Closed Bug 1528436 Opened 1 year ago Closed 1 year ago

Perma TEST-UNEXPECTED-FAIL : xperf: File 'c:\windows\system32\dxgi.dll' (normalized from 'C:\Windows\System32\dxgi.dll') was accessed and we were not expecting it. DiskReadCount: 2, DiskWriteCount: 0, DiskReadBytes: 32768, DiskWriteBytes: 0

Categories

(Core :: Audio/Video: MediaStreamGraph, defect, P5)

defect

Tracking

()

RESOLVED FIXED
mozilla67
Tracking Status
firefox-esr60 --- unaffected
firefox65 --- unaffected
firefox66 --- fixed
firefox67 --- fixed

People

(Reporter: intermittent-bug-filer, Assigned: rhunt)

References

Details

(Keywords: intermittent-failure, regression)

Attachments

(1 file)

#[markdown(off)]
Filed by: csabou [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=228501142&repo=mozilla-beta

https://queue.taskcluster.net/v1/task/ATd_IYvoQ3yWkziR7u-OJA/runs/0/artifacts/public/logs/live_backing.log

22:09:58 INFO - PID 6152 | [Parent 6152, Gecko_IOThread] WARNING: pipe error: 109: file z:/build/build/src/ipc/chromium/src/chrome/common/ipc_channel_win.cc, line 332
22:09:58 INFO - PID 6152 | [Child 3664, Chrome_ChildThread] WARNING: pipe error: 109: file z:/build/build/src/ipc/chromium/src/chrome/common/ipc_channel_win.cc, line 332
22:09:58 INFO - PID 6152 | [Child 3664, Chrome_ChildThread] WARNING: pipe error: 109: file z:/build/build/src/ipc/chromium/src/chrome/common/ipc_channel_win.cc, line 332
22:09:59 INFO - TEST-INFO | 6152: exit 0
22:10:02 INFO - The trace you have just captured "Z:\task_1550179438\build\test.etl.kernel" may contain personally identifiable information, including but not necessarily limited to paths to files accessed, paths to registry accessed and process names. Exact information depends on the events that were logged. Please be aware of this when sharing out this trace with other people.
22:10:02 INFO - The trace you have just captured "Z:\task_1550179438\build\test.etl.user" may contain personally identifiable information, including but not necessarily limited to paths to files accessed, paths to registry accessed and process names. Exact information depends on the events that were logged. Please be aware of this when sharing out this trace with other people.
22:10:12 INFO - Merged Etl: test.etl
22:12:20 INFO - reading etl filename: test.etl
22:12:20 INFO - etlparser: in readfile: test.etl.csv
22:12:20 INFO - TEST-UNEXPECTED-FAIL : xperf: File 'c:\windows\system32\dxgi.dll' (normalized from 'C:\Windows\System32\dxgi.dll') was accessed and we were not expecting it. DiskReadCount: 2, DiskWriteCount: 0, DiskReadBytes: 32768, DiskWriteBytes: 0
22:12:22 INFO - extending with xperf!
22:12:22 INFO - Detected a regression for tp5n
22:12:22 INFO - TEST-UNEXPECTED-FAIL | tp5n | Talos has found a regression, if you have questions ask for help in irc on #perf
22:12:22 ERROR - Traceback (most recent call last):
22:12:22 INFO - File "Z:\task_1550179438\build\tests\talos\talos\run_tests.py", line 300, in run_tests
22:12:22 INFO - talos_results.add(mytest.runTest(browser_config, test))
22:12:22 INFO - File "Z:\task_1550179438\build\tests\talos\talos\ttest.py", line 64, in runTest
22:12:22 INFO - return self._runTest(browser_config, test_config, setup)
22:12:22 INFO - File "Z:\task_1550179438\build\tests\talos\talos\ttest.py", line 268, in _runTest
22:12:22 INFO - 'Talos has found a regression, if you have questions'
22:12:22 INFO - TalosRegression: Talos has found a regression, if you have questions ask for help in irc on #perf
22:12:22 INFO - TEST-INFO took 255704ms
22:12:22 INFO - SUITE-END | took 255s
22:12:22 ERROR - Return code: 1
22:12:22 WARNING - setting return code to 1
22:12:22 ERROR - # TBPL WARNING #

I'm a bit confused on why this is failing on beta, but not central. Does this test only run on beta?

Joel, is this something that should be fixed by adding an entry to the xperf whitelist? Or is a new DLL being loaded considered a regression?

I see a similar issue in bug 1491434.

Flags: needinfo?(rhunt) → needinfo?(jmaher)

We have different configs in beta which means there are different patterns that we see from the test results. In this case there are a small number of beta specific failures in the xperf suite which we do just add to xperf_whitelist.json. I believe dxdi.dll is related to directX. For any of the c:\windows* .dll files, I typically add them in- it is a good idea to ensure that the patch in question isn't causing undue work (i.e. one patch was loading 5 additional font files). In this case as it is beta only, I would just add it to the file and move on.

Flags: needinfo?(jmaher)

Thanks Joel. This DLL is related to DirectX and being used for HDR telemetry. The telemetry is submitted on startup, so I think this is okay to whitelist.

Pushed by rhunt@eqrion.net:
https://hg.mozilla.org/integration/autoland/rev/66fe745ebaec
Add dxgi.dll to xperf_whitelist.json r=jmaher
Status: NEW → RESOLVED
Closed: 1 year ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla67
Blocks: 1514840
Whiteboard: [checkin-needed-beta]
Whiteboard: [checkin-needed-beta]
You need to log in before you can comment on or make changes to this bug.