[openfabrics-ewg] ipath patches for RC3

[Michael S. Tsirkin]

Quoting r. Bryan O'Sullivan <bos at pathscale.com>:
> Subject: Re: [openfabrics-ewg] ipath patches for RC3
> 
> On Wed, 2006-08-30 at 18:26 +0300, Tziporet Koren wrote:
> 
> > Michael and me tried to apply the rest of your patches but they won't apply.
> 
> Yeah, I have a suspicion that the dropping of the first three patches
> will make applying the rest a pain.
> 
> > Note also that the patches you send where not named correctly:
> > 1. Please start all your patches with ipath-
> > 2. Since patches are applied in alphabetic order, please make sure that 
> > the name of the patches will be in the order that they need to be 
> > applied. (we added numbers according to the order in the series file).
> 
> OK.
> 
> > Please open the tgz, add the patches (except the 5 above) to the 
> > appropriate directories and see that everything can applied & compiles. 
> 
> Fixing the patches up may end up taking a lot of work.  I'll do what I
> can.  Thanks for the help.

So what I did was git log master..for-2.6.19
and then I git-cherry-picked ipath patches from Roland's for-2.6.19 tree.

It was actually almost fully automatic: only 3 commits had trivial conflicts:
f031cd51d1c4f08320a8d2c82f8bf7d90d436d37
a3b8d106f3e66a748d24d7dd0471c7cb9885c727
ff7de577cc7ec42651a9c94c57fefe8f58edd8a9

That's out of 22. Check the result out from

git://www.mellanox.co.il/~git/infiniband ipath_update

Bryan, now if you can
1. test this at least minimally (I didn't even compile)
2. provide backport patches against this
3. put backport patches prefixes with ipath and numbered
   approrpiately in appropriate backport directories
4. and post the tarball with backports

Then that's what we'll have for RC3: I'll convert git branch
to patch series in fixes, and add the backports that you send.

Please note we don't work Friday so must finish all changes
we want in RC3, today.

-- 
MST