Re: [PATCH wireless-next v4 05/21] wifi: nl80211: define interfaces for SMD BSS Transition context transfer

[Pooventhiran G]

On 11/11/2026 12:15 am, Johannes Berg wrote:
> On Fri, 2026-11-06 at 13:28 +0530, Pooventhiran G wrote:
>>
>> +/**
>> + * enum nl80211_smd_ctx_type - SMD Request Type values
> 
> This is more "ST request type" or something.
> 
> You're conflating "SMD" with "SMD transition" throughout, and the "BSS"
> in the subject isn't really supposed to be there
> 

I will change it to SMD Transition (and ST) and remove "BSS".

>> + * Used with %NL80211_ATTR_SMD_CTX_TYPE.
>> + *
>> + * @NL80211_SMD_CTX_TYPE_PREP: SMD BSS Transition (ST) Preparation.
> 
> (as here)
> 
>> + * @NL80211_SMD_CTX_TYPE_EXEC: ST Execution.
>> + * @NUM_SMD_CTX_TYPES: internal.
>> + * @NL80211_SMD_CTX_TYPE_MAX: Max type of ST context.
>> + */
>> +enum nl80211_smd_ctx_type {
>> +	NL80211_SMD_CTX_TYPE_PREP,
>> +	NL80211_SMD_CTX_TYPE_EXEC,
>> +
>> +	NUM_SMD_CTX_TYPES,
>> +	NL80211_SMD_CTX_TYPE_MAX = NUM_SMD_CTX_TYPES - 1,
>> +};
>> +
>> +/**
>> + * enum nl80211_smd_ctx_ba_param - SMD Block Ack Context attributes
>> + *
>> + * @__NL80211_SMD_CTX_BA_ATTR_INVALID: Invalid attribute.
>> + * @NL80211_SMD_CTX_BA_ATTR_BUFF_SIZE: Required (u16) BA buffer size.
>> + * @NL80211_SMD_CTX_BA_ATTR_POLICY: Required (u8) BA policy.
>> + * @NL80211_SMD_CTX_BA_ATTR_AMSDU_SUPPORT: Flag attribute to indicate
>> + *	AMSDU support.
>> + * @NL80211_SMD_CTX_BA_ATTR_TIMEOUT: Required (u16) Block Ack timeout.
>> + * @NL80211_SMD_CTX_BA_ATTR_EXT_NO_FRAG: Flag attribute to indicate
>> + *	fragmentation support.
>> + * @NL80211_SMD_CTX_BA_ATTR_EXT_FRAG_LEVEL: Required (u8) HE fragmentation
>> + *	level.
>> + * @NL80211_SMD_CTX_BA_ATTR_EXT_BUFF_SIZE: Required (u16) extended buffer size.
> 
> Same here as in the other patch - splitting the buffer size into two
> variables or attributes seems ridiculous.
> 

Sure.

>> + * @NL80211_SMD_CTX_DL_ATTR_VALID_TID_BITMAP: Optional (u8) TIDs included.
> 
> That doesn't really make sense, it just introduces a place where things
> can drift - the per TID nested values below already implicitly say which
> TIDs are valid.
> 

Ouch! Yes, it is not needed in the nl interface. It was originally added for
GET_SMD_CTX but that does not need it anymore too. I will clean this up.

>> +/**
>> + * enum nl80211_smd_ctx_ul - SMD Dynamic Context - UL direction
> 
> Right now