[dependencies] update go-store, go-mutexes (#422)

* update go-store, go-mutexes

Signed-off-by: kim <grufwub@gmail.com>

* update vendored code

Signed-off-by: kim <grufwub@gmail.com>

go.mod

@@ -5,7 +5,7 @@ go 1.17
 require (
 	codeberg.org/gruf/go-errors v1.0.5
 	codeberg.org/gruf/go-runners v1.2.0
-	codeberg.org/gruf/go-store v1.3.3
+	codeberg.org/gruf/go-store v1.3.6
 	github.com/ReneKroon/ttlcache v1.7.0
 	github.com/buckket/go-blurhash v1.1.0
 	github.com/coreos/go-oidc/v3 v3.1.0
@@ -49,10 +49,11 @@ require (
 
 require (
 	codeberg.org/gruf/go-bytes v1.0.2 // indirect
+	codeberg.org/gruf/go-fastcopy v1.1.1 // indirect
 	codeberg.org/gruf/go-fastpath v1.0.2 // indirect
 	codeberg.org/gruf/go-format v1.0.3 // indirect
 	codeberg.org/gruf/go-hashenc v1.0.1 // indirect
-	codeberg.org/gruf/go-mutexes v1.1.0 // indirect
+	codeberg.org/gruf/go-mutexes v1.1.2 // indirect
 	codeberg.org/gruf/go-pools v1.0.2 // indirect
 	github.com/aymerick/douceur v0.2.0 // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect

go.sum

@@ -53,6 +53,8 @@ codeberg.org/gruf/go-bytes v1.0.2/go.mod h1:1v/ibfaosfXSZtRdW2rWaVrDXMc9E3bsi/M9
 codeberg.org/gruf/go-cache v1.1.2/go.mod h1:/Dbc+xU72Op3hMn6x2PXF3NE9uIDFeS+sXPF00hN/7o=
 codeberg.org/gruf/go-errors v1.0.5 h1:rxV70oQkfasUdggLHxOX2QAoJOMFM7XWxHQR45Zx/Fg=
 codeberg.org/gruf/go-errors v1.0.5/go.mod h1:n03EpmvcmfzU3/xJKC0XXtleXXJUNFpT2fgISODvZ1Y=
+codeberg.org/gruf/go-fastcopy v1.1.1 h1:HhPCeFdVR5pwiSVDnQEGJ+J2ny9b5QgfiESc0zrWQAY=
+codeberg.org/gruf/go-fastcopy v1.1.1/go.mod h1:GDDYR0Cnb3U/AIfGM3983V/L+GN+vuwVMvrmVABo21s=
 codeberg.org/gruf/go-fastpath v1.0.1/go.mod h1:edveE/Kp3Eqi0JJm0lXYdkVrB28cNUkcb/bRGFTPqeI=
 codeberg.org/gruf/go-fastpath v1.0.2 h1:O3nuYPMXnN89dsgAwVFU5iCGINtPJdITWmbRe2an/iQ=
 codeberg.org/gruf/go-fastpath v1.0.2/go.mod h1:edveE/Kp3Eqi0JJm0lXYdkVrB28cNUkcb/bRGFTPqeI=
@@ -60,8 +62,8 @@ codeberg.org/gruf/go-format v1.0.3 h1:WoUGzTwZe6SIhILNvtr0qNIA7BOOCgdBlk5bUrfeii
 codeberg.org/gruf/go-format v1.0.3/go.mod h1:k3TLXp1dqAXdDqxlon0yEM+3FFHdNn0D6BVJTwTy5As=
 codeberg.org/gruf/go-hashenc v1.0.1 h1:EBvNe2wW8IPMUqT1XihB6/IM6KMJDLMFBxIUvmsy1f8=
 codeberg.org/gruf/go-hashenc v1.0.1/go.mod h1:IfHhPCVScOiYmJLqdCQT9bYVS1nxNTV4ewMUvFWDPtc=
-codeberg.org/gruf/go-mutexes v1.1.0 h1:kMVWHLxdfGEZTetNVRncdBMeqS4M8dSJxSGbRYXyvKk=
-codeberg.org/gruf/go-mutexes v1.1.0/go.mod h1:1j/6/MBeBQUedAtAtysLLnBKogfOZAxdym0E3wlaBD8=
+codeberg.org/gruf/go-mutexes v1.1.2 h1:AMC1CFV6kMi+iBjR3yQv8yIagG3lWm68U6sQHYFHEf4=
+codeberg.org/gruf/go-mutexes v1.1.2/go.mod h1:1j/6/MBeBQUedAtAtysLLnBKogfOZAxdym0E3wlaBD8=
 codeberg.org/gruf/go-nowish v1.0.0/go.mod h1:70nvICNcqQ9OHpF07N614Dyk7cpL5ToWU1K1ZVCec2s=
 codeberg.org/gruf/go-nowish v1.1.0/go.mod h1:70nvICNcqQ9OHpF07N614Dyk7cpL5ToWU1K1ZVCec2s=
 codeberg.org/gruf/go-pools v1.0.2 h1:B0X6yoCL9FVmnvyoizb1SYRwMYPWwEJBjPnBMM5ILos=
@@ -69,8 +71,8 @@ codeberg.org/gruf/go-pools v1.0.2/go.mod h1:MjUV3H6IASyBeBPCyCr7wjPpSNu8E2N87LG4
 codeberg.org/gruf/go-runners v1.1.1/go.mod h1:9gTrmMnO3d+50C+hVzcmGBf+zTuswReS278E2EMvnmw=
 codeberg.org/gruf/go-runners v1.2.0 h1:tkoPrwYMkVg1o/C4PGTR1YbC11XX4r06uLPOYajBsH4=
 codeberg.org/gruf/go-runners v1.2.0/go.mod h1:9gTrmMnO3d+50C+hVzcmGBf+zTuswReS278E2EMvnmw=
-codeberg.org/gruf/go-store v1.3.3 h1:fAP9FXy6HiLPxdD7cmpSzyfKXmVvZLjqn0m7HhxVT5M=
-codeberg.org/gruf/go-store v1.3.3/go.mod h1:g4+9h3wbwZ6IW0uhpw57xywcqiy4CIj0zQLqqtjEU1M=
+codeberg.org/gruf/go-store v1.3.6 h1:OKzdvfUC+nvsWV5FiSKdk+85yvxF2Tb7K5ZtRqlDBDU=
+codeberg.org/gruf/go-store v1.3.6/go.mod h1:a4vJtZf61UyrsejskX8q+s0lZeNGy7cJLUZt+fH00wo=
 dmitri.shuralyov.com/gpu/mtl v0.0.0-20190408044501-666a987793e9/go.mod h1:H6x//7gZCb22OMCxBHrMx7a5I7Hp++hsVxbQ4BYO7hU=
 github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
 github.com/BurntSushi/xgb v0.0.0-20160522181843-27f122750802/go.mod h1:IVnqGOEym/WlBOVXweHU+Q+/VP0lqqI8lqeDx9IjBqo=

