quota deadlock on 6.1-RC1

David Kirchner dpk at dpk.net
Fri May 5 17:56:26 UTC 2006

On 5/5/06, Pawel Jakub Dawidek <pjd at freebsd.org> wrote:
> It isn't good to release a software with known, documented bugs, but its
> better than shipping an untested software with god-one-knows unknown
> bugs.

There's another reasonable option: the known buggy code could be
disabled by default, and left to the user to enable if they want to
test it and provide feedback. The feature is in no way critical, so
the harm would be minimal.

Another option is to delay the release until the bugs are fixed and
patched, but the impression I get here is that releases absolutely
must occur, so that option won't be explored.

The only reason I have been pushing this issue is because the bug
affects the core subsystem of the OS, and because when the bug occurs,
the average user is left with no feedback and no options other than to
hard-reset the box and cross his fingers. I don't think this should be
acceptable. I'll continue to apply the workaround locally, and will
continue to recommend others do the same, and I guess that'll be that.

Thank you for the feedback. I'll re-file my findings in a new PR so
the issue can be tracked outside this thread (which has strayed
significantly from the original subject, my apologies).

David 'dpk' Kirchner

More information about the freebsd-stable mailing list