[PATCH 1/1] io_uring: move zcrx into a separate branch

[PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

Keep zcrx next changes in a separate branch. It was more productive this
way past month and will simplify the workflow for already lined up
changes requiring cross tree patches, specifically netdev. The current
changes can still target the generic io_uring tree as there are no
strong reasons to keep it separate. It'll also be using the io_uring
mailing list.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 MAINTAINERS | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index fe168477caa4..d9e4ab173000 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12870,6 +12870,14 @@ F:	include/uapi/linux/io_uring.h
 F:	include/uapi/linux/io_uring/
 F:	io_uring/
 
+IO_URING ZCRX
+M:	Pavel Begunkov <asml.silence@gmail.com>
+L:	io-uring@vger.kernel.org
+T:	git https://github.com/isilence/linux.git zcrx/for-next
+T:	git git://git.kernel.dk/linux-block
+S:	Maintained
+F:	io_uring/zcrx.*
+
 IPMI SUBSYSTEM
 M:	Corey Minyard <corey@minyard.net>
 L:	openipmi-developer@lists.sourceforge.net (moderated for non-subscribers)
-- 
2.49.0

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On 8/17/25 4:43 PM, Pavel Begunkov wrote:
> Keep zcrx next changes in a separate branch. It was more productive this
> way past month and will simplify the workflow for already lined up
> changes requiring cross tree patches, specifically netdev. The current
> changes can still target the generic io_uring tree as there are no
> strong reasons to keep it separate. It'll also be using the io_uring
> mailing list.

I'm fine with this, as long as it doesn't bifurcate the management of
the overall/main branch - eg patches will flow from the zcrx branch into
my main branch, always. I've got many years of experience of managing
downstream branches with upstream trees, and whenever there are
dependencies, it's a pain in the butt. Don't want to add to that pain.

-- 
Jens Axboe

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 8/18/25 16:24, Jens Axboe wrote:
> On 8/17/25 4:43 PM, Pavel Begunkov wrote:
>> Keep zcrx next changes in a separate branch. It was more productive this
>> way past month and will simplify the workflow for already lined up
>> changes requiring cross tree patches, specifically netdev. The current
>> changes can still target the generic io_uring tree as there are no
>> strong reasons to keep it separate. It'll also be using the io_uring
>> mailing list.
> 
> I'm fine with this, as long as it doesn't bifurcate the management of
> the overall/main branch - eg patches will flow from the zcrx branch into
> my main branch, always. I've got many years of experience of managing
> downstream branches with upstream trees, and whenever there are
> dependencies, it's a pain in the butt. Don't want to add to that pain.

That's the assumed workflow. Apart from conflicts I don't want to
deal with, it'll likely be inconsistent and low volume; doesn't
make sense sending it to Linus directly for many reasons.

-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
> Keep zcrx next changes in a separate branch. It was more productive this
> way past month and will simplify the workflow for already lined up
> changes requiring cross tree patches, specifically netdev. The current
> changes can still target the generic io_uring tree as there are no
> strong reasons to keep it separate. It'll also be using the io_uring
> mailing list.
> 
> [...]

Applied, thanks!

[1/1] io_uring: move zcrx into a separate branch
      commit: 6d136abdd1bdfce0c7108c8e0af6fd95e3d353ad

Best regards,
-- 
Jens Axboe

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 8/20/25 19:20, Jens Axboe wrote:
> 
> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>> Keep zcrx next changes in a separate branch. It was more productive this
>> way past month and will simplify the workflow for already lined up
>> changes requiring cross tree patches, specifically netdev. The current
>> changes can still target the generic io_uring tree as there are no
>> strong reasons to keep it separate. It'll also be using the io_uring
>> mailing list.
>>
>> [...]
> 
> Applied, thanks!

Did it get dropped in the end? For some reason I can't find it.

> 
> [1/1] io_uring: move zcrx into a separate branch
>        commit: 6d136abdd1bdfce0c7108c8e0af6fd95e3d353ad
> 
> Best regards,

-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On 10/17/25 6:37 AM, Pavel Begunkov wrote:
> On 8/20/25 19:20, Jens Axboe wrote:
>>
>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>> Keep zcrx next changes in a separate branch. It was more productive this
>>> way past month and will simplify the workflow for already lined up
>>> changes requiring cross tree patches, specifically netdev. The current
>>> changes can still target the generic io_uring tree as there are no
>>> strong reasons to keep it separate. It'll also be using the io_uring
>>> mailing list.
>>>
>>> [...]
>>
>> Applied, thanks!
> 
> Did it get dropped in the end? For some reason I can't find it.

