On Tue, Nov 22, 2022 at 10:53:00AM +0100, Carlos Maiolino wrote: > On Mon, Nov 21, 2022 at 04:25:25PM -0800, Darrick J. Wong wrote: > > On Mon, Nov 21, 2022 at 03:35:47PM +0100, Carlos Maiolino wrote: > > > Hello. > > > > > > The xfsprogs, for-next branch located at: > > > > > > https://git.kernel.org/pub/scm/fs/xfs/xfsprogs-dev.git/log/?h=for-next > > > > > > has just been updated. > > > > > > The new head is: > > > > > > b827e2318 xfs: fix sb write verify for lazysbcount > > > > > > > > > This update contains only the libxfs-sync for Linux 6.1, and will serve as a > > > base for the xfsprogs 6.1 release. > > > Please, let me know if any issue. > > > > > > > > > The following commits are now in the for-next tree: > > > > > > [b827e2318] xfs: fix sb write verify for lazysbcount > > > > Why was this commit merged for xfsprogs 6.1? That patch is queued for > > kernel 6.2 in for-next, but the merge window is not open yet. > > > > Since 6.1 isn't out yet, I did a `libxfs-apply <last libxfscommit>..`, my fault > for not spotting you've patches already queued for 6.2. I'm planning to release > xfsprogs-6.1 some time later after linux 6.1 is out, so this patch will already > be in Linus's tree by the time. I can also get this patch out of the tree when I > push the other patches I have queued up, although I'm trying to avoid doing > forced updates to the tree. Is it ok for you to leave it in the tree, or better > remove it by now? For this one patch it doesn't matter since the race only happens when there are multiple threads updating the inode summary counts. Userspace doesn't do any concurrent inobt updates, so the fix arriving early ought to be benign. (That said, the only thing you have to do if you push -f is let everyone know that for-next has been **REBASED**, but I agree with the sentiment of only doing it when it's really important.) --D > Cheers > > -- > Carlos Maiolino
