Sync from upstream (2024-12-03) (#5)

* Fix unwinding at syscall on aarch64 (open-telemetry#218)

* Add PID as an attribute in each sample (open-telemetry#212)

* ebpf: increase number of stack delta buckets (open-telemetry#231)

Signed-off-by: Florian Lehner <florian.lehner@elastic.co>

* reporter: use htlhash attribute for profiling specific hash (open-telemetry#236)

Signed-off-by: Florian Lehner <florian.lehner@elastic.co>

* reporter: drop fifo (open-telemetry#239)

Signed-off-by: Florian Lehner <florian.lehner@elastic.co>

* Drop more unused code (open-telemetry#240)

* reporter: do not add empty attributes (open-telemetry#233)

Signed-off-by: Florian Lehner <florian.lehner@elastic.co>

* controller: fix reporter interval mix up with monitor interval (open-telemetry#242)

* Extract reporter runloop (open-telemetry#228)

Co-authored-by: Christos Kalkanis <christos.kalkanis@elastic.co>

* Extract lookupcgroupv2 out of the otlp reporter (open-telemetry#227)

* Add CPU id to trace and trace metadata (open-telemetry#249)

* reporter: don't expire actively used executables (open-telemetry#247)

* Remove legacy code from libpf.UnixTime64 (open-telemetry#252)

* Turn kernel module file parsing errors into warnings (open-telemetry#255)

Co-authored-by: Florian Lehner <florianl@users.noreply.github.com>

* readatbuf: add missing check when reading from tail chunk (open-telemetry#259)

* metrics: Don't send counters with 0 values (open-telemetry#246)

---------

Signed-off-by: Florian Lehner <florian.lehner@elastic.co>
Co-authored-by: umanwizard <brennan@umanwizard.com>
Co-authored-by: Bhavna Jindal <bhavna.jindal@yahoo.com>
Co-authored-by: Florian Lehner <florianl@users.noreply.github.com>
Co-authored-by: Tim Rühsen <tim.ruhsen@elastic.co>
Co-authored-by: Nayef Ghattas <nayef.ghattas@datadoghq.com>
Co-authored-by: Damien Mathieu <42@dmathieu.com>
Co-authored-by: Christos Kalkanis <christos.kalkanis@elastic.co>
Co-authored-by: Tommy Reilly <gnurizen@gmail.com>
Co-authored-by: Joel Höner <joel@elastic.co>

README.md

@@ -91,9 +91,9 @@ devfiler spins up a local server that listens on `0.0.0.0:11000`.
 
 To run it, simply download and unpack the archive from the following URL:
 
-https://upload.elastic.co/d/87e7697991940ec37f0c6e39ac38d213f65e8dc1ef9dbedff6aab9cba0adfaba
+https://upload.elastic.co/d/f8aa0c386baa808a616ca29f86b34c726edb5af36f8840a4cf28468ad534a4b5
 
-Authentication token: `c74dfc4db2212015`
+Authentication token: `2635c0750bf8ea69`
 
 
 The archive contains a build for each of the following platforms:

go.mod

@@ -3,12 +3,12 @@ module go.opentelemetry.io/ebpf-profiler
 go 1.22.2
 
 require (
+	github.com/aws/aws-sdk-go-v2 v1.30.5
 	github.com/aws/aws-sdk-go-v2/config v1.27.35
-	github.com/aws/aws-sdk-go-v2/feature/s3/manager v1.17.21
 	github.com/aws/aws-sdk-go-v2/service/s3 v1.62.0
 	github.com/cespare/xxhash/v2 v2.3.0
 	github.com/cilium/ebpf v0.16.0
-	github.com/elastic/go-freelru v0.15.0
+	github.com/elastic/go-freelru v0.16.0
 	github.com/elastic/go-perf v0.0.0-20241016160959-1342461adb4a
 	github.com/google/uuid v1.6.0
 	github.com/jsimonetti/rtnetlink v1.4.2
@@ -29,7 +29,6 @@ require (
 )
 
 require (
-	github.com/aws/aws-sdk-go-v2 v1.30.5 // indirect
 	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.4 // indirect
 	github.com/aws/aws-sdk-go-v2/credentials v1.17.33 // indirect
 	github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.16.13 // indirect

