Message ID | 1349339900.4260.10.camel@pasglop (mailing list archive) |
---|---|
State | Not Applicable |
Headers | show |
On Thu, Oct 4, 2012 at 1:38 AM, Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote: > > First, however, a note about the pull request details ... the diffstat looks > completely on crack, any idea what's up ? It sees piles of files modified > in various other archs & generic code but I see no patch in that branch > that touches any of them. So this happens if you have "reverse merges" (ie you've pulled my tree, or some other tree I've pulled), and there is no longer a single clear common point that you started from. In that case, there is no simple diff for the "what has changed since that original point", and to get the diff for the merge you actually have to do the merge and check the end result. git-request-pull doesn't do that, it just assumes it's the simple case of some single common point. The fact that you haven't seen it until now just means that you've generally done a good job at keeping your powerpc tree "clean" from other trees, and containing only your own work. Linus
On Thu, Oct 4, 2012 at 1:38 AM, Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote: > > The following changes since commit 271fd03a3013b106ccc178d54219c1be0c9759b7: > > powerpc/powernv: I/O and memory alignment for P2P bridges (2012-09-11 16:59:47 -0600) > > are available in the git repository at: > > git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc.git Hmm. There's nothing there. Did you mean for me to pull some branch/tag you didn't mention? Linus
On Thu, 2012-10-04 at 09:00 -0700, Linus Torvalds wrote: > So this happens if you have "reverse merges" (ie you've pulled my > tree, or some other tree I've pulled), and there is no longer a single > clear common point that you started from. In that case, there is no > simple diff for the "what has changed since that original point", and > to get the diff for the merge you actually have to do the merge and > check the end result. git-request-pull doesn't do that, it just > assumes it's the simple case of some single common point. That was more/less my guess. Thanks. > The fact that you haven't seen it until now just means that you've > generally done a good job at keeping your powerpc tree "clean" from > other trees, and containing only your own work. Heh, yeah I try to :-) In this specific case, this was a topic branch from Bjorn which some pre-requisite patches to the generic PCI code that Gavin did to allow us to get rid of some custom resource allocation crap, so the "getting rid of" had a dependency on that topic branch. Unfortunately, Bjorn had also based that branch on top of some larger cleanup patch so I ended up pulling that as well from him, which tripped the whole thing. Cheers, Ben.
On Thu, 2012-10-04 at 09:02 -0700, Linus Torvalds wrote: > Hmm. There's nothing there. > > Did you mean for me to pull some branch/tag you didn't mention? Heh, yeah, the mirrors hadn't updated yet so git request-pull didn't put the branch name in. It's common but I usually fix it up by hand. Since I was probably too tired for thinking straight, I forgot that and forgot to put the proper branch name in the email subject as well... argh :-) It's in my next branch. Mirrors should be good by now, it's at: git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc.git next Sorry for the misshap. Cheers, Ben.