2008-07-23 21:53:42

by [email protected]

[permalink] [raw]
Subject: Re: [PATCH 0/8] rpcbind v4 support in net/sunrpc/svc*

On Mon, Jul 21, 2008 at 03:07:48PM -0400, Chuck Lever wrote:
> On Sun, Jul 20, 2008 at 5:17 PM, J. Bruce Fields <[email protected]> wrote:
> > On Fri, Jul 18, 2008 at 09:07:38PM -0400, J. Bruce Fields wrote:
> >> On Mon, Jun 30, 2008 at 06:45:21PM -0400, Chuck Lever wrote:
> >> > Hi Bruce, Trond -
> >> >
> >> > These implement basic support in net/sunrpc/svc* for registering IPv6-
> >> > enabled kernel RPC services with the local user-space rpcbind. These
> >> > all depend on the new rpcb_v4_register() API I just posted.
> >> >
> >> > The last patch enables IPv6 support for the NFSv4 callback service, even
> >> > though it doesn't register with rpcbind.
> >>
> >> There's some conflicts against my latest tree. Would it be possible to
> >> rebase to
> >>
> >> git://linux-nfs.org/~bfields/linux.git for-2.6.27
> >>
> >> sometime?
> >
> > Actually, never mind, I fixed up the conflicts in the first couple and
> > staged them here:
> >
> > git://linux-nfs.org/~bfields/linux.git for-2.6.28
> >
> > (I think the rest'll still apply cleanly to that.)
> >
> > We could maybe still try to get the rest of this in at the end of the
> > 2.6.27 merge window, but I'm inclined to wait. Apologies, partly my
> > fault for not getting around to these sooner.
>
> Given the complexity and controversy of these changes, the number of
> other changes that went into .27 and will require us to rewrite some
> of these, and the pending holiday/travel for our team of maintainers,
> I think waiting for .28 is our only choice.

They don't seem as controversial as that! But, OK, I guess .28 it is.
Please let me know as soon as there's another set ready.

--b.


2008-07-23 22:47:28

by Chuck Lever III

[permalink] [raw]
Subject: Re: [PATCH 0/8] rpcbind v4 support in net/sunrpc/svc*

On Wed, Jul 23, 2008 at 5:53 PM, J. Bruce Fields <[email protected]> wrote:
> On Mon, Jul 21, 2008 at 03:07:48PM -0400, Chuck Lever wrote:
>> On Sun, Jul 20, 2008 at 5:17 PM, J. Bruce Fields <[email protected]> wrote:
>> > On Fri, Jul 18, 2008 at 09:07:38PM -0400, J. Bruce Fields wrote:
>> >> On Mon, Jun 30, 2008 at 06:45:21PM -0400, Chuck Lever wrote:
>> >> > Hi Bruce, Trond -
>> >> >
>> >> > These implement basic support in net/sunrpc/svc* for registering IPv6-
>> >> > enabled kernel RPC services with the local user-space rpcbind. These
>> >> > all depend on the new rpcb_v4_register() API I just posted.
>> >> >
>> >> > The last patch enables IPv6 support for the NFSv4 callback service, even
>> >> > though it doesn't register with rpcbind.
>> >>
>> >> There's some conflicts against my latest tree. Would it be possible to
>> >> rebase to
>> >>
>> >> git://linux-nfs.org/~bfields/linux.git for-2.6.27
>> >>
>> >> sometime?
>> >
>> > Actually, never mind, I fixed up the conflicts in the first couple and
>> > staged them here:
>> >
>> > git://linux-nfs.org/~bfields/linux.git for-2.6.28
>> >
>> > (I think the rest'll still apply cleanly to that.)
>> >
>> > We could maybe still try to get the rest of this in at the end of the
>> > 2.6.27 merge window, but I'm inclined to wait. Apologies, partly my
>> > fault for not getting around to these sooner.
>>
>> Given the complexity and controversy of these changes, the number of
>> other changes that went into .27 and will require us to rewrite some
>> of these, and the pending holiday/travel for our team of maintainers,
>> I think waiting for .28 is our only choice.
>
> They don't seem as controversial as that! But, OK, I guess .28 it is.
> Please let me know as soon as there's another set ready.

Well, Sankar just informed me that the lock recovery bug we discussed
last week was most likely another problem not related to my IPv6
changes. Bug found on RHEL 5, and fixed there, needs to go
upstream... he may have a patch I can apply to my repo so we can
confirm that NFSv2/v3 lock recovery is not affected by my IPv6-related
changes.

I have mostly finished clean ups as suggested during review, and as
soon as Sankar and I can get a working set up we will confirm all is
well, and I will repost what I have.

Shooting for .28 will give us a chance to get these into linux-next
for broader verification that NFS over IPv4 is unaffected by the IPv6
support.

--
"Alright guard, begin the unnecessarily slow-moving dipping mechanism."
--Dr. Evil