Re: [PATCH v15 4/5] perf test: Add inject ASLR test

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Missing unresolved symbol address filtering in user-space ASLR tests causes spurious test failures.
--

commit 0ac54fddb12723dd235fe2bd67c473f8ec3fe19a
Author: Ian Rogers <irogers@google.com>

perf test: Add inject ASLR test

Adds a new shell test to verify ASLR injection, covering basic address
remapping, pipe mode, callchains, and kernel address remapping.

> diff --git a/tools/perf/tests/shell/inject_aslr.sh b/tools/perf/tests/shell/inject_aslr.sh
> new file mode 100755
> index 0000000000000..d8ded16ba9052
> --- /dev/null
> +++ b/tools/perf/tests/shell/inject_aslr.sh

[ ... ]

> +test_report_aslr() {
> +  echo "Test perf report consistency"

[ ... ]

> +  perf report -i "${data_clean}" --stdio > "${report1}"
> +  perf report -i "${data2}" --stdio > "${report2}"
> +
> +  # Strip headers and compare lines with percentages
> +  grep '%' "${report1}" | grep -v '^#' | sort > "${report1_clean}" || true
> +  grep '%' "${report2}" | grep -v '^#' | sort > "${report2_clean}" || true

[Severity: Medium]
Does this code need to filter out unresolved symbol addresses?

Because --aslr intentionally shifts mappings, unresolved symbols (which perf
report displays as raw hexadecimal addresses) will differ between the two
reports.

Can this cause the diff to fail and the test to flake if perf record catches
samples in the dynamic loader or other unmapped regions?

I noticed test_kernel_report_aslr() avoids this by filtering out hex addresses
with grep -v -E '0x[0-9a-f]{8,}'.

> +
> +  diff -u -w "${report1_clean}" "${report2_clean}" > "${diff_file}" || true

[ ... ]

> +test_pipe_report_aslr() {
> +  echo "Test pipe mode perf report consistency"

[ ... ]

> +  perf report -i "${data_clean}" --stdio > "${report1}"
> +  perf report -i "${data2}" --stdio > "${report2}"
> +
> +  # Strip headers and compare lines with percentages
> +  grep '%' "${report1}" | grep -v '^#' | sort > "${report1_clean}" || true
> +  grep '%' "${report2}" | grep -v '^#' | sort > "${report2_clean}" || true

[Severity: Medium]
Will this comparison also suffer from the same unresolved symbol flakiness
described in test_report_aslr() above?

> +
> +  diff -u -w "${report1_clean}" "${report2_clean}" > "${diff_file}" || true

[ ... ]

> +test_pipe_out_report_aslr() {
> +  echo "Test pipe output mode perf report consistency"

[ ... ]

> +  perf report -i "${data_clean}" --stdio > "${report1}"
> +  perf inject -b --aslr -i "${data}" -o - | perf report -i - --stdio > "${report2}"
> +
> +  # Strip headers and compare lines with percentages
> +  grep '%' "${report1}" | grep -v '^#' | sort > "${report1_clean}" || true
> +  grep '%' "${report2}" | grep -v '^#' | sort > "${report2_clean}" || true

[Severity: Medium]
Could this filter also fail to strip raw hex addresses for unresolved symbols
and cause similar test flakes?

> +
> +  diff -u -w "${report1_clean}" "${report2_clean}" > "${diff_file}" || true

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260606072125.2786845-1-irogers@google.com?part=4