vendor/codeberg.org/gruf/go-fastcopy/README.md

@@ -0,0 +1,3 @@
+# go-fastcopy
+
+An `io.Copy()` implementation that uses a memory pool for the copy buffer.

vendor/codeberg.org/gruf/go-fastcopy/copy.go

@@ -0,0 +1,134 @@
+package fastcopy
+
+import (
+	"io"
+	"sync"
+	_ "unsafe" // link to io.errInvalidWrite.
+)
+
+var (
+	// global pool instance.
+	pool = CopyPool{size: 4096}
+
+	//go:linkname errInvalidWrite io.errInvalidWrite
+	errInvalidWrite error
+)
+
+// CopyPool provides a memory pool of byte
+// buffers for io copies from readers to writers.
+type CopyPool struct {
+	size int
+	pool sync.Pool
+}
+
+// See CopyPool.Buffer().
+func Buffer(sz int) int {
+	return pool.Buffer(sz)
+}
+
+// See CopyPool.CopyN().
+func CopyN(dst io.Writer, src io.Reader, n int64) (int64, error) {
+	return pool.CopyN(dst, src, n)
+}
+
+// See CopyPool.Copy().
+func Copy(dst io.Writer, src io.Reader) (int64, error) {
+	return pool.Copy(dst, src)
+}
+
+// Buffer sets the pool buffer size to allocate. Returns current size.
+// Note this is NOT atomically safe, please call BEFORE other calls to CopyPool.
+func (cp *CopyPool) Buffer(sz int) int {
+	if sz > 0 {
+		// update size
+		cp.size = sz
+	} else if cp.size < 1 {
+		// default size
+		return 4096
+	}
+	return cp.size
+}
+
+// CopyN performs the same logic as io.CopyN(), with the difference
+// being that the byte buffer is acquired from a memory pool.
+func (cp *CopyPool) CopyN(dst io.Writer, src io.Reader, n int64) (int64, error) {
+	written, err := cp.Copy(dst, io.LimitReader(src, n))
+	if written == n {
+		return n, nil
+	}
+	if written < n && err == nil {
+		// src stopped early; must have been EOF.
+		err = io.EOF
+	}
+	return written, err
+}
+
+// Copy performs the same logic as io.Copy(), with the difference
+// being that the byte buffer is acquired from a memory pool.
+func (cp *CopyPool) Copy(dst io.Writer, src io.Reader) (int64, error) {
+	// Prefer using io.WriterTo to do the copy (avoids alloc + copy)
+	if wt, ok := src.(io.WriterTo); ok {
+		return wt.WriteTo(dst)
+	}
+
+	// Prefer using io.ReaderFrom to do the copy.
+	if rt, ok := dst.(io.ReaderFrom); ok {
+		return rt.ReadFrom(src)
+	}
+
+	var buf []byte
+
+	if b, ok := cp.pool.Get().([]byte); ok {
+		// Acquired buf from pool
+		buf = b
+	} else {
+		// Allocate new buffer of size
+		buf = make([]byte, cp.Buffer(0))
+	}
+
+	// Defer release to pool
+	defer cp.pool.Put(buf)
+
+	var n int64
+	for {
+		// Perform next read into buf
+		nr, err := src.Read(buf)
+		if nr > 0 {
+			// We error check AFTER checking
+			// no. read bytes so incomplete
+			// read still gets written up to nr.
+
+			// Perform next write from buf
+			nw, ew := dst.Write(buf[0:nr])
+
+			// Check for valid write
+			if nw < 0 || nr < nw {
+				if ew == nil {
+					ew = errInvalidWrite
+				}
+				return n, ew
+			}
+
+			// Incr total count
+			n += int64(nw)
+
+			// Check write error
+			if ew != nil {
+				return n, ew
+			}
+
+			// Check unequal read/writes
+			if nr != nw {
+				return n, io.ErrShortWrite
+			}
+		}
+
+		// Return on err
+		if err != nil {
+			if err == io.EOF {
+				err = nil // expected
+			}
+			return n, err
+		}
+	}
+}

