On Wed, Aug 09, 2017 at 05:13:35PM -0500, Dan Rue wrote:
On Wed, Aug 09, 2017 at 02:44:17PM -0700, Greg KH wrote:
- ltp-syscalls-tests/pselect01 -- failing since build v4.9.41, from Aug. 7, 2017, 11 a.m.
A new failure? Did this run actually catch something "real"?
There's pselect01 and pselect01_64. It seems that one of the 5 tests within pselect01 and pselect01_64 fail every time. i.e. v4.9.41 failed pselect01_64 and passed pselect01, but the last two builds failed pselect01 and passed pselect01_64.
v4.9.41: See https://lkft.validation.linaro.org/scheduler/job/17875#L7048
pselect01.c:131: FAIL: pselect() slept for too long 3111143us, expected 3000000us, threshold 78000v4.9.41-94-g80d017f: See https://lkft.validation.linaro.org/scheduler/job/17683#L7060
pselect01.c:131: FAIL: pselect() slept for too long 3109769us, expected 3000000us, threshold 78000Here's the test source (we're running the latest ltp release, but i notice this test has since been refactored in master): https://github.com/linux-test-project/ltp/blob/20170516/testcases/kernel/sys...
Hm, this didn't answer my question, is this a new bug I just created, or is it something that has always been there?
Can I be "trusting" these test results yet? If not, that's fine, I will not, otherwise it would be good to verify I didn't break anything here.
thanks,
greg k-h