On 06/21/2001 03:08 PM, Chris Mason wrote: > > On Thursday, June 21, 2001 03:22:03 AM +0200 Manuel Krause > <[EMAIL PROTECTED]> wrote: > > >>Hi, >> >>at the moment I'm running kernel-2.4.5 with the following patches applied >>(when it came by mailing-list only - named according to the subject and >>then marked) >> >>patch-2.4.5-reiserfs-leak-in-error-paths.mail.patch >>patch-2.4.5-reiserfs-mark-journal-new.mail.patch >> > > These were included in 2.4.6-pre3 > > >>patch-2.4.5-reiserfs-umount-fix.patch aka "super.c patch" >> > > 2.4.6-pre1 > > >>patch-2.4.5-reiserfs-writepage-prealloc-race.mail.patch >> > > 2.4.6-pre4 > > >>patch-2.4.5-truncate_inode_pages.patch (like in lkml and Mr. Reisers >>request for context) >> > > 2.4.6-pre3 > > >>without >>patch-2.4.5-reiserfs-memory_leak.mail.patch aka "Don't worry ?"-patch >>yet - and >> > > This has not been sent for inclusion yet. > > >>without >>problems/errors/oopses related since the last was out. Please, don't ask, >>I like these procedures as it brings back some of the old times' feeling >>for ReiserFS... Hopefully I didn't miss anything. ;-)) >> >> > > You've got every reiserfs patch that has gone to Linus, except for the > knfsd fixes from Neil Brown. If you aren't using NFS, you're up to date. > > >>Except perhaps ... for some df differences before/after "cp -a"ing my >>reiserfs-partitions for backup-purposes (what makes a difference < +2 >>percent of disk space after copying) so I thought to ignore this as I >>don't miss anything and it is below normal technical/natural tolerances). >>:-) >> > > Different usage of the tail option? No, I usually create a new filesystem with identical size and type ( -h r5 -v 2) with the latest mkreiserfs on my backup disk and mount it in the same way (-o noatime,notail) and copy from the original partition to the new one. In this step I usually get the df differences. Then I create a new reiserfs in the original place and copy everything back and don't get df differences. If I didn't use notail I would save more than 100MB of disk space with that installation - what would be very obvious then. This procedure may be a bit complicated. Maybe I do it just to see my reiserfs "growing"... ;-) This complete procedure takes about 15 minutes for the 2 * 1,67GB. This partitions is normally mounted /root and I start the system on a different partition with the same kernel configuration for these copy-sessions. > > >>- >> >>So, now, I don't know if I should give 2.4.6-pre4 a try because of the >>apm<->pci related topics from 2.4.6-pre3 that confused my (mind's) <-> >>(computer's hardware) - relation. >> >>What did go into the new pre-patch 2.4.6-pre4 extra to: >>"-pre4: >> - Chris Mason: ReiserFS pre-allocation locking bugfix" >> >> > > Nothing reiserfs related ;-) pre5 quickly followed pre4, so I'm assuming > there was something wrong with pre4. Haven't read l-k yet though, it might > just be that Linus forgot a patch and wanted it in. > > -chris > Many thanks for your detailed answers! Manuel