Re: [PATCH v19 00/12] Remus/Libxl: Remus network buffering and drbd disk

[Hongyang Yang <yanghy@cn.fujitsu.com>]

I have done testing and make some fix on top of you branch. See
below link for clear image of what I have fixed:
https://github.com/macrosheep/xen/commits/wip.remus.v21

And thank you for the detailed manual on rearrange patches.
I have rearranged the patches using autosquash and finished
testing on the final remus-v21 patches which I have posted
on list already.
https://github.com/macrosheep/xen/commits/remus-v21

Thank you very much for the work!

Thanks,
Yang.

在 09/26/2014 01:14 AM, Ian Jackson 写道:
> Hongyang Yang writes ("Re: [PATCH v19 00/12] Remus/Libxl: Remus network bufferi> I have sent a v20 which was rebased, you can rework based on that.
>
> I have prepared a branch which I think addresses the comments I had on
> v20:
>
>   git://xenbits.xen.org/people/iwj/xen.git refs/tags/remus-v20..wip.remus.v20
>
>   http://xenbits.xen.org/gitweb/?p=people/iwj/xen.git;a=shortlog;h=refs/heads/wip.remus.v20
>
> Rather than rebasing your series, I have put a pile of individual
> commits on top of it, tagged with the special syntax which
>    git rebase -i --autosquash
> uses to reorder and combine commits.
>
> That is how I would have prepared the series anyway, but if I stop now
> (before actually running the rebase --autosquash) you can see what I
> have done as individual commits, and it has enabled me to write useful
> commit messages for you.  I have also spared the list another
> patchbomb.
>
> I think the next step is for you to test this branch head.  I have
> checked that it compiles but I have NOT EXECUTED IT AT ALL I am
> afraid.  You can then add your own commits on the end with the same
> autosquash syntax, to fix bugs I have introduced.
>
> And then you will need to actually rebase and squash the series to
> make v21, using a rune like:
>    git rebase -i --autosquash remus-v20
> Alternatively you can do that rebase first.
>
> You will probably find that there are some conflicts but hopefully
> they will be straightforward.
>
> I'll now discuss my actual changes:
>
>
> The two commits `libxl: multidev: ' need to be bubbled to the top of
> your series.  The rest of my changes to `Introduce an abstract Remus
> device layer' depend on them.
>
>
> You can allow git to combine the rest of the changes marked with
> `squash!' with whatever commit they belong to.  You'll proabably want
> to edit the commit message to combine my comments with your own commit
> message so that the whole thing looks sane.
>
> I hope the commit messages I have provided make sense to you.
>
>
> The top commit on my branch, squash-VARIOUS, swaps
> LIBXL__DEVICE_KIND_{VIF,VBD} for LIBXL__DEVICE_REMUS_KIND_{NIC,DISK}.
> Sorry for bundling that all together, but it's a simple
> search-and-replace.  It might be easier to do that to your whole
> series with git-filter-branch.  So I have done that for you here:
>
>   git://xenbits.xen.org/people/iwj/xen.git refs/tags/remus-v20..wip.remus.v20.filtered
>
> You'll still need to bubble that top patch, which is very small in the
> filtered version, down the series.  See below for the
> git-filter-branch rune I used.
>
>
> I hope this is all of some use.
>
> Regards,
> Ian.
>
>
> mariner:xen.git> git-checkout -b wip.remus.v20.filtered
> mariner:xen.git> git-reset --hard wip.remus.v20
> mariner:xen.git> cat ~/work/t
> #!/bin/sh
> perl -i -pe 's/LIBXL__DEVICE_KIND_REMUS_DISK/LIBXL__DEVICE_KIND_VBD/g;
> s/LIBXL__DEVICE_KIND_REMUS_NIC/LIBXL__DEVICE_KIND_VIF/g' tools/libxl/*
> mariner:xen.git> git-filter-branch --tree-filter ~/work/t hongyang/remus-v20..wip.remus.v20.filtered
> [stuff]
> .
>

-- 
Thanks,
Yang.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel