Re: [PATCH net] net/smc: fix mismatch between CLC header and proposal extensions

[Alexandra Winter <wintera@linux.ibm.com>]

On 31.10.25 04:18, D. Wythe wrote:
> The current CLC proposal message construction uses a mix of
> `ini->smc_type_v1/v2` and `pclc_base->hdr.typev1/v2` to decide whether
> to include optional extensions (IPv6 prefix extension for v1, and v2
> extension). This leads to a critical inconsistency: when
> `smc_clc_prfx_set()` fails - for example, in IPv6-only environments with
> only link-local addresses, or when the local IP address and the outgoing
> interface’s network address are not in the same subnet.
> 
> As a result, the proposal message is assembled using the stale
> `ini->smc_type_v1` value—causing the IPv6 prefix extension to be
> included even though the header indicates v1 is not supported.
> The peer then receives a malformed CLC proposal where the header type
> does not match the payload, and immediately resets the connection.
> 
> Fix this by consistently using `pclc_base->hdr.typev1` and
> `pclc_base->hdr.typev2`—the authoritative fields that reflect the
> actual capabilities advertised in the CLC header—when deciding whether
> to include optional extensions, as required by the SMC-R v2
> specification ("V1 IP Subnet Extension and V2 Extension only present if
> applicable").


Just thinking out loud:
It seems to me that the 'ini' structure exists once per socket and is used
to pass information between many functions involved with the handshake.
Did you consider updating ini->smc_type_v1/v2 when `smc_clc_prfx_set()` fails,
and using ini as the authoritative source?
With your patch, it seems to me `ini->smc_type_v1` still contains a stale value,
which may lead to issues in other places or future code.