vendor/codeberg.org/gruf/go-mutexes/debug.go

@@ -0,0 +1,39 @@
+package mutexes
+
+// func init() {
+// 	log.SetFlags(log.Flags() | log.Lshortfile)
+// }
+
+// type debugMutex sync.Mutex
+
+// func (mu *debugMutex) Lock() {
+// 	log.Output(2, "Lock()")
+// 	(*sync.Mutex)(mu).Lock()
+// }
+
+// func (mu *debugMutex) Unlock() {
+// 	log.Output(2, "Unlock()")
+// 	(*sync.Mutex)(mu).Unlock()
+// }
+
+// type debugRWMutex sync.RWMutex
+
+// func (mu *debugRWMutex) Lock() {
+// 	log.Output(2, "Lock()")
+// 	(*sync.RWMutex)(mu).Lock()
+// }
+
+// func (mu *debugRWMutex) Unlock() {
+// 	log.Output(2, "Unlock()")
+// 	(*sync.RWMutex)(mu).Unlock()
+// }
+
+// func (mu *debugRWMutex) RLock() {
+// 	log.Output(2, "RLock()")
+// 	(*sync.RWMutex)(mu).RLock()
+// }
+
+// func (mu *debugRWMutex) RUnlock() {
+// 	log.Output(2, "RUnlock()")
+// 	(*sync.RWMutex)(mu).RUnlock()
+// }

vendor/codeberg.org/gruf/go-mutexes/map.go

@@ -6,260 +6,347 @@
 	"sync/atomic"
 )
 
-// locktype defines maskable mutexmap lock types.
-type locktype uint8
-
 const (
 	// possible lock types.
-	lockTypeRead  = locktype(1) << 0
-	lockTypeWrite = locktype(1) << 1
-	lockTypeMap   = locktype(1) << 2
+	lockTypeRead  = uint8(1) << 0
+	lockTypeWrite = uint8(1) << 1
+	lockTypeMap   = uint8(1) << 2
 
 	// possible mutexmap states.
 	stateUnlockd = uint8(0)
 	stateRLocked = uint8(1)
 	stateLocked  = uint8(2)
 	stateInUse   = uint8(3)
+
+	// default values.
+	defaultWake = 1024
 )
 
-// permitLockType returns if provided locktype is permitted to go ahead in current state.
-func permitLockType(state uint8, lt locktype) bool {
+// acquireState attempts to acquire required map state for lockType.
+func acquireState(state uint8, lt uint8) (uint8, bool) {
 	switch state {
 	// Unlocked state
 	// (all allowed)
 	case stateUnlockd:
-		return true
 
 	// Keys locked, no state lock.
 	// (don't allow map locks)
 	case stateInUse:
-		return lt&lockTypeMap == 0
+		if lt&lockTypeMap != 0 {
+			return 0, false
+		}
 
 	// Read locked
 	// (only allow read locks)
 	case stateRLocked:
-		return lt&lockTypeRead != 0
+		if lt&lockTypeRead == 0 {
+			return 0, false
+		}
 
 	// Write locked
 	// (none allowed)
 	case stateLocked:
-		return false
+		return 0, false
 