go.sum

@@ -8,8 +8,6 @@ github.com/aws/aws-sdk-go-v2/credentials v1.17.33 h1:lBHAQQznENv0gLHAZ73ONiTSkCt
 github.com/aws/aws-sdk-go-v2/credentials v1.17.33/go.mod h1:MBuqCUOT3ChfLuxNDGyra67eskx7ge9e3YKYBce7wpI=
 github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.16.13 h1:pfQ2sqNpMVK6xz2RbqLEL0GH87JOwSxPV2rzm8Zsb74=
 github.com/aws/aws-sdk-go-v2/feature/ec2/imds v1.16.13/go.mod h1:NG7RXPUlqfsCLLFfi0+IpKN4sCB9D9fw/qTaSB+xRoU=
-github.com/aws/aws-sdk-go-v2/feature/s3/manager v1.17.21 h1:sV0doPPsRT7gMP0BnDPwSsysVTV/nKpB/nFmMnz8goE=
-github.com/aws/aws-sdk-go-v2/feature/s3/manager v1.17.21/go.mod h1:ictvfJWqE2gkUFDRJVp5VU/TrytuzK88DYcpan7UYuA=
 github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.17 h1:pI7Bzt0BJtYA0N/JEC6B8fJ4RBrEMi1LBrkMdFYNSnQ=
 github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.17/go.mod h1:Dh5zzJYMtxfIjYW+/evjQ8uj2OyR/ve2KROHGHlSFqE=
 github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.17 h1:Mqr/V5gvrhA2gvgnF42Zh5iMiQNcOYthFYwCyrnuWlc=
@@ -43,10 +41,8 @@ github.com/cilium/ebpf v0.16.0/go.mod h1:L7u2Blt2jMM/vLAVgjxluxtBKlz3/GWjB0dMOEn
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/elastic/go-freelru v0.13.0 h1:TKKY6yCfNNNky7Pj9xZAOEpBcdNgZJfihEftOb55omg=
-github.com/elastic/go-freelru v0.13.0/go.mod h1:bSdWT4M0lW79K8QbX6XY2heQYSCqD7THoYf82pT/H3I=
-github.com/elastic/go-freelru v0.15.0 h1:Jo1aY8JAvpyxbTDJEudrsBfjFDaALpfVv8mxuh9sfvI=
-github.com/elastic/go-freelru v0.15.0/go.mod h1:bSdWT4M0lW79K8QbX6XY2heQYSCqD7THoYf82pT/H3I=
+github.com/elastic/go-freelru v0.16.0 h1:gG2HJ1WXN2tNl5/p40JS/l59HjvjRhjyAa+oFTRArYs=
+github.com/elastic/go-freelru v0.16.0/go.mod h1:bSdWT4M0lW79K8QbX6XY2heQYSCqD7THoYf82pT/H3I=
 github.com/elastic/go-perf v0.0.0-20241016160959-1342461adb4a h1:ymmtaN4bVCmKKeu4XEf6JEWNZKRXPMng1zjpKd+8rCU=
 github.com/elastic/go-perf v0.0.0-20241016160959-1342461adb4a/go.mod h1:Nt+pnRYvf0POC+7pXsrv8ubsEOSsaipJP0zlz1Ms1RM=
 github.com/go-quicktest/qt v1.101.0 h1:O1K29Txy5P2OK0dGo59b7b0LR6wKfIhttaAhHUyn7eI=

host/host.go

@@ -55,4 +55,5 @@ type Trace struct {
 	TID              libpf.PID
 	APMTraceID       libpf.APMTraceID
 	APMTransactionID libpf.APMTransactionID
+	CPU              int
 }

internal/controller/controller.go

@@ -56,8 +56,8 @@ func (c *Controller) Start(ctx context.Context) error {
 	traceHandlerCacheSize :=
 		traceCacheSize(c.config.MonitorInterval, c.config.SamplesPerSecond, uint16(presentCores))
 
-	intervals := times.New(c.config.MonitorInterval,
-		c.config.ReporterInterval, c.config.ProbabilisticInterval)
+	intervals := times.New(c.config.ReporterInterval, c.config.MonitorInterval,
+		c.config.ProbabilisticInterval)
 
 	// Start periodic synchronization with the realtime clock
 	times.StartRealtimeSync(ctx, c.config.ClockSyncInterval)

libpf/cgroupv2.go

@@ -0,0 +1,59 @@
+// Copyright The OpenTelemetry Authors
+// SPDX-License-Identifier: Apache-2.0
+
+package libpf // import "go.opentelemetry.io/ebpf-profiler/libpf"
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"regexp"
+
+	lru "github.com/elastic/go-freelru"
+	log "github.com/sirupsen/logrus"
+)
+
+var (
+	cgroupv2PathPattern = regexp.MustCompile(`0:.*?:(.*)`)
+)
+
+// LookupCgroupv2 returns the cgroupv2 ID for pid.
+func LookupCgroupv2(cgrouplru *lru.SyncedLRU[PID, string], pid PID) (string, error) {
+	id, ok := cgrouplru.Get(pid)
+	if ok {
+		return id, nil
+	}
+
+	// Slow path
+	f, err := os.Open(fmt.Sprintf("/proc/%d/cgroup", pid))
+	if err != nil {
+		return "", err
+	}
+	defer f.Close()
+
+	var genericCgroupv2 string
+	scanner := bufio.NewScanner(f)
+	buf := make([]byte, 512)
+	// Providing a predefined buffer overrides the internal buffer that Scanner uses (4096 bytes).
+	// We can do that and also set a maximum allocation size on the following call.
+	// With a maximum of 4096 characters path in the kernel, 8192 should be fine here. We don't
+	// expect lines in /proc/<PID>/cgroup to be longer than that.
+	scanner.Buffer(buf, 8192)
+	var pathParts []string
+	for scanner.Scan() {
+		line := scanner.Text()
+		pathParts = cgroupv2PathPattern.FindStringSubmatch(line)
+		if pathParts == nil {
+			log.Debugf("Could not extract cgroupv2 path from line: %s", line)
+			continue
+		}
+		genericCgroupv2 = pathParts[1]
+		break
+	}
+
+	// Cache the cgroupv2 information.
+	// To avoid busy lookups, also empty cgroupv2 information is cached.
+	cgrouplru.Add(pid, genericCgroupv2)
+
+	return genericCgroupv2, nil
+}

