Skip to contentSkip to conversation

perf test: Skip failing 'perf test aslr' test case

PATCH Applied Thomas Richter 3 participants 3 messages Jun 29 – Jun 30

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 perf test: Skip failing 'perf test aslr' test case

V2: In response to sashiko:

Quoted 3 lines

[Severity: Medium]
Does unconditionally setting err=2 here mask prior test failures?

Do not change the return code when skipping the test. The failure is printed on stdout anyway and does not hide any previouly failed tests.

Added correct commit-id as suggested by Namhyung Kim

The test case 'perf test aslr' fails on s390. The root cause of the failure is subtest test_callchain_aslr. This test invokes command

 # ./perf record -g -e task-clock:u -o /tmp/perf-test-aslr \
	-- perf test -w noploop 3

to generate a call stack using event task-clock:u. On s390 this defaults to '--call-graph dwarf' whereas on x86_64 this defaults to framepointer (fp) format. The command

 # ./perf inject --aslr -i /tmp/perf-test-aslr

now scans all SAMPLE entries recorded in the perf.data file to convert possible addresses. This is done in aslr_tool__process_sample() looking at sample_type bits PERF_SAMPLE_IP, PERF_SAMPLE_TID, PERF_SAMPLE_TIME, PERF_SAMPLE_PERIOD, PERF_SAMPLE_CALLCHAIN, PERF_SAMPLE_REGS_USER and PERF_SAMPLE_STACK_USER.
On s390 the samples do not contain FP entries
of type PERF_SAMPLE_CALLCHAIN (the bit is set in sample_type, but the number of FP entries is 0).
The processing enters the PERF_SAMPLE_STACK_USER portion where the data is copied to the newly constructed sample and then aborted with this warning:

/* TODO: can this be less conservative? */ pr_debug("Dropping stack user sample as possible ASLR leak\n");

With command line option '--call-graph dwarf' the new output file does not contain any samples at all. This leads to a missing $new_addr value in the shell script and a failure.

Fix this and skip this subtest. Emit a hint that this subtest is currently unsupported on all platform when option --call-graph dwarf is selected.
Since one subtest is skipped, the complete test is reported as skipped.

Fixes 190c45463844 ("perf test: Add inject ASLR test")
---
 tools/perf/tests/shell/inject_aslr.sh | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
tools/perf/tests/shell/inject_aslr.sh +8 -2
index c00461828ea7..d83c2591db8f 100755
@@ -135,8 +135,14 @@ test_callchain_aslr() {
     echo "Callchain ASLR test [Failed - no noploop samples in original file]"
     err=1
   elif [ -z "$new_addr" ]; then
-    echo "Callchain ASLR test [Failed - could not find remapped address]"
-    err=1
+    if perf evlist -v -i "${data}" | grep -q 'sample_type:.*STACK_USER'; then
+      echo "Dropping stack user sample as possible ASLR leak"
+      echo "Call-graph dwarf not supported with 'perf inject --aslr'"
+      echo "Callchain ASLR test [Skip]"
+    else
+      echo "Callchain ASLR test [Failed - could not find remapped address]"
+      err=1
+    fi
   elif [ "$orig_addr" = "$new_addr" ]; then
     echo "Callchain ASLR test [Failed - addresses are not remapped]"
     err=1
Signature · 1 line
2.54.0
#1 lore raw
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] Commit message contradicts script behavior regarding test skip exit code, hiding skipped status…

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Commit message contradicts script behavior regarding test skip exit code, hiding skipped status from the test runner. --

commit 0ec9bd8ca22ab0ece501382558440bfde5e6fdd0
Author: Thomas Richter <[email protected]>

perf test: Skip failing 'perf test aslr' test case

This commit addresses a failure in the 'perf test aslr' test case on s390 due to missing dwarf call-graph support in perf inject. It modifies the script to drop the stack user sample and emit a skip message instead of failing.

Quoted 5 lines
tools/perf/tests/shell/inject_aslr.sh +0 -0
index c00461828ea79..d83c2591db8f6 100755

tools/perf/tests/shell/inject_aslr.sh test_callchain_aslr() { · line 141

Quoted 9 lines
tools/perf/tests/shell/inject_aslr.sh +4 -2
@@ -135,8 +135,14 @@ test_callchain_aslr() {
     echo "Callchain ASLR test [Failed - no noploop samples in original file]"
     err=1
   elif [ -z "$new_addr" ]; then
-    echo "Callchain ASLR test [Failed - could not find remapped address]"
-    err=1
+    if perf evlist -v -i "${data}" | grep -q 'sample_type:.*STACK_USER'; then
+      echo "Dropping stack user sample as possible ASLR leak"
+      echo "Call-graph dwarf not supported with 'perf inject --aslr'"
+      echo "Callchain ASLR test [Skip]"

[Severity: Medium]
The commit message states "Since one subtest is skipped, the complete test is reported as skipped." However, it appears err is not set to 2 (the standard skip code) here, meaning the script will exit with 0 (Ok) rather than being reported as skipped by the test runner.

If setting err=2 was intentionally avoided because the script's cleanup function currently assumes any non-zero exit code is a failure and leaves temporary directories behind, would it be better to redesign the error handling to support returning 2 without leaking temp files, or should the commit message be corrected to reflect the actual exit status?

Quoted 7 lines
tools/perf/tests/shell/inject_aslr.sh +4 -0
+    else
+      echo "Callchain ASLR test [Failed - could not find remapped address]"
+      err=1
+    fi
   elif [ "$orig_addr" = "$new_addr" ]; then
     echo "Callchain ASLR test [Failed - addresses are not remapped]"
     err=1
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
maintainer Applied
#2 lore raw
In reply to Thomas Richter
Quoted 9 lines · Thomas Richter

V2: In response to sashiko:

Quoted 3 lines

[Severity: Medium]
Does unconditionally setting err=2 here mask prior test failures?

Do not change the return code when skipping the test. The failure is printed on stdout anyway and does not hide any previouly failed tests.

[...]

Applied to perf-tools-next, thanks!

Best regards,
Namhyung