On Mon, Jul 01, 2019 at 01:41:29PM -0400, Theodore Ts'o wrote: > On Mon, Jul 01, 2019 at 10:01:59AM -0700, Darrick J. Wong wrote: > > Note that this is not the final format of the patches, because I intend > > to pick a point towards the end of the merge window (after everyone > > else's merges have landed), rebase this series atop that, and push it > > back to Linus. > > So normally Linus isn't psyched about pulling branches that were > rebased at the last minute. I guess we could ask him ahead of time if > he's OK with this plan. Or have you done that already? I've not done so yet, since this is the first time this cleanup has been posted. Assuming I don't hear any loud complaining from anyone, I was going to send the existing iomap 5.3 patches (all three of them) to Linus at the start of the merge window and let him know that I'd like to do the quick cleanup during the second week. I've done tree cleanups this way before without hearing any complaints, though they've never gone outside the xfs tree. > Alternatively you could rebase this on top of v5.3-rc2, after the > merge window closes, and get agreement from the 4 file systems which > are currently iomap users: ext2, ext4, gfs2, and xfs to start their > development trees on top of that common branch for the 5.4 merge > window. After all, it's just moving code around and there are no > substantive changes in this patch series, right? So there's no rush > as I understand things for this to hit mainline. <shrug> I prefer to get this done quickly at the end of the 5.3 merge window to reduce the risk that I'll have to reconcile the cleanup with whatever iomap fixes land during -rc1 to -rc7. I don't know if you've been following the "lift xfs writeback to iomap" thread lately but I've also been thinking about putting out a work branch for 5.4 with the sample writeback code in fs/iomap/writeback.c to see if anyone bites, so we're likely to get such a tree anyway. --D > Cheers, > > - Ted
