On Wed, Apr 10, 2024 at 06:00:58AM +0200, Christoph Hellwig wrote: > On Tue, Apr 09, 2024 at 05:05:28PM -0700, Darrick J. Wong wrote: > > From: Darrick J. Wong <djwong@xxxxxxxxxx> > > > > Per some very late review comments, capture the generation numbers of > > both inodes involved in a file content exchange operation so that we > > don't accidentally target files with have been reallocated. > > > > Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx> > > --- > > I'm throwing this one on the pile since I guess it's not so hard to add > > the generation number to a brand new log item. > > It does looks fine to me, but it leaves the question open: why here > and not elsewhere. And the answer based on the previous discussions > is that this is the first new log item after the problem was known > and we'll need to eventually rev the other ino based items as well. > Maybe capture this in a comment? /* * This log intent item targets inodes, which means that it effectively * contains a file handle. Check that the generation numbers match the * intent item like we do for other file handles. This is the first * new log intent item to be defined after this validation weakness was * identified, which is why recovery for other items do not check this. */ How about that? --D
