On Tue, Oct 17, 2017 at 01:22:27PM +0200, Arnd Bergmann wrote:
On Tue, Oct 17, 2017 at 1:13 PM, Greg KH gregkh@google.com wrote:
On Mon, Oct 16, 2017 at 02:50:22PM +0200, Arnd Bergmann wrote:
On Mon, Oct 16, 2017 at 12:11 PM, Greg KH gregkh@google.com wrote:
There are three issues here:
- you looked at the wrong link. The board that produced the failure
was 'x15' (32-bit ARM), while the test-run you linked to is for 64-bit x86, and fcntl36 did not actually fail there. This is obviously confusing, since we run the same test suite once for each board, and the results are in different places.
Huh? There was only one link here, where else should I be looking other than the provided link in the email, and then scrolling down to click on the red "fail" button for the test itself?
What did I miss on the page that I should have looked at instead?
The link you clicked was presumably this one
https://qa-reports.linaro.org/lkft/linux-stable-rc-4.4-oe/build/v4.4.92
From there, you can see all the results grouped by test suite (boot, kselftest, ltp-containers-tests, ltp-syscalls-test, ...) and by target machine (juno-r2, x15, dell-poweredge-r200, possibly hikey and others as well depending on kernel release).
However, you get all the combinations, so in this report there are three test runs for ltp-syscalls-test, one for each board we had a booting kernel for. Each of the ltp-syscalls-test runs had some succeeding and some known-failing test cases inside it, but only the x15 one was flagged as a regression in the output
Ah, ugh, that's a mess, yeah, having different servers does not make that simple... I'll try to be more careful next time, but I did click on the failed-test button, so it might have been the wrong arch/server, but it was still a failure :)
thanks,
greg k-h