From mboxrd@z Thu Jan 1 00:00:00 1970 From: "David Dabbs" Subject: fsck.reiser4 problem (was: reiser4 corruption problem) Date: Mon, 30 Aug 2004 21:07:41 -0000 Message-ID: <20040830210735.6104715C98@mail03.powweb.com> Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com Content-Disposition: inline List-Id: Content-Type: text/plain; charset="iso-8859-1" To: reiserfs-list@namesys.com Michael Weissenbacher wrote:=0D =0D >i've investigated this problem further the last days and came to the =0D >following conclusions:=0D >[...]=0D >fsck does not like all contain german umlauts. but otoh there are =0D >filenames with umlauts that are ok!=0D >=0D >here are some filenames that fail:=0D 0123456789012345=0D >BewerbungF=FCrAnw altsb=FCro.doc=0D >Graphik01Marken identit=E4t.sxd=0D >SkriptumzumSemi narF=FChren.doc=0D >WieTeamseffizie ntwerdenk=F6nnen.doc=0D =0D 0123456789012345=0D >but otoh these work:=0D >WerbeK=E4rntnerBa llonwerbung.sxw=0D >Vo 08 - 29. M=E4r z 2001.pdf=0D >Tsch=F6ranKirche. jpg=0D >V=F6lkermarktKirc heGross.jpg=0D >=0D =0D Even though both file sets contain umlauts, or perhaps more accurately exte= nded ASCII chartacters, there is something distinctive in the "failure" set= : the umlauts/extended characters appear after the 15th character. If you a= re using REISER4_LARGE_KEYS, the first fifteen characters will be shifted i= nto the second and third key elements with the final key el containing the = hash of the remaining characters =0D =0D key =3D { [dirhash], [hash_bit+fibre_bits+1st 7 chars], [next 8 chars], [ha= sh] }=0D =0D Code in fs/reiser4/kassign.c assembles the key and uses your chosen hash, R= 5 being the default. If you created the files without failure, could read/o= pened them okay but then FSCK reported problems, could this point to a diff= erence in the hash code (w.r.t. extended ASCII)? I'm on holiday now, so can= not check to see if this suspicion holds any water. =0D =0D David=0D =0D p.s. One other possibility is that there is some extended ASCII variance in= the the fibration code, but this seems unlikely. =0D =0D