Re: [PATCH mptcp-next v6 3/7] mptcp: pm: improve error messages

[Geliang Tang <geliang@kernel.org>]

Happy New Year Matt!

On Mon, 2025-01-06 at 09:45 +0100, Matthieu Baerts wrote:
> Hi Geliang,
> 
> On 06/01/2025 09:41, Geliang Tang wrote:
> > On Mon, 2024-12-30 at 14:24 +0100, Matthieu Baerts (NGI0) wrote:
> > > Some error messages were:
> > > 
> > >  - too generic: "missing input", "invalid request"
> > > 
> > >  - not precise enough: "limit greater than maximum" but what's
> > > the
> > > max?
> > > 
> > >  - missing: subflow not found, or connect error.
> > > 
> > > This can be easily improved by being more precise, or adding new
> > > error
> > > messages.
> 
> (...)
> 
> > > diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
> > > index
> > > 9c0bec588c498e0aa78acf4018db509abe90cad2..a8f73b082460ba00f7fa998
> > > cf76
> > > 46368f8201e2e 100644
> > > --- a/net/mptcp/pm_userspace.c
> > > +++ b/net/mptcp/pm_userspace.c
> 
> (...)
> 
> > > @@ -634,6 +638,9 @@ int mptcp_userspace_pm_set_flags(struct
> > > sk_buff
> > > *skb, struct genl_info *info)
> > >  	ret = mptcp_pm_nl_mp_prio_send_ack(msk, &loc.addr,
> > > &rem.addr, bkup);
> > >  	release_sock(sk);
> > >  
> > > +	if (ret)
> > > +		GENL_SET_ERR_MSG(info, "subflow not found");
> > 
> > I guess you don't want to use "subflow not found" here. I changed
> > it as
> > "mp_prio send ack failed" in v7.
> 
> No, I wanted to use "subflow not found":
> mptcp_pm_nl_mp_prio_send_ack()
> will fail only if it cannot find the subflow matching the local and
> remote attributes.

My bad, how about using "subflow not found in set_flags" here, and use
"subflow not found in subflow_destroy" for
mptcp_pm_nl_subflow_destroy_doit?

> 
> Cheers,
> Matt