From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hans Reiser Subject: Re: [2.4.18]: Reiserfs: vs-2120: add_save_link: insert_item returned -28 Date: Fri, 16 Jan 2004 16:29:40 +0300 Message-ID: <4007E744.4060102@namesys.com> References: <200401091622.41352.lkml@kcore.org> <4007A54A.3030005@namesys.com> <1074244260.2250.45.camel@tribesman.namesys.com> <200401161229.30831.lkml@kcore.org> <4007DBC6.4040606@namesys.com> <1074257413.2250.53.camel@tribesman.namesys.com> <4007DF3E.6010307@namesys.com> <1074259779.29546.79.camel@tiny.suse.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <1074259779.29546.79.camel@tiny.suse.com> List-Id: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Chris Mason Cc: Vladimir Saveliev , Jan De Luyck , reiserfs-list@namesys.com Chris Mason wrote: >On Fri, 2004-01-16 at 07:55, Hans Reiser wrote: > > > >>>At this case we do not return ENOSPC. Operation which failed due to >>>ENOSPC was "save link" insertion. The major operation (unlink or >>>truncate) just continues without "save links". (save link are used to be >>>able to complete removal of unlinked/truncated files on reboot if they >>>did not manage to get properly removed/truncated due to unclean >>>shutdown). >>> >>> > > > >>so it is a real bug which can result in space leaks? >> >> > >It's a limitation. The only way around it would be to somehow reserve >the space for the savelink when you create the file. It's a rare case >and the FS carries on as best it can. > >-chris > > > > > > does it leak space? -- Hans