrAlreadyInProgress
}
c.inProgress[r] = true
return nil
}
// MarkNotInProgress will release the lock on a given request. This should be
// called after put.
func (c *SkipSlotCache) MarkNotInProgress(r [32]byte) error {
if c.disabled {
return nil
}
c.lock.Lock()
defer c.lock.Unlock()
delete(c.inProgress, r)
return nil
}
// Put the response in the cache.
func (c *SkipSlotCache) Put(_ context.Context, r [32]byte, state *stateTrie.BeaconState) error {
if c.disabled {
return nil
}
// Copy state so cached value is not mutated.
c.cache.Add(r, state.Copy())
return nil
}