libpf/convenience.go

@@ -5,34 +5,14 @@ package libpf // import "go.opentelemetry.io/ebpf-profiler/libpf"
 
 import (
 	"context"
-	"fmt"
 	"math/rand/v2"
-	"os"
 	"reflect"
 	"time"
 	"unsafe"
 
 	log "github.com/sirupsen/logrus"
 )
 
-// WriteTempFile writes a data buffer to a temporary file on the filesystem. It
-// is the callers responsibility to clean up that file again. The function returns
-// the filename if successful.
-func WriteTempFile(data []byte, directory, prefix string) (string, error) {
-	file, err := os.CreateTemp(directory, prefix)
-	if err != nil {
-		return "", err
-	}
-	defer file.Close()
-	if _, err := file.Write(data); err != nil {
-		return "", fmt.Errorf("failed to write data to temporary file: %w", err)
-	}
-	if err := file.Sync(); err != nil {
-		return "", fmt.Errorf("failed to synchronize file data: %w", err)
-	}
-	return file.Name(), nil
-}
-
 // SleepWithJitter sleeps for baseDuration +/- jitter (jitter is [0..1])
 func SleepWithJitter(baseDuration time.Duration, jitter float64) {
 	time.Sleep(AddJitter(baseDuration, jitter))

libpf/libpf.go

@@ -5,8 +5,6 @@ package libpf // import "go.opentelemetry.io/ebpf-profiler/libpf"
 
 import (
 	"encoding/json"
-	"fmt"
-	"math"
 	"time"
 )
 
@@ -32,36 +30,9 @@ func NowAsUInt32() uint32 {
 	return uint32(time.Now().Unix())
 }
 
-// UnixTime64 represents nanoseconds or (reduced precision) seconds since epoch.
+// UnixTime64 represents nanoseconds since epoch.
 type UnixTime64 uint64
 
-func (t UnixTime64) MarshalJSON() ([]byte, error) {
-	if t > math.MaxUint32 {
-		// Nanoseconds, ES does not support 'epoch_nanoseconds' so
-		// we have to pass it a value formatted as 'strict_date_optional_time_nanos'.
-		out := []byte(fmt.Sprintf("%q",
-			time.Unix(0, int64(t)).UTC().Format(time.RFC3339Nano)))
-		return out, nil
-	}
-
-	// Reduced precision seconds-since-the-epoch, ES 'epoch_second' formatter will match these.
-	out := []byte(fmt.Sprintf("%d", t))
-	return out, nil
-}
-
-// Unix returns the value as seconds since epoch.
-func (t UnixTime64) Unix() int64 {
-	if t > math.MaxUint32 {
-		// Nanoseconds, convert to seconds-since-the-epoch
-		return time.Unix(0, int64(t)).Unix()
-	}
-
-	return int64(t)
-}
-
-// Compile-time interface checks
-var _ json.Marshaler = (*UnixTime64)(nil)
-
 // AddressOrLineno represents a line number in an interpreted file or an offset into
 // a native file.
 type AddressOrLineno uint64

libpf/libpf_test.go

@@ -4,12 +4,9 @@
 package libpf
 
 import (
-	"fmt"
-	"strconv"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
 )
 
 func TestTraceType(t *testing.T) {
@@ -45,35 +42,3 @@ func TestTraceType(t *testing.T) {
 		assert.Equal(t, test.str, test.ty.String())
 	}
 }
-
-func TestUnixTime64_MarshalJSON(t *testing.T) {
-	tests := []struct {
-		name string
-		time UnixTime64
-		want []byte
-	}{
-		{
-			name: "zero",
-			time: UnixTime64(0),
-			want: []byte(strconv.Itoa(0)),
-		},
-		{
-			name: "non-zero, seconds since the epoch",
-			time: UnixTime64(1710349106),
-			want: []byte(strconv.Itoa(1710349106)),
-		},
-		{
-			name: "non-zero, nanoseconds since the epoch",
-			time: UnixTime64(1710349106864964685),
-			want: []byte(fmt.Sprintf("%q", "2024-03-13T16:58:26.864964685Z")),
-		},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			b, err := test.time.MarshalJSON()
-			require.NoError(t, err)
-			assert.Equal(t, test.want, b)
-		})
-	}
-}