From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Date: Wed, 29 May 2013 00:19:04 +0200 Subject: [Buildroot] [PATCH] fs/ext2: further fix to the UUID In-Reply-To: <1369779097-28915-1-git-send-email-yann.morin.1998@free.fr> (Yann E. MORIN's message of "Wed, 29 May 2013 00:11:37 +0200") References: <1369779097-28915-1-git-send-email-yann.morin.1998@free.fr> Message-ID: <8738t69293.fsf@dell.be.48ers.dk> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net >>>>> "Yann" == Yann E MORIN writes: Yann> From: "Yann E. MORIN" Yann> Turned out that setting a nil-UUID is no better than clearing it. Yann> What currently happens is as follows: Yann> - first, genext2fs does not generate a UUID Yann> - then we tune2fs to upgrade the filesystem Yann> - then we run fsck, which generates a random UUID Yann> - then we re-run tune2fs to set a nil-UUID Yann> So, on the target, if the file system is improperly unmounted (eg. Yann> with a power failure), on next boot, fsck may be run, and a new Yann> random UUID will be generated. Yann> *However*, fsck improperly updates the filesystem when it adds the Yann> UUID, and there are a few group descriptor checksum errors. Yann> Those errors will go undetected until the next fsck, which will then Yann> block for user input (bad on embedded systems, bad). Yann> Fix that by systematically generating a random UUID _before_ we call Yann> to fsck. Yann> A random UUID is not so bad, after all, since there are already so Yann> many sources of unpredictability in the filesystem: files date and Yann> ordering, files content (date, paths...) which renders a fixed UUID Yann> unneeded. Yann> And it is still possible to set the UUID in a post-image script if Yann> needed, anyway. Committed, thanks. -- Bye, Peter Korsgaard