Simplify parallelReader.Read() (#7109)

Simplify parallelReader.Read() which also fixes previous 
implementation where it was returning before all the parallel 
reading go-routines had terminated which caused race conditions.

cmd/erasure-decode.go

@@ -19,6 +19,7 @@ package cmd
 import (
 	"context"
 	"io"
+	"sync"
 
 	"github.com/minio/minio/cmd/logger"
 )
@@ -58,74 +59,74 @@ func (p *parallelReader) canDecode(buf [][]byte) bool {
 
 // Read reads from readers in parallel. Returns p.dataBlocks number of bufs.
 func (p *parallelReader) Read() ([][]byte, error) {
-	type errIdx struct {
-		idx int
-		buf []byte
-		err error
-	}
-
-	errCh := make(chan errIdx)
-	currReaderIndex := 0
 	newBuf := make([][]byte, len(p.readers))
+	var newBufLK sync.RWMutex
 
 	if p.offset+p.shardSize > p.shardFileSize {
 		p.shardSize = p.shardFileSize - p.offset
 	}
 
-	read := func(currReaderIndex int) {
-		if p.buf[currReaderIndex] == nil {
-			p.buf[currReaderIndex] = make([]byte, p.shardSize)
-		}
-		p.buf[currReaderIndex] = p.buf[currReaderIndex][:p.shardSize]
-		_, err := p.readers[currReaderIndex].ReadAt(p.buf[currReaderIndex], p.offset)
-		errCh <- errIdx{currReaderIndex, p.buf[currReaderIndex], err}
+	readTriggerCh := make(chan bool, len(p.readers))
+	for i := 0; i < p.dataBlocks; i++ {
+		// Setup read triggers for p.dataBlocks number of reads so that it reads in parallel.
+		readTriggerCh <- true
 	}
 
-	readerCount := 0
-	for _, r := range p.readers {
-		if r != nil {
-			readerCount++
-		}
-	}
-	if readerCount < p.dataBlocks {
-		return nil, errXLReadQuorum
-	}
-
-	readerCount = 0
-	for i, r := range p.readers {
-		if r == nil {
-			continue
+	readerIndex := 0
+	var wg sync.WaitGroup
+	// if readTrigger is true, it implies next disk.ReadAt() should be tried
+	// if readTrigger is false, it implies previous disk.ReadAt() was successful and there is no need
+	// to try reading the next disk.
+	for readTrigger := range readTriggerCh {
+		newBufLK.RLock()
+		canDecode := p.canDecode(newBuf)
+		newBufLK.RUnlock()
+		if canDecode {
+			break
 		}
-		go read(i)
-		readerCount++
-		if readerCount == p.dataBlocks {
-			currReaderIndex = i + 1
+		if readerIndex == len(p.readers) {
 			break
 		}
-	}
-
-	for errVal := range errCh {
-		if errVal.err == nil {
-			newBuf[errVal.idx] = errVal.buf
-			if p.canDecode(newBuf) {
-				p.offset += int64(p.shardSize)
-				return newBuf, nil
-			}
+		if !readTrigger {
 			continue
 		}
-		p.readers[errVal.idx] = nil
-		for currReaderIndex < len(p.readers) {
-			if p.readers[currReaderIndex] != nil {
-				break
+		wg.Add(1)
+		go func(i int) {
+			defer wg.Done()
+			disk := p.readers[i]
+			if disk == nil {
+				// Since disk is nil, trigger another read.
+				readTriggerCh <- true
+				return
 			}
-			currReaderIndex++
-		}
+			if p.buf[i] == nil {
+				// Reading first time on this disk, hence the buffer needs to be allocated.
+				// Subsequent reads will re-use this buffer.
+				p.buf[i] = make([]byte, p.shardSize)
+			}
+			// For the last shard, the shardsize might be less than previous shard sizes.
+			// Hence the following statement ensures that the buffer size is reset to the right size.
+			p.buf[i] = p.buf[i][:p.shardSize]
+			_, err := disk.ReadAt(p.buf[i], p.offset)
+			if err != nil {
+				p.readers[i] = nil
+				// Since ReadAt returned error, trigger another read.
+				readTriggerCh <- true
+				return
+			}
+			newBufLK.Lock()
+			newBuf[i] = p.buf[i]
+			newBufLK.Unlock()
+			// Since ReadAt returned success, there is no need to trigger another read.
+			readTriggerCh <- false
+		}(readerIndex)
+		readerIndex++
+	}
+	wg.Wait()
 
-		if currReaderIndex == len(p.readers) {
-			break
-		}
-		go read(currReaderIndex)
-		currReaderIndex++
+	if p.canDecode(newBuf) {
+		p.offset += p.shardSize
+		return newBuf, nil
 	}
 
 	return nil, errXLReadQuorum