diff --git a/.chloggen/pkg-stanza-input-udp-async-fix-data-corruption.yaml b/.chloggen/pkg-stanza-input-udp-async-fix-data-corruption.yaml new file mode 100644 index 000000000000..0bcf8f4a5e1f --- /dev/null +++ b/.chloggen/pkg-stanza-input-udp-async-fix-data-corruption.yaml @@ -0,0 +1,27 @@ +# Use this changelog template to create an entry for release notes. + +# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix' +change_type: bug_fix + +# The name of the component, or a single word describing the area of concern, (e.g. filelogreceiver) +component: pkg/stanza + +# A brief description of the change. Surround your text with quotes ("") if it needs to start with a backtick (`). +note: Fix data-corruption/race-condition issue in udp async (reuse of buffer); use buffer pool isntead. + +# Mandatory: One or more tracking issues related to the change. You can use the PR number here if no issue exists. +issues: [27613] + +# (Optional) One or more lines of additional information to render under the primary note. +# These lines will be padded with 2 spaces and then inserted directly into the document. +# Use pipe (|) for multiline entries. +subtext: + +# If your change doesn't affect end users or the exported elements of any package, +# you should instead start your pull request title with [chore] or use the "Skip Changelog" label. +# Optional: The change log or logs in which this entry should be included. +# e.g. '[user]' or '[user, api]' +# Include 'user' if the change is relevant to end users. +# Include 'api' if there is a change to a library API. +# Default: '[user]' +change_logs: [] diff --git a/pkg/stanza/operator/input/udp/udp.go b/pkg/stanza/operator/input/udp/udp.go index 4593f9ead5e5..ac7126f0eec8 100644 --- a/pkg/stanza/operator/input/udp/udp.go +++ b/pkg/stanza/operator/input/udp/udp.go @@ -138,6 +138,12 @@ func (c Config) Build(logger *zap.SugaredLogger) (operator.Operator, error) { if c.AsyncConfig != nil { udpInput.messageQueue = make(chan messageAndAddress, c.AsyncConfig.MaxQueueLength) + udpInput.readBufferPool = sync.Pool{ + New: func() interface{} { + buffer := make([]byte, MaxUDPSize) + return &buffer + }, + } } return udpInput, nil } @@ -159,13 +165,15 @@ type Input struct { splitFunc bufio.SplitFunc resolver *helper.IPResolver - messageQueue chan messageAndAddress - stopOnce sync.Once + messageQueue chan messageAndAddress + readBufferPool sync.Pool + stopOnce sync.Once } type messageAndAddress struct { - Message []byte - RemoteAddr net.Addr + Message *[]byte + RemoteAddr net.Addr + MessageLength int } // Start will start listening for messages on a socket. @@ -206,9 +214,12 @@ func (u *Input) readAndProcessMessages(ctx context.Context) { defer u.wg.Done() dec := decode.New(u.encoding) - buf := make([]byte, 0, MaxUDPSize) + readBuffer := make([]byte, MaxUDPSize) + scannerBuffer := make([]byte, 0, MaxUDPSize) for { - message, remoteAddr, err := u.readMessage() + message, remoteAddr, bufferLength, err := u.readMessage(readBuffer) + message = u.removeTrailingCharactersAndNULsFromBuffer(message, bufferLength) + if err != nil { select { case <-ctx.Done(): @@ -219,11 +230,11 @@ func (u *Input) readAndProcessMessages(ctx context.Context) { break } - u.processMessage(ctx, message, remoteAddr, dec, buf) + u.processMessage(ctx, message, remoteAddr, dec, scannerBuffer) } } -func (u *Input) processMessage(ctx context.Context, message []byte, remoteAddr net.Addr, dec *decode.Decoder, buf []byte) { +func (u *Input) processMessage(ctx context.Context, message []byte, remoteAddr net.Addr, dec *decode.Decoder, scannerBuffer []byte) { if u.OneLogPerPacket { log := truncateMaxLog(message) u.handleMessage(ctx, remoteAddr, dec, log) @@ -231,7 +242,7 @@ func (u *Input) processMessage(ctx context.Context, message []byte, remoteAddr n } scanner := bufio.NewScanner(bytes.NewReader(message)) - scanner.Buffer(buf, MaxUDPSize) + scanner.Buffer(scannerBuffer, MaxUDPSize) scanner.Split(u.splitFunc) @@ -247,8 +258,10 @@ func (u *Input) readMessagesAsync(ctx context.Context) { defer u.wg.Done() for { - message, remoteAddr, err := u.readMessage() + readBuffer := u.readBufferPool.Get().(*[]byte) // Can't reuse the same buffer since same references would be written multiple times to the messageQueue (and cause data override of previous entries) + message, remoteAddr, bufferLength, err := u.readMessage(*readBuffer) if err != nil { + u.readBufferPool.Put(readBuffer) select { case <-ctx.Done(): return @@ -259,8 +272,9 @@ func (u *Input) readMessagesAsync(ctx context.Context) { } messageAndAddr := messageAndAddress{ - Message: message, - RemoteAddr: remoteAddr, + Message: &message, + MessageLength: bufferLength, + RemoteAddr: remoteAddr, } // Send the message to the message queue for processing @@ -272,7 +286,7 @@ func (u *Input) processMessagesAsync(ctx context.Context) { defer u.wg.Done() dec := decode.New(u.encoding) - buf := make([]byte, 0, MaxUDPSize) + scannerBuffer := make([]byte, 0, MaxUDPSize) for { // Read a message from the message queue. @@ -281,7 +295,9 @@ func (u *Input) processMessagesAsync(ctx context.Context) { return // Channel closed, exit the goroutine. } - u.processMessage(ctx, messageAndAddr.Message, messageAndAddr.RemoteAddr, dec, buf) + trimmedMessage := u.removeTrailingCharactersAndNULsFromBuffer(*messageAndAddr.Message, messageAndAddr.MessageLength) + u.processMessage(ctx, trimmedMessage, messageAndAddr.RemoteAddr, dec, scannerBuffer) + u.readBufferPool.Put(messageAndAddr.Message) } } @@ -331,17 +347,22 @@ func (u *Input) handleMessage(ctx context.Context, remoteAddr net.Addr, dec *dec } // readMessage will read log messages from the connection. -func (u *Input) readMessage() ([]byte, net.Addr, error) { - n, addr, err := u.connection.ReadFrom(u.buffer) +func (u *Input) readMessage(buffer []byte) ([]byte, net.Addr, int, error) { + n, addr, err := u.connection.ReadFrom(buffer) if err != nil { - return nil, nil, err + return nil, nil, 0, err } + return buffer, addr, n, nil +} + +// This will remove trailing characters and NULs from the buffer +func (u *Input) removeTrailingCharactersAndNULsFromBuffer(buffer []byte, n int) []byte { // Remove trailing characters and NULs - for ; (n > 0) && (u.buffer[n-1] < 32); n-- { // nolint + for ; (n > 0) && (buffer[n-1] < 32); n-- { // nolint } - return u.buffer[:n], addr, nil + return buffer[:n] } // Stop will stop listening for udp messages.