[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGXJAmwr5u7Don-nKivvTuj6C9pGT2A5cAXvhKo6sAm_hHy72A@mail.gmail.com>
Date: Tue, 19 Nov 2024 13:13:36 -0800
From: John Ousterhout <ouster@...stanford.edu>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: netdev@...r.kernel.org, linux-api@...r.kernel.org
Subject: Re: [PATCH net-next v2 00/12] Begin upstreaming Homa transport protocol
On Thu, Nov 14, 2024 at 8:59 AM John Ousterhout <ouster@...stanford.edu> wrote:
>
> On Wed, Nov 13, 2024 at 9:08 AM Cong Wang <xiyou.wangcong@...il.com> wrote:
>
> > 2. Please consider adding socket diagnostics, see net/ipv4/inet_diag.c.
>
> I wasn't familiar with them before your email; I'll take a look.
I have taken a look at socket diagnostics, and I agree that Homa
should support them. However, this looks like a nontrivial task: there
appears to be a fair amount of complexity there, not much
documentation, and the current diagnostics appear pretty TCP-specific
so it may take some work (and advice) to figure out how to map them
onto Homa. Thus it feels like it will take a fair amount of
back-and-forth to get this right. I would prefer to defer this until
after the main body of Homa has been upstreamed, then have a patch
series that is dedicated to socket diagnostics support, rather than
lumping socket diagnostics into the current patch series. Is that OK
with you?
-John-
Powered by blists - more mailing lists