 	// shouldn't reach here
 	default:
 		panic("unexpected state")
 	}
-}
-
-// MutexMap is a structure that allows having a map of self-evicting mutexes
-// by key. You do not need to worry about managing the contents of the map,
-// only requesting RLock/Lock for keys, and ensuring to call the returned
-// unlock functions.
-type MutexMap struct {
-	mus   map[string]RWMutex
-	mapMu sync.Mutex
-	pool  sync.Pool
-	queue []func()
-	evict []func()
-	count int32
-	maxmu int32
-	state uint8
-}
-
-// NewMap returns a new MutexMap instance with provided max no. open mutexes.
-func NewMap(max int32) MutexMap {
-	if max < 1 {
-		// Default = 128 * GOMAXPROCS
-		procs := runtime.GOMAXPROCS(0)
-		max = int32(procs * 128)
-	}
-	return MutexMap{
-		mus: make(map[string]RWMutex),
-		pool: sync.Pool{
-			New: func() interface{} {
-				return NewRW()
-			},
-		},
-		maxmu: max,
-	}
-}
-
-// acquire will either acquire a mutex from pool or alloc.
-func (mm *MutexMap) acquire() RWMutex {
-	return mm.pool.Get().(RWMutex)
-}
-
-// release will release provided mutex to pool.
-func (mm *MutexMap) release(mu RWMutex) {
-	mm.pool.Put(mu)
-}
-
-// spinLock will wait (using a mutex to sleep thread) until 'cond()' returns true,
-// returning with map lock. Note that 'cond' is performed within a map lock.
-func (mm *MutexMap) spinLock(cond func() bool) {
-	mu := mm.acquire()
-	defer mm.release(mu)
-
-	for {
-		// Get map lock
-		mm.mapMu.Lock()
-
-		// Check if return
-		if cond() {
-			return
-		}
-
-		// Queue ourselves
-		unlock := mu.Lock()
-		mm.queue = append(mm.queue, unlock)
-		mm.mapMu.Unlock()
-
-		// Wait on notify
-		mu.Lock()()
-	}
-}
-
-// lockMutex will acquire a lock on the mutex at provided key, handling earlier allocated mutex if provided. Unlocks map on return.
-func (mm *MutexMap) lockMutex(key string, lt locktype) func() {
-	var unlock func()
-
-	// Incr counter
-	mm.count++
-
-	// Check for existing mutex at key
-	mu, ok := mm.mus[key]
-	if !ok {
-		// Alloc from pool
-		mu = mm.acquire()
-		mm.mus[key] = mu
-
-		// Queue mutex for eviction
-		mm.evict = append(mm.evict, func() {
-			delete(mm.mus, key)
-			mm.pool.Put(mu)
-		})
-	}
-
-	// If no state, set in use.
-	// State will already have been
-	// set if this is from LockState{}
-	if mm.state == stateUnlockd {
-		mm.state = stateInUse
-	}
 
 	switch {
-	// Read lock
+	// If unlocked and not a map
+	// lock request, set in use
+	case lt&lockTypeMap == 0:
+		if state == stateUnlockd {
+			state = stateInUse
+		}
+
+	// Set read lock state
 	case lt&lockTypeRead != 0:
-		unlock = mu.RLock()
+		state = stateRLocked
 
-	// Write lock
+	// Set write lock state
 	case lt&lockTypeWrite != 0:
-		unlock = mu.Lock()
+		state = stateLocked
 
-	// shouldn't reach here
 	default:
 		panic("unexpected lock type")
 	}
 
-	// Unlock map + return
-	mm.mapMu.Unlock()
-	return func() {
-		mm.mapMu.Lock()
-		unlock()
-		go mm.onUnlock()
+	return state, true
+}
+
+// MutexMap is a structure that allows read / write locking key, performing
+// as you'd expect a map[string]*sync.RWMutex to perform. The differences
+// being that the entire map can itself be read / write locked, it uses memory
+// pooling for the mutex (not quite) structures, and it is self-evicting. The
+// core configurations of maximum no. open locks and wake modulus* are user
+// definable.
+//
+// * The wake modulus is the number that the current number of open locks is
+// modulused against to determine how often to notify sleeping goroutines.
+// These are goroutines that are attempting to lock a key / whole map and are
+// awaiting a permissible state (.e.g no key write locks allowed when the
+// map is read locked).
+type MutexMap struct {
+	qpool pool
+	queue []*sync.Mutex
+
+	mumap map[string]*rwmutex
+	mpool pool
+	evict []*rwmutex
+
+	count int32
+	maxmu int32
+	wake  int32
+
+	mapmu sync.Mutex
+	state uint8
+}
+
+// NewMap returns a new MutexMap instance with provided max no. open mutexes.
+func NewMap(max, wake int32) MutexMap {
+	// Determine wake mod.
+	if wake < 1 {
+		wake = defaultWake
+	}
+
+	// Determine max no. mutexes
+	if max < 1 {
+		procs := runtime.GOMAXPROCS(0)
+		max = wake * int32(procs)
+	}
+
+	return MutexMap{
+		qpool: pool{
+			alloc: func() interface{} {
+				return &sync.Mutex{}
+			},
+		},
+		mumap: make(map[string]*rwmutex, max),
+		mpool: pool{
+			alloc: func() interface{} {
+				return &rwmutex{}
+			},
+		},
+		maxmu: max,
+		wake:  wake,
 	}
 }
 
-// onUnlock is performed as the final (async) stage of releasing an acquired key / map mutex.
-func (mm *MutexMap) onUnlock() {
-	// Decr counter
-	mm.count--
+// MAX sets the MutexMap max open locks and wake modulus, returns current values.
+// For values less than zero defaults are set, and zero is non-op.
+func (mm *MutexMap) SET(max, wake int32) (int32, int32) {
+	mm.mapmu.Lock()
 
-	if mm.count < 1 {
-		// Perform all queued evictions
-		for i := 0; i < len(mm.evict); i++ {
-			mm.evict[i]()
-		}
+	switch {
+	// Set default wake
+	case wake < 0:
+		mm.wake = defaultWake
 
-		// Notify all waiting goroutines
-		for i := 0; i < len(mm.queue); i++ {
-			mm.queue[i]()
-		}
-
-		// Reset the map state
-		mm.evict = nil
-		mm.queue = nil
-		mm.state = stateUnlockd
+	// Set supplied wake
+	case wake > 0:
+		mm.wake = wake
 	}
 
