<html> <head> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8"> </head> <body> On 4/8/20 3:19 AM, Po-Hsu Lin wrote: <blockquote type="cite" cite="mid:CAMy_GT_8WRGK5Bwdyj81CrJYQkY7iz+QjrCAhZ+YZHoskznD0g@mail.gmail.com"> <pre class="moz-quote-pre" wrap="">Hi Paolo, for this kind of small change which is limited to a specific test case, not sure if this is a good idea, but I think at least you can just use our jenkins server directly as you need to check it on different arches. </pre> </blockquote> Patching Autotest is the correct way to tackle this, since the partitioning for logic for sfdisk is stored in autotest itself. Each of the tests call the autotest API (client/partitioning.py) to handle their storage devices. -sfeole <blockquote type="cite" cite="mid:CAMy_GT_8WRGK5Bwdyj81CrJYQkY7iz+QjrCAhZ+YZHoskznD0g@mail.gmail.com"> <pre class="moz-quote-pre" wrap=""> Just remember to let Sean and I know, and don't forget to watch the test result in case it causes some interruption for the ongoing tests. Note that it looks like this is for ubuntu_xfstests_*, as for the limited test resources and the time it needs for the test, we only have amd64 tested with it. On Mon, Apr 6, 2020 at 6:38 PM Paolo Pisati <a class="moz-txt-link-rfc2396E" href="mailto:paolo.pisati@canonical.com"><paolo.pisati@canonical.com></a> wrote: </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap=""> I tested it in Xenial, Bionic and Focal but only on amd64 - let me spin it on some other arch. BTW, if i have an invasive autotest patch, is there a way for me to spin a private autotest run on several arches/releases? Or do i have to spin every single impacted test manually? Create/allocate the target arch/release, install autotest/autotest/client-tests, run the test, wash, rinse and repeat for every arch/release target pair. On Mon, Apr 6, 2020 at 12:25 PM Colin Ian King <a class="moz-txt-link-rfc2396E" href="mailto:colin.king@canonical.com"><colin.king@canonical.com></a> wrote: </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap=""> On 06/04/2020 11:22, Paolo Pisati wrote: </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">BugLink: <a class="moz-txt-link-freetext" href="https://bugs.launchpad.net/bugs/1864642">https://bugs.launchpad.net/bugs/1864642</a> Every selftest unit test instantiating a virtual_partition class: $ git grep virtual_partition ubuntu_xfstests_btrfs/control:vp = partition.virtual_partition(file_img=file_img, file_size=1024*1024) ubuntu_xfstests_ext4/control:vp = partition.virtual_partition(file_img=file_img, file_size=1024*1024) ubuntu_xfstests_xfs/control:vp = partition.virtual_partition(file_img=file_img, file_size=1024*1024) xfstests/control:vp = partition.virtual_partition(file_img=file_img, file_size=1024*1024) will fail during _init_ with: </pre> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">Created a new DOS disklabel with disk identifier 0x9b4a9cf7. </pre> </blockquote> </blockquote> </blockquote> <pre class="moz-quote-pre" wrap="">/dev/loop1p1: Start sector 0 out of range. Failed to add #1 partition: Numerical result out of range Leaving. because we hardcoded 0 as the starting sector: fix it by letting sfdisk pick the first available sector. Paolo Pisati (1): UBUNTU: SAUCE: sfdisk: let sfdisk decide the first available sector client/partition.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) </pre> </blockquote> <pre class="moz-quote-pre" wrap="">Does this work on every release with non-powerpc arches too? </pre> </blockquote> <pre class="moz-quote-pre" wrap=""> -- bye, p. -- kernel-team mailing list <a class="moz-txt-link-abbreviated" href="mailto:kernel-team@lists.ubuntu.com">kernel-team@lists.ubuntu.com</a> <a class="moz-txt-link-freetext" href="https://lists.ubuntu.com/mailman/listinfo/kernel-team">https://lists.ubuntu.com/mailman/listinfo/kernel-team</a> </pre> </blockquote> <pre class="moz-quote-pre" wrap=""> </pre> </blockquote> <pre class="moz-signature" cols="72">-- -Sean Feole</pre> </body> </html>