-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prevents panics during ELF parsing from crashing process #30725
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM for agent-security files
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Not sure why but on your branch I got errors saying pcap.h couldn't be found. It's probably unrelated but if that means anything to you then do with that information what you will.
// Any panic during parsing is turned into an error. This is necessary since | ||
// there are a bunch of unfixed bugs in debug/elf. | ||
// | ||
// https://github.com/golang/go/issues?q=is%3Aissue+is%3Aopen+debug%2Felf+in%3Atitle |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
which issues from that list actually concern you?
debug/elf switched to use saferio under the hood and I think now it is considered a safe package
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There have been panics fixed as recently as Go 1.22 due to slice bounds. This is a defensive approach, since we are parsing uncontrolled input.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we are using go 1.22 for the datadog-agent repo. I still don't see which item from the list you provided in the comment concerns you
I am not sure i understand the decision to introduce the panic wrapper specifically for the elf parsing case.
if we want to avoid panics we can make a general wrapper for the agent
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I still don't see which item from the list you provided in the comment concerns you
This comment comes from cilium/ebpf
which I grabbed some of the code from. I do share the concern though. Someone from the Go team even responded to me with this:
But running go-fuzz on the package is likely to re-reproduce it (good idea anyway b/c there are likely more bugs after this one)
Separately:
if we want to avoid panics we can make a general wrapper for the agent
It is not possible to have a general panic/recover wrapper and still maintain correct state or control flow. You must do it deep and where you have knowledge of the called code behavior. In this case we know the debug/elf
parsing is not using Goroutines, so we are safe to use panic/recover.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Regardless of the above, this is a defensive approach that costs us nothing.
Go Package Import DifferencesBaseline: 5ed352e
|
Test changes on VMUse this command from test-infra-definitions to manually test this PR changes on a VM: inv create-vm --pipeline-id=48860913 --os-family=ubuntu Note: This applies to commit 8a34060 |
Regression DetectorRegression Detector ResultsMetrics dashboard Baseline: 5ed352e Optimization Goals: ✅ No significant changes detected
|
perf | experiment | goal | Δ mean % | Δ mean % CI | trials | links |
---|---|---|---|---|---|---|
➖ | pycheck_lots_of_tags | % cpu utilization | +1.11 | [-2.30, +4.53] | 1 | Logs |
➖ | quality_gate_idle_all_features | memory utilization | +0.78 | [+0.68, +0.88] | 1 | Logs bounds checks dashboard |
➖ | otel_to_otel_logs | ingress throughput | +0.52 | [-0.16, +1.21] | 1 | Logs |
➖ | quality_gate_idle | memory utilization | +0.41 | [+0.36, +0.46] | 1 | Logs bounds checks dashboard |
➖ | file_to_blackhole_1000ms_latency | egress throughput | +0.12 | [-0.36, +0.60] | 1 | Logs |
➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | +0.07 | [-0.65, +0.80] | 1 | Logs |
➖ | tcp_dd_logs_filter_exclude | ingress throughput | -0.00 | [-0.01, +0.01] | 1 | Logs |
➖ | uds_dogstatsd_to_api | ingress throughput | -0.00 | [-0.09, +0.08] | 1 | Logs |
➖ | file_to_blackhole_100ms_latency | egress throughput | -0.01 | [-0.32, +0.31] | 1 | Logs |
➖ | file_to_blackhole_1000ms_latency_linear_load | egress throughput | -0.01 | [-0.49, +0.48] | 1 | Logs |
➖ | file_to_blackhole_300ms_latency | egress throughput | -0.03 | [-0.22, +0.17] | 1 | Logs |
➖ | file_to_blackhole_0ms_latency | egress throughput | -0.03 | [-0.49, +0.43] | 1 | Logs |
➖ | file_to_blackhole_500ms_latency | egress throughput | -0.08 | [-0.32, +0.17] | 1 | Logs |
➖ | tcp_syslog_to_blackhole | ingress throughput | -0.09 | [-0.13, -0.04] | 1 | Logs |
➖ | file_tree | memory utilization | -0.09 | [-0.21, +0.03] | 1 | Logs |
➖ | basic_py_check | % cpu utilization | -3.72 | [-7.39, -0.06] | 1 | Logs |
Bounds Checks: ❌ Failed
perf | experiment | bounds_check_name | replicates_passed | links |
---|---|---|---|---|
❌ | quality_gate_idle | memory_usage | 0/10 | bounds checks dashboard |
❌ | file_to_blackhole_0ms_latency | lost_bytes | 2/10 | |
❌ | quality_gate_idle_all_features | memory_usage | 9/10 | bounds checks dashboard |
✅ | file_to_blackhole_0ms_latency | 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 | memory_usage | 10/10 | |
✅ | file_to_blackhole_500ms_latency | 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".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't fully understand code in util/safeelf
but I don't see anything strange either 👍 I would suggest to add codeowners for util/safeelf
if you consider adding / modify this folder later
@guyarb I've addressed your comments. |
/merge |
Devflow running:
|
/merge |
Devflow running:
|
b272967
to
8a34060
Compare
/merge |
Devflow running:
|
What does this PR do?
safeelf
package everywhere, which handles panics during ELF parsing.debug/elf
withdepguard
.elf.Section.ReaderAt
isnil
, which is possible.Motivation
The
debug/elf
package is known to have several bugs/limitations which result in panics. Since system-probe parses many arbitrary binaries, this change will prevent those binaries from crashing system-probe if they are malformed in some way.Describe how to test/QA your changes
Existing automated tests pass.
Possible Drawbacks / Trade-offs
Additional Notes