usm: shared-libraries: Ignore O_WRONLY opens #32402
Conversation
If a file is opened with O_WRONLY, ignore it since it's never going to be mmap(2)'d in to the process. This prevents trying to hook files which are being copied or created, such as when containerd pulls or imports a container image.
vitkyrka
added
changelog/no-changelog
team/usm
github-actions
bot
added
component/system-probe
short review
github-actions
bot
added
medium review
|
[Fast Unit Tests Report] On pipeline 51759871 (CI Visibility). The following jobs did not run any unit tests: Jobs:
If you modified Go files and expected unit tests to run in these jobs, please double check the job logs. If you think tests should have been executed reach out to #agent-devx-help |
Uncompressed package size comparisonComparison with ancestor Diff per package
Decision |
Co-authored-by: Guy Arbitman <guy20495@gmail.com>
eBPF complexity changesSummary result: ✅ - stable
shared_libraries detailsshared_libraries [programs with changes]
shared_libraries [programs without changes]
This report was generated based on the complexity data for the current branch vincent.whitchurch/sharedlib-wronly (pipeline 51759871, commit 0d6b4a7) and the base branch main (commit 9905116). Objects without changes are not reported. Contact #ebpf-platform if you have any questions/feedback. Table complexity legend: 🔵 - new; ⚪ - unchanged; 🟢 - reduced; 🔴 - increased |
|
@vitkyrka I do recommend on adding a UT to ensure the logic is working as expected |
Test changes on VMUse this command from test-infra-definitions to manually test this PR changes on a VM: inv aws.create-vm --pipeline-id=51759871 --os-family=ubuntuNote: This applies to commit 0d6b4a7 |
Regression DetectorRegression Detector ResultsMetrics dashboard Baseline: 9905116 Optimization Goals: ✅ No significant changes detected
|
| perf | experiment | goal | Δ mean % | Δ mean % CI | trials | links |
|---|---|---|---|---|---|---|
| ➖ | quality_gate_logs | % cpu utilization | +0.83 | [-2.35, +4.02] | 1 | Logs |
| ➖ | quality_gate_idle | memory utilization | +0.17 | [+0.13, +0.20] | 1 | Logs bounds checks dashboard |
| ➖ | file_to_blackhole_1000ms_latency | egress throughput | +0.15 | [-0.65, +0.95] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | +0.05 | [-0.62, +0.73] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency | egress throughput | +0.04 | [-0.80, +0.89] | 1 | Logs |
| ➖ | file_to_blackhole_100ms_latency | egress throughput | +0.04 | [-0.64, +0.72] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http2 | egress throughput | +0.03 | [-0.81, +0.88] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http1 | egress throughput | +0.02 | [-0.85, +0.89] | 1 | Logs |
| ➖ | file_to_blackhole_300ms_latency | egress throughput | +0.01 | [-0.64, +0.65] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api | ingress throughput | -0.00 | [-0.12, +0.12] | 1 | Logs |
| ➖ | tcp_dd_logs_filter_exclude | ingress throughput | -0.00 | [-0.01, +0.01] | 1 | Logs |
| ➖ | file_to_blackhole_1000ms_latency_linear_load | egress throughput | -0.03 | [-0.50, +0.44] | 1 | Logs |
| ➖ | file_to_blackhole_500ms_latency | egress throughput | -0.11 | [-0.89, +0.67] | 1 | Logs |
| ➖ | file_tree | memory utilization | -0.35 | [-0.47, -0.22] | 1 | Logs |
| ➖ | quality_gate_idle_all_features | memory utilization | -0.44 | [-0.53, -0.36] | 1 | Logs bounds checks dashboard |
| ➖ | tcp_syslog_to_blackhole | ingress throughput | -0.46 | [-0.53, -0.40] | 1 | Logs |
Bounds Checks: ✅ Passed
| perf | experiment | bounds_check_name | replicates_passed | links |
|---|---|---|---|---|
| ✅ | file_to_blackhole_0ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http2 | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http2 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency_linear_load | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | memory_usage | 10/10 | |
| ✅ | quality_gate_idle | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_idle_all_features | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_logs | lost_bytes | 10/10 | |
| ✅ | quality_gate_logs | memory_usage | 10/10 |
Explanation
Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%
Performance changes are noted in the perf column of each table:
- ✅ = significantly better comparison variant performance
- ❌ = significantly worse comparison variant performance
- ➖ = no significant change in performance
A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".
For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:
-
Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
-
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
-
Its configuration does not mark it "erratic".
CI Pass/Fail Decision
✅ Passed. All Quality Gates passed.
- quality_gate_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_logs, bounds check lost_bytes: 10/10 replicas passed. Gate passed.
- quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
| @@ -133,6 +153,8 @@ int tracepoint__syscalls__sys_exit_openat(exit_sys_ctx *args) { | |||
| SEC("tracepoint/syscalls/sys_enter_openat2") | |||
| int tracepoint__syscalls__sys_enter_openat2(enter_sys_openat2_ctx *args) { | |||
| CHECK_BPF_PROGRAM_BYPASSED() | |||
| // Unlike the other variants, openat2(2) has the flags embedded inside the | |||
| // how argument; we don't bother trying to accessing it for now. | |||
There was a problem hiding this comment.
Why skip it? From what I know, libc prefers openat2 so you would get more benefit from covering this syscall.
What does this PR do?
If a file is opened with O_WRONLY, ignore it since it's never going to be mmap(2)'d in to the process. This prevents trying to hook files which are being copied or created, such as when containerd pulls or imports a container image.
Motivation
https://datadoghq.atlassian.net/wiki/spaces/UT/pages/4105764896
Describe how you validated your changes
Automated test added.
containerd scenario manually verified with:
(The import/export is only to avoid repeatedly download the image.) Everytime the pull or import is executed after removing the image, containerd writes to a
libsslfile.Before the patch, the libssl path gets blocked (due to the file being partially written when we try to analyze it), checked with the debug endpoint
sudo curl --unix /opt/datadog-agent/run/sysprobe.sock http://unix/network_tracer/debug/usm/blocked_processes.After the patch, the path is not blocked.
Possible Drawbacks / Trade-offs
Additional Notes