https://bugzilla.redhat.com/show_bug.cgi?id=2449216
--- Comment #25 from [email protected] --- (In reply to Brad Smith from comment #23) > I narrowed down the tests that are skipped with the following notes: > > %{cargo_test -f capi -- -- %{shrink: > --skip _global_fs_dir > --skip _opath > --skip utils::sysctl::tests > }} > > The _global_fs_dir string will ignore a few additional tests not just > failing. The _opath string skips significantly more tests (failing and ok). > Both sets of failures are due to 'Invalid cross-device' link errors. I could > not see a way to tighten these skips without adding a large number of lines. There is an issue/regression with the O_PATH resolver when running in Fedora's infrastructure (or actually, most container workloads) that I hadn't noticed before -- a read-only /proc/sys (default on most container setups) is correctly detected as an overmount (EXDEV/"Invalid cross-device link") when checking the value of /proc/sys/fs/protected_symlinks but that incorrectly causes the lookup to fail, which is the issue you're seeing. I opened an issue about it the other day[1], and I'll include a fix in 0.2.5. Sorry about that, I really should set up a CI job to run the tests in containers as well. :/ Ideally we would also permit those kinds of overmounts (i.e., ones that mount the same underlying path but with ro set) in general but implementing that properly is a little difficult on pre-6.8 kernels (with statmount it should be fairly trivial to do in a race-free way) so I might punt that to 0.2.6. > The sysctl tests triggered panics in rs code without and obvious root cause. > I tried adding additional debug info but that did not help me (not familiar > with rust). I suspect this is also an issue with `/proc/sys`, and I should adjust the tests to have an xfail/skip if /proc/sys is mounted read-only (you would expect an error in this case) but if you could leave a comment in [1] with the error trace that would be great! [1]: https://github.com/cyphar/libpathrs/issues/372 -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component https://bugzilla.redhat.com/show_bug.cgi?id=2449216 Report this comment as SPAM: https://bugzilla.redhat.com/enter_bug.cgi?product=Bugzilla&format=report-spam&short_desc=Report%20of%20Bug%202449216%23c25 -- _______________________________________________ package-review mailing list -- [email protected] To unsubscribe send an email to [email protected] Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/[email protected] Do not reply to spam, report it: https://forge.fedoraproject.org/infra/tickets/issues/new
