On Mon, Dec 02, 2019 at 03:22:31PM -0800, Linus Torvalds wrote: > On Sun, Dec 1, 2019 at 10:48 AM Darrick J. Wong <djwong@xxxxxxxxxx> wrote: > > > > FYI, Stephen Rothwell reported a merge conflict with the y2038 tree at > > the end of October[1]. His resolution looked pretty straightforward, > > though the current y2038 for-next branch no longer changes fs/ioctl.c > > (and the changes that were in it are not in upstream master), so that > > may not be necessary. > > The changes and conflicts are definitely still there (now upstream), > I'm not sure what made you not see them. But thanks for the note, I > compared my end result with linux-next to verify. Aha! I pulled master yesterday morning, tried a test merge with xfs, saw the lack of merge conflicts, and sent you the xfs pull request. A few hours later you pulled in the compat ioctl changes from Arnd's git tree, but the branch in his repo that feeds the -next tree doesn't contain the compat ioctl changes, so I assumed that meant he wasn't going to send them for 5.5... and then thought better of myself and attached an FYI anyway. > My resolution is different from Stephen's. All my non-x86-64 FS_IOC_* > cases just do "goto found_handler", because the compat case is > identical for the native case outside of the special x86-64 alignment > behavior, and I think that's what Arnd meant to happen. Yeah, that looks correct to me. Stephen's solution backed out the changes that Arnd made for the !x86_64 compat ioctl case, so I or someone would have had to re-apply them. > There was some other minor difference too, but it's also possible I > could have messed up, so cc'ing Stephen and Arnd on this just in case > they have comments. <nod> Thanks for sorting this out. --D > > > Linus