A bit hazy, but I probably did with the discussions on the netdev side
too as they were ongoing.

-- 
Jens Axboe

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 10/20/25 18:07, Jens Axboe wrote:
> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>> On 8/20/25 19:20, Jens Axboe wrote:
>>>
>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>> way past month and will simplify the workflow for already lined up
>>>> changes requiring cross tree patches, specifically netdev. The current
>>>> changes can still target the generic io_uring tree as there are no
>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>> mailing list.
>>>>
>>>> [...]
>>>
>>> Applied, thanks!
>>
>> Did it get dropped in the end? For some reason I can't find it.
> 
> A bit hazy, but I probably did with the discussions on the netdev side
> too as they were ongoing.

The ones where my work is maliciously blocked with a good email
trace to prove that? How is that relevant though?

-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On 10/20/25 11:41 AM, Pavel Begunkov wrote:
> On 10/20/25 18:07, Jens Axboe wrote:
>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>
>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>> way past month and will simplify the workflow for already lined up
>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>> changes can still target the generic io_uring tree as there are no
>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>> mailing list.
>>>>>
>>>>> [...]
>>>>
>>>> Applied, thanks!
>>>
>>> Did it get dropped in the end? For some reason I can't find it.
>>
>> A bit hazy, but I probably did with the discussions on the netdev side
>> too as they were ongoing.
> 
> The ones where my work is maliciously blocked with a good email
> trace to prove that? How is that relevant though?

I have no horse in that game so don't know which thread(s) that is (nor
does it sound like I need to know), I just recall Mina and/or someone
else having patches for this too. Hence I dropped it to get everyone
come to an agreement on what the appropriate entry would be.

FWIW, I don't think there's much point to listing a separate branch.
It's all supposed to go upstream to the main tree sooner rather than
later, once it's ready. And since it's basically just you working on the
zcrx bits, there's no risk of conflicts. If there was a conflict, then
yeah we'd just resolve it and route things appropriately. But it also
doesn't really matter to me what's listed in there, as long as things
are on the list as well. And iirc the netdev side was the same, they
just want to ensure they see patches for zcrx. For me, the entry was
more about ensuring you get CC'ed on relevant patches.

Which is why I figured that you and the netdev side would discuss this
and come up with an entry that everybody was happy with, then we can get
that upstream.

-- 
Jens Axboe

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 10/20/25 19:01, Jens Axboe wrote:
> On 10/20/25 11:41 AM, Pavel Begunkov wrote:
>> On 10/20/25 18:07, Jens Axboe wrote:
>>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>>
>>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>>> way past month and will simplify the workflow for already lined up
>>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>>> changes can still target the generic io_uring tree as there are no
>>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>>> mailing list.
>>>>>>
>>>>>> [...]
>>>>>
>>>>> Applied, thanks!
>>>>
>>>> Did it get dropped in the end? For some reason I can't find it.
>>>
>>> A bit hazy, but I probably did with the discussions on the netdev side
>>> too as they were ongoing.
>>
>> The ones where my work is maliciously blocked with a good email
>> trace to prove that? How is that relevant though?
> 
> I have no horse in that game so don't know which thread(s) that is (nor
> does it sound like I need to know), I just recall Mina and/or someone
> else having patches for this too. Hence I dropped it to get everyone
> come to an agreement on what the appropriate entry would be.
> 
> FWIW, I don't think there's much point to listing a separate branch.

I sent this patch because last cycle I was waiting for roughly a
month for zcrx to be merged, and hence I started managing a branch
anyway, which also turned out to be simpler and more convenient for
me than the usual workflow. Not blaming anyone, but that's how it went.
And there were a couple of (trivial) patches from folks.

> It's all supposed to go upstream to the main tree sooner rather than
> later, once it's ready. And since it's basically just you working on the
> zcrx bits, there's no risk of conflicts. If there was a conflict, then
> yeah we'd just resolve it and route things appropriately. But it also
> doesn't really matter to me what's listed in there, as long as things
> are on the list as well. And iirc the netdev side was the same, they
> just want to ensure they see patches for zcrx. For me, the entry was

Taking everything into account, I have no doubts what started the
discussion was purely about blocking my work, but I'll spare the
details from this thread and since you said you don't need to know.

> more about ensuring you get CC'ed on relevant patches.
> 
> Which is why I figured that you and the netdev side would discuss this
> and come up with an entry that everybody was happy with, then we can get
> that upstream.
-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 10/20/25 19:34, Pavel Begunkov wrote:
> On 10/20/25 19:01, Jens Axboe wrote:
>> On 10/20/25 11:41 AM, Pavel Begunkov wrote:
>>> On 10/20/25 18:07, Jens Axboe wrote:
>>>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>>>
>>>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>>>> way past month and will simplify the workflow for already lined up
>>>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>>>> changes can still target the generic io_uring tree as there are no
>>>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>>>> mailing list.
>>>>>>>
>>>>>>> [...]
>>>>>>
>>>>>> Applied, thanks!
>>>>>
>>>>> Did it get dropped in the end? For some reason I can't find it.
>>>>
>>>> A bit hazy, but I probably did with the discussions on the netdev side
>>>> too as they were ongoing.
>>>
>>> The ones where my work is maliciously blocked with a good email
>>> trace to prove that? How is that relevant though?
>>
>> I have no horse in that game so don't know which thread(s) that is (nor
>> does it sound like I need to know), I just recall Mina and/or someone
>> else having patches for this too. Hence I dropped it to get everyone
>> come to an agreement on what the appropriate entry would be.
>>
>> FWIW, I don't think there's much point to listing a separate branch.
> 
> I sent this patch because last cycle I was waiting for roughly a
> month for zcrx to be merged, and hence I started managing a branch
> anyway, which also turned out to be simpler and more convenient for
> me than the usual workflow. Not blaming anyone, but that's how it went.
> And there were a couple of (trivial) patches from folks.

To get some clarity, are you going to pick it up?

-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On 10/21/25 7:44 AM, Pavel Begunkov wrote:
> On 10/20/25 19:34, Pavel Begunkov wrote:
>> On 10/20/25 19:01, Jens Axboe wrote:
>>> On 10/20/25 11:41 AM, Pavel Begunkov wrote:
>>>> On 10/20/25 18:07, Jens Axboe wrote:
>>>>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>>>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>>>>
>>>>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>>>>> way past month and will simplify the workflow for already lined up
>>>>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>>>>> changes can still target the generic io_uring tree as there are no
>>>>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>>>>> mailing list.
>>>>>>>>
>>>>>>>> [...]
>>>>>>>
>>>>>>> Applied, thanks!
>>>>>>
>>>>>> Did it get dropped in the end? For some reason I can't find it.
>>>>>
>>>>> A bit hazy, but I probably did with the discussions on the netdev side
>>>>> too as they were ongoing.
>>>>
>>>> The ones where my work is maliciously blocked with a good email
>>>> trace to prove that? How is that relevant though?
>>>
>>> I have no horse in that game so don't know which thread(s) that is (nor
>>> does it sound like I need to know), I just recall Mina and/or someone
>>> else having patches for this too. Hence I dropped it to get everyone
>>> come to an agreement on what the appropriate entry would be.
>>>
>>> FWIW, I don't think there's much point to listing a separate branch.
>>
>> I sent this patch because last cycle I was waiting for roughly a
>> month for zcrx to be merged, and hence I started managing a branch
>> anyway, which also turned out to be simpler and more convenient for
>> me than the usual workflow. Not blaming anyone, but that's how it went.
>> And there were a couple of (trivial) patches from folks.
> 
> To get some clarity, are you going to pick it up?

Wasn't planning on it, please work with Jakub/Mina/netdev crew for an
entry that everybody is happy with.

-- 
Jens Axboe

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Pavel Begunkov]

On 10/21/25 16:42, Jens Axboe wrote:
> On 10/21/25 7:44 AM, Pavel Begunkov wrote:
>> On 10/20/25 19:34, Pavel Begunkov wrote:
>>> On 10/20/25 19:01, Jens Axboe wrote:
>>>> On 10/20/25 11:41 AM, Pavel Begunkov wrote:
>>>>> On 10/20/25 18:07, Jens Axboe wrote:
>>>>>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>>>>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>>>>>
>>>>>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>>>>>> way past month and will simplify the workflow for already lined up
>>>>>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>>>>>> changes can still target the generic io_uring tree as there are no
>>>>>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>>>>>> mailing list.
>>>>>>>>>
>>>>>>>>> [...]
>>>>>>>>
>>>>>>>> Applied, thanks!
>>>>>>>
>>>>>>> Did it get dropped in the end? For some reason I can't find it.
>>>>>>
>>>>>> A bit hazy, but I probably did with the discussions on the netdev side
>>>>>> too as they were ongoing.
>>>>>
>>>>> The ones where my work is maliciously blocked with a good email
>>>>> trace to prove that? How is that relevant though?
>>>>
>>>> I have no horse in that game so don't know which thread(s) that is (nor
>>>> does it sound like I need to know), I just recall Mina and/or someone
>>>> else having patches for this too. Hence I dropped it to get everyone
>>>> come to an agreement on what the appropriate entry would be.
>>>>
>>>> FWIW, I don't think there's much point to listing a separate branch.
>>>
>>> I sent this patch because last cycle I was waiting for roughly a
>>> month for zcrx to be merged, and hence I started managing a branch
>>> anyway, which also turned out to be simpler and more convenient for
>>> me than the usual workflow. Not blaming anyone, but that's how it went.
>>> And there were a couple of (trivial) patches from folks.
>>
>> To get some clarity, are you going to pick it up?
> 
> Wasn't planning on it, please work with Jakub/Mina/netdev crew for an
> entry that everybody is happy with.

Not really a crew, only Jakub has a problem. And I still don't see
how it's relevant to the project written by me and David with zero
contribution from Jakub apart from endless delays. But I can
understand why you'd be standing for a fellow maintainer, even if
he's abusing the maintainership, while screwing and impeding my
work on zcrx.

-- 
Pavel Begunkov

Re: [PATCH 1/1] io_uring: move zcrx into a separate branch

[Jens Axboe]

On 10/21/25 12:45 PM, Pavel Begunkov wrote:
> On 10/21/25 16:42, Jens Axboe wrote:
>> On 10/21/25 7:44 AM, Pavel Begunkov wrote:
>>> On 10/20/25 19:34, Pavel Begunkov wrote:
>>>> On 10/20/25 19:01, Jens Axboe wrote:
>>>>> On 10/20/25 11:41 AM, Pavel Begunkov wrote:
>>>>>> On 10/20/25 18:07, Jens Axboe wrote:
>>>>>>> On 10/17/25 6:37 AM, Pavel Begunkov wrote:
>>>>>>>> On 8/20/25 19:20, Jens Axboe wrote:
>>>>>>>>>
>>>>>>>>> On Sun, 17 Aug 2025 23:43:14 +0100, Pavel Begunkov wrote:
>>>>>>>>>> Keep zcrx next changes in a separate branch. It was more productive this
>>>>>>>>>> way past month and will simplify the workflow for already lined up
>>>>>>>>>> changes requiring cross tree patches, specifically netdev. The current
>>>>>>>>>> changes can still target the generic io_uring tree as there are no
>>>>>>>>>> strong reasons to keep it separate. It'll also be using the io_uring
>>>>>>>>>> mailing list.
>>>>>>>>>>
>>>>>>>>>> [...]
>>>>>>>>>
>>>>>>>>> Applied, thanks!
>>>>>>>>
>>>>>>>> Did it get dropped in the end? For some reason I can't find it.
>>>>>>>
>>>>>>> A bit hazy, but I probably did with the discussions on the netdev side
>>>>>>> too as they were ongoing.
>>>>>>
>>>>>> The ones where my work is maliciously blocked with a good email
>>>>>> trace to prove that? How is that relevant though?
>>>>>
>>>>> I have no horse in that game so don't know which thread(s) that is (nor
>>>>> does it sound like I need to know), I just recall Mina and/or someone
>>>>> else having patches for this too. Hence I dropped it to get everyone
>>>>> come to an agreement on what the appropriate entry would be.
>>>>>
>>>>> FWIW, I don't think there's much point to listing a separate branch.
>>>>
>>>> I sent this patch because last cycle I was waiting for roughly a
>>>> month for zcrx to be merged, and hence I started managing a branch
>>>> anyway, which also turned out to be simpler and more convenient for
>>>> me than the usual workflow. Not blaming anyone, but that's how it went.
>>>> And there were a couple of (trivial) patches from folks.
>>>
>>> To get some clarity, are you going to pick it up?
>>
>> Wasn't planning on it, please work with Jakub/Mina/netdev crew for an
>> entry that everybody is happy with.
> 
> Not really a crew, only Jakub has a problem. And I still don't see
> how it's relevant to the project written by me and David with zero
> contribution from Jakub apart from endless delays. But I can
> understand why you'd be standing for a fellow maintainer, even if
> he's abusing the maintainership, while screwing and impeding my
> work on zcrx.

Let's try and keep it cordial...

Just send out an updated patch that fits the criteria to the best of
your understanding, with the folks CC'ed. I'll be happy to apply it,
even if it needs some massaging.

-- 
Jens Axboe