Re: [PATCH mptcp-next v14 00/25] userspace pm enhancements

[Geliang Tang <geliang.tang@linux.dev>]

Hi Matt,

On Wed, 2023-12-27 at 12:38 +0100, Matthieu Baerts wrote:
> Hi Geliang,
> 
> On 08/12/2023 11:07, Geliang Tang wrote:
> > v14:
> >  - implement flush operation in user space as Mat suggested.
> >  - update selftests.
> >  - Now this series includes five parts:
> > 
> > Part 1: dump for userspace pm (patches 1-9)
> > Part 2: fixes for creating id 0 subflow (patches 10-13)
> > Part 3: v4-mapped addr support (patches 14-15)
> > Part 4: flush for userspace pm (patches 16-18)
> > Part 5: address entry refcount for userspace pm (patches 19-25)
> 
> For tracking purposes, which patches can we drop on Patchwork?
> 
> https://patchwork.kernel.org/project/mptcp/list/?series=808195&state=7
> 
> - Part 1: If I understood correctly, patches 1-9 are supposed to be
> replaced by a new series "dump for userspace pm", right? Then can we
> drop patches 1, 2 and 5 from Patchwork? (Patch 3 has been marked as
> "Rejected", I guess that's OK.)

Patches 1, 2 and 3 can been marked as "Rejected".
Patches 4-9 can been marked as "Changes Requested", they will
been included in the next version.

> 
> - Part 2: does it depend on part 1, or can it be applied separatelly
> in
> -net because they are fixes? Or at least the modifications in the
> code,
> the selftests tests could go only in -next if it is difficult to get
> them in -net.
> 
> - Part 3: Is it also a fix? I mean: did we forget to support v4-
> mapped
> addr in one command, but others have this support? Can it be sent
> separately to -net?

Yes, patches 10, 11, 12, 14 are fixes. They don't depend on part 1 but
will conflict with part 1. I'll make them a new series for -net.

> 
> - Part 4: it is not clear to me if they were needed? They have been
> marked as "Rejected", but please tell me if they are still needed and
> just the title here was wrong.

Patch 18 can be marked as "Changes Requested". We can test flushing
addresses in userspace in it.

Thanks,
-Geliang

> 
> - Part 5: I need to re-check the code before invalidating them (and
> issue #403).
> 
> Cheers,
> Matt