-	// Finally, unlock
-	mm.mapMu.Unlock()
+	switch {
+	// Set default max
+	case max < 0:
+		procs := runtime.GOMAXPROCS(0)
+		mm.maxmu = wake * int32(procs)
+
+	// Set supplied max
+	case max > 0:
+		mm.maxmu = max
+	}
+
+	// Fetch values
+	max = mm.maxmu
+	wake = mm.wake
+
+	mm.mapmu.Unlock()
+	return max, wake
+}
+
+// spinLock will wait (using a mutex to sleep thread) until conditional returns true.
+func (mm *MutexMap) spinLock(cond func() bool) {
+	var mu *sync.Mutex
+
+	for {
+		// Acquire map lock
+		mm.mapmu.Lock()
+
+		if cond() {
+			// Release mu if needed
+			if mu != nil {
+				mm.qpool.Release(mu)
+			}
+			return
+		}
+
+		// Alloc mu if needed
+		if mu == nil {
+			v := mm.qpool.Acquire()
+			mu = v.(*sync.Mutex)
+		}
+
+		// Queue ourselves
+		mm.queue = append(mm.queue, mu)
+		mu.Lock()
+
+		// Unlock map
+		mm.mapmu.Unlock()
+
+		// Wait on notify
+		mu.Lock()
+		mu.Unlock()
+	}
+}
+
+// lock will acquire a lock of given type on the 'mutex' at key.
+func (mm *MutexMap) lock(key string, lt uint8) func() {
+	var ok bool
+	var mu *rwmutex
+
+	// Spin lock until returns true
+	mm.spinLock(func() bool {
+		// Check not overloaded
+		if !(mm.count < mm.maxmu) {
+			return false
+		}
+
+		// Attempt to acquire usable map state
+		state, ok := acquireState(mm.state, lt)
+		if !ok {
+			return false
+		}
+
+		// Update state
+		mm.state = state
+
+		// Ensure mutex at key
+		// is in lockable state
+		mu, ok = mm.mumap[key]
+		return !ok || mu.CanLock(lt)
+	})
+
+	// Incr count
+	mm.count++
+
+	if !ok {
+		// No mutex found for key
+
+		// Alloc from pool
+		v := mm.mpool.Acquire()
+		mu = v.(*rwmutex)
+		mm.mumap[key] = mu
+
+		// Set our key
+		mu.key = key
+
+		// Queue for eviction
+		mm.evict = append(mm.evict, mu)
+	}
+
+	// Lock mutex
+	mu.Lock(lt)
+
+	// Unlock map
+	mm.mapmu.Unlock()
+
+	return func() {
+		mm.mapmu.Lock()
+		mu.Unlock()
+		go mm.cleanup()
+	}
+}
+
+// lockMap will lock the whole map under given lock type.
+func (mm *MutexMap) lockMap(lt uint8) {
+	// Spin lock until returns true
+	mm.spinLock(func() bool {
+		// Attempt to acquire usable map state
+		state, ok := acquireState(mm.state, lt)
+		if !ok {
+			return false
+		}
+
+		// Update state
+		mm.state = state
+
+		return true
+	})
+
+	// Incr count
+	mm.count++
+
+	// State acquired, unlock
+	mm.mapmu.Unlock()
+}
+
+// cleanup is performed as the final stage of unlocking a locked key / map state, finally unlocks map.
+func (mm *MutexMap) cleanup() {
+	// Decr count
+	mm.count--
+
+	if mm.count%mm.wake == 0 {
+		// Notify queued routines
+		for _, mu := range mm.queue {
+			mu.Unlock()
+		}
+
+		// Reset queue
+		mm.queue = mm.queue[:0]
+	}
+
+	if mm.count < 1 {
+		// Perform evictions
+		for _, mu := range mm.evict {
+			key := mu.key
+			mu.key = ""
+			delete(mm.mumap, key)
+			mm.mpool.Release(mu)
+		}
+
+		// Reset map state
+		mm.evict = mm.evict[:0]
+		mm.state = stateUnlockd
+		mm.mpool.GC()
+		mm.qpool.GC()
+	}
+
+	// Unlock map
+	mm.mapmu.Unlock()
 }
 
 // RLockMap acquires a read lock over the entire map, returning a lock state for acquiring key read locks.
 // Please note that the 'unlock()' function will block until all keys locked from this state are unlocked.
 func (mm *MutexMap) RLockMap() *LockState {
-	return mm.getMapLock(lockTypeRead)
+	mm.lockMap(lockTypeRead | lockTypeMap)
+	return &LockState{
+		mmap: mm,
+		ltyp: lockTypeRead,
+	}
 }
 
 // LockMap acquires a write lock over the entire map, returning a lock state for acquiring key read/write locks.
 // Please note that the 'unlock()' function will block until all keys locked from this state are unlocked.
 func (mm *MutexMap) LockMap() *LockState {
-	return mm.getMapLock(lockTypeWrite)
+	mm.lockMap(lockTypeWrite | lockTypeMap)
+	return &LockState{
+		mmap: mm,
+		ltyp: lockTypeWrite,
+	}
 }
 
 // RLock acquires a mutex read lock for supplied key, returning an RUnlock function.
 func (mm *MutexMap) RLock(key string) (runlock func()) {
-	return mm.getLock(key, lockTypeRead)
+	return mm.lock(key, lockTypeRead)
 }
 
 // Lock acquires a mutex write lock for supplied key, returning an Unlock function.
 func (mm *MutexMap) Lock(key string) (unlock func()) {
-	return mm.getLock(key, lockTypeWrite)
-}
-
-// getLock will fetch lock of provided type, for given key, returning unlock function.
-func (mm *MutexMap) getLock(key string, lt locktype) func() {
-	// Spin until achieve lock
-	mm.spinLock(func() bool {
-		return permitLockType(mm.state, lt) &&
-			mm.count < mm.maxmu // not overloaded
-	})
-
-	// Perform actual mutex lock
-	return mm.lockMutex(key, lt)
-}
-
-// getMapLock will acquire a map lock of provided type, returning a LockState session.
-func (mm *MutexMap) getMapLock(lt locktype) *LockState {
-	// Spin until achieve lock
-	mm.spinLock(func() bool {
-		return permitLockType(mm.state, lt|lockTypeMap) &&
-			mm.count < mm.maxmu // not overloaded
-	})
-
-	// Incr counter
-	mm.count++
-
-	switch {
-	// Set read lock state
-	case lt&lockTypeRead != 0:
-		mm.state = stateRLocked
-
-	// Set write lock state
-	case lt&lockTypeWrite != 0:
-		mm.state = stateLocked
-
-	default:
-		panic("unexpected lock type")
-	}
-
-	// Unlock + return
-	mm.mapMu.Unlock()
-	return &LockState{
-		mmap: mm,
-		ltyp: lt,
-	}
+	return mm.lock(key, lockTypeWrite)
 }
 
 // LockState represents a window to a locked MutexMap.
@@ -267,56 +354,113 @@ type LockState struct {
 	wait sync.WaitGroup
 	mmap *MutexMap
 	done uint32
-	ltyp locktype
+	ltyp uint8
 }
 
 // Lock: see MutexMap.Lock() definition. Will panic if map only read locked.
 func (st *LockState) Lock(key string) (unlock func()) {
-	return st.getLock(key, lockTypeWrite)
+	return st.lock(key, lockTypeWrite)
 }
 
 // RLock: see MutexMap.RLock() definition.
 func (st *LockState) RLock(key string) (runlock func()) {
-	return st.getLock(key, lockTypeRead)
+	return st.lock(key, lockTypeRead)
+}
+
+// lock: see MutexMap.lock() definition.
+func (st *LockState) lock(key string, lt uint8) func() {
+	st.wait.Add(1) // track lock
+
+	if atomic.LoadUint32(&st.done) == 1 {
+		panic("called (r)lock on unlocked state")
+	} else if lt&lockTypeWrite != 0 &&
+		st.ltyp&lockTypeWrite == 0 {
+		panic("called lock on rlocked map")
+	}
+
+	var ok bool
+	var mu *rwmutex
+
+	// Spin lock until returns true
+	st.mmap.spinLock(func() bool {
+		// Check not overloaded
+		if !(st.mmap.count < st.mmap.maxmu) {
+			return false
+		}
+
+		// Ensure mutex at key
+		// is in lockable state
+		mu, ok = st.mmap.mumap[key]
+		return !ok || mu.CanLock(lt)
+	})
+
+	// Incr count
+	st.mmap.count++
+
+	if !ok {
+		// No mutex found for key
+
+		// Alloc from pool
+		v := st.mmap.mpool.Acquire()
+		mu = v.(*rwmutex)
+		st.mmap.mumap[key] = mu
+
+		// Set our key
+		mu.key = key
+
+		// Queue for eviction
+		st.mmap.evict = append(st.mmap.evict, mu)
+	}
+
+	// Lock mutex
+	mu.Lock(lt)
+
+	// Unlock map
+	st.mmap.mapmu.Unlock()
+
+	return func() {
+		st.mmap.mapmu.Lock()
+		mu.Unlock()
+		go st.mmap.cleanup()
+		st.wait.Add(-1)
+	}
 }
 
 // UnlockMap will close this state and release the currently locked map.
 func (st *LockState) UnlockMap() {
-	// Set state to finished (or panic if already done)
 	if !atomic.CompareAndSwapUint32(&st.done, 0, 1) {
-		panic("called UnlockMap() on expired state")
+		panic("called unlockmap on expired state")
 	}
-
-	// Wait until done
 	st.wait.Wait()
-
-	// Async reset map
-	st.mmap.mapMu.Lock()
-	go st.mmap.onUnlock()
+	st.mmap.mapmu.Lock()
+	go st.mmap.cleanup()
 }
 
-// getLock: see MutexMap.getLock() definition.
-func (st *LockState) getLock(key string, lt locktype) func() {
-	st.wait.Add(1) // track lock
+// rwmutex is a very simple *representation* of a read-write
+// mutex, though not one in implementation. it works by
+// tracking the lock state for a given map key, which is
+// protected by the map's mutex.
+type rwmutex struct {
+	rcnt uint32
+	lock uint8
+	key  string
+}
 
-	// Check if closed, or if write lock is allowed
-	if atomic.LoadUint32(&st.done) == 1 {
-		panic("map lock closed")
-	} else if lt&lockTypeWrite != 0 &&
-		st.ltyp&lockTypeWrite == 0 {
-		panic("called .Lock() on rlocked map")
-	}
+func (mu *rwmutex) CanLock(lt uint8) bool {
+	return mu.lock == 0 ||
+		(mu.lock&lockTypeRead != 0 && lt&lockTypeRead != 0)
+}
 
-	// Spin until achieve map lock
-	st.mmap.spinLock(func() bool {
-		return st.mmap.count < st.mmap.maxmu
-	}) // i.e. not overloaded
-
-	// Perform actual mutex lock
-	unlock := st.mmap.lockMutex(key, lt)
-
-	return func() {
-		unlock()
-		st.wait.Done()
+func (mu *rwmutex) Lock(lt uint8) {
+	mu.lock = lt
+	if lt&lockTypeRead != 0 {
+		mu.rcnt++
+	}
+}
+
+func (mu *rwmutex) Unlock() {
+	mu.rcnt--
+	if mu.rcnt == 0 {
+		mu.lock = 0
 	}
 }

vendor/codeberg.org/gruf/go-mutexes/mutex.go

@@ -41,24 +41,24 @@ func WithFuncRW(mu RWMutex, onLock, onRLock, onUnlock, onRUnlock func()) RWMutex
 }
 
 // baseMutex simply wraps a sync.Mutex to implement our Mutex interface
-type baseMutex struct{ mu sync.Mutex }
+type baseMutex sync.Mutex
 
 func (mu *baseMutex) Lock() func() {
-	mu.mu.Lock()
-	return mu.mu.Unlock
+	(*sync.Mutex)(mu).Lock()
+	return (*sync.Mutex)(mu).Unlock
 }
 
 // baseRWMutex simply wraps a sync.RWMutex to implement our RWMutex interface
-type baseRWMutex struct{ mu sync.RWMutex }
+type baseRWMutex sync.RWMutex
 
 func (mu *baseRWMutex) Lock() func() {
-	mu.mu.Lock()
-	return mu.mu.Unlock
+	(*sync.RWMutex)(mu).Lock()
+	return (*sync.RWMutex)(mu).Unlock
 }
 
 func (mu *baseRWMutex) RLock() func() {
-	mu.mu.RLock()
-	return mu.mu.RUnlock
+	(*sync.RWMutex)(mu).RLock()
+	return (*sync.RWMutex)(mu).RUnlock
 }
 
 // fnMutex wraps a Mutex to add hooks for Lock and Unlock

vendor/codeberg.org/gruf/go-mutexes/mutex_safe.go

@@ -1,6 +1,8 @@
 package mutexes
 
-import "sync"
+import (
+	"sync/atomic"
+)
 
 // WithSafety wrapps the supplied Mutex to protect unlock fns
 // from being called multiple times
@@ -19,8 +21,7 @@ type safeMutex struct{
 
 func (mu *safeMutex) Lock() func() {
 	unlock := mu.mu.Lock()
-	once := sync.Once{}
-	return func() { once.Do(unlock) }
+	return once(unlock)
 }
 
 // safeRWMutex simply wraps a RWMutex to add multi-unlock safety
@@ -28,12 +29,22 @@ type safeRWMutex struct{
 
 func (mu *safeRWMutex) Lock() func() {
 	unlock := mu.mu.Lock()
-	once := sync.Once{}
-	return func() { once.Do(unlock) }
+	return once(unlock)
 }
 
 func (mu *safeRWMutex) RLock() func() {
 	unlock := mu.mu.RLock()
-	once := sync.Once{}
-	return func() { once.Do(unlock) }
+	return once(unlock)
+}
+
+// once will perform 'do' only once, this is safe for unlocks
+// as 2 functions calling 'unlock()' don't need absolute guarantees
+// that by the time it is completed the unlock was finished.
+func once(do func()) func() {
+	var done uint32
+	return func() {
+		if atomic.CompareAndSwapUint32(&done, 0, 1) {
+			do()
+		}
+	}
 }

vendor/codeberg.org/gruf/go-mutexes/mutex_timeout.go

@@ -97,7 +97,9 @@ func mutexTimeout(d time.Duration, unlock func(), fn func()) func() {
 // timerPool is the global &timer{} pool.
 var timerPool = sync.Pool{
 	New: func() interface{} {
-		return newtimer()
+		t := time.NewTimer(time.Minute)
+		t.Stop()
+		return &timer{t: t, c: make(chan struct{})}
 	},
 }
 
@@ -107,13 +109,6 @@ type timer struct {
 	c chan struct{}
 }
 
-// newtimer returns a new timer instance.
-func newtimer() *timer {
-	t := time.NewTimer(time.Minute)
-	t.Stop()
-	return &timer{t: t, c: make(chan struct{})}
-}
-
 // Start will start the timer with duration 'd', performing 'fn' on timeout.
 func (t *timer) Start(d time.Duration, fn func()) {
 	t.t.Reset(d)

vendor/codeberg.org/gruf/go-mutexes/pool.go

@@ -0,0 +1,40 @@
+package mutexes
+
+// pool is a very simply memory pool.
+type pool struct {
+	current []interface{}
+	victim  []interface{}
+	alloc   func() interface{}
+}
+
+// Acquire will returns a sync.RWMutex from pool (or alloc new).
+func (p *pool) Acquire() interface{} {
+	// First try the current queue
+	if l := len(p.current) - 1; l >= 0 {
+		v := p.current[l]
+		p.current = p.current[:l]
+		return v
+	}
+
+	// Next try the victim queue.
+	if l := len(p.victim) - 1; l >= 0 {
+		v := p.victim[l]
+		p.victim = p.victim[:l]
+		return v
+	}
+
+	// Lastly, alloc new.
+	return p.alloc()
+}
+
+// Release places a sync.RWMutex back in the pool.
+func (p *pool) Release(v interface{}) {
+	p.current = append(p.current, v)
+}
+
+// GC will clear out unused entries from the pool.
+func (p *pool) GC() {
+	current := p.current
+	p.current = nil
+	p.victim = current
+}

vendor/codeberg.org/gruf/go-store/kv/store.go

@@ -45,7 +45,7 @@ func OpenStorage(storage storage.Storage) (*KVStore, error) {
 
 	// Return new KVStore
 	return &KVStore{
-		mutex:   mutexes.NewMap(-1),
+		mutex:   mutexes.NewMap(-1, -1),
 		storage: storage,
 	}, nil
 }

vendor/codeberg.org/gruf/go-store/storage/disk.go

@@ -10,7 +10,7 @@
 	"syscall"
 
 	"codeberg.org/gruf/go-bytes"
-	"codeberg.org/gruf/go-pools"
+	"codeberg.org/gruf/go-fastcopy"
 	"codeberg.org/gruf/go-store/util"
 )
 
@@ -81,10 +81,10 @@ func getDiskConfig(cfg *DiskConfig) DiskConfig {
 
 // DiskStorage is a Storage implementation that stores directly to a filesystem
 type DiskStorage struct {
-	path   string           // path is the root path of this store
-	bufp   pools.BufferPool // bufp is the buffer pool for this DiskStorage
-	config DiskConfig       // cfg is the supplied configuration for this store
-	lock   *Lock            // lock is the opened lockfile for this storage instance
+	path   string            // path is the root path of this store
+	cppool fastcopy.CopyPool // cppool is the prepared io copier with buffer pool
+	config DiskConfig        // cfg is the supplied configuration for this store
+	lock   *Lock             // lock is the opened lockfile for this storage instance
 }
 
 // OpenFile opens a DiskStorage instance for given folder path and configuration
@@ -147,13 +147,17 @@ func OpenFile(path string, cfg *DiskConfig) (*DiskStorage, error) {
 		return nil, err
 	}
 
-	// Return new DiskStorage
-	return &DiskStorage{
+	// Prepare DiskStorage
+	st := &DiskStorage{
 		path:   storePath,
-		bufp:   pools.NewBufferPool(config.WriteBufSize),
 		config: config,
 		lock:   lock,
-	}, nil
+	}
+
+	// Set copypool buffer size
+	st.cppool.Buffer(config.WriteBufSize)
+
+	return st, nil
 }
 
 // Clean implements Storage.Clean()
@@ -271,13 +275,8 @@ func (st *DiskStorage) WriteStream(key string, r io.Reader) error {
 	}
 	defer cFile.Close()
 
-	// Acquire write buffer
-	buf := st.bufp.Get()
-	defer st.bufp.Put(buf)
-	buf.Grow(st.config.WriteBufSize)
-
-	// Copy reader to file
-	_, err = io.CopyBuffer(cFile, r, buf.B)
+	// Copy provided reader to file
+	_, err = st.cppool.Copy(cFile, r)
 	return err
 }
 

vendor/modules.txt

@@ -4,6 +4,9 @@ codeberg.org/gruf/go-bytes
 # codeberg.org/gruf/go-errors v1.0.5
 ## explicit; go 1.15
 codeberg.org/gruf/go-errors
+# codeberg.org/gruf/go-fastcopy v1.1.1
+## explicit; go 1.17
+codeberg.org/gruf/go-fastcopy
 # codeberg.org/gruf/go-fastpath v1.0.2
 ## explicit; go 1.14
 codeberg.org/gruf/go-fastpath
@@ -13,7 +16,7 @@ codeberg.org/gruf/go-format
 # codeberg.org/gruf/go-hashenc v1.0.1
 ## explicit; go 1.16
 codeberg.org/gruf/go-hashenc
-# codeberg.org/gruf/go-mutexes v1.1.0
+# codeberg.org/gruf/go-mutexes v1.1.2
 ## explicit; go 1.14
 codeberg.org/gruf/go-mutexes
 # codeberg.org/gruf/go-pools v1.0.2
@@ -22,7 +25,7 @@ codeberg.org/gruf/go-pools
 # codeberg.org/gruf/go-runners v1.2.0
 ## explicit; go 1.14
 codeberg.org/gruf/go-runners
-# codeberg.org/gruf/go-store v1.3.3
+# codeberg.org/gruf/go-store v1.3.6
 ## explicit; go 1.14
 codeberg.org/gruf/go-store/kv
 codeberg.org/gruf/go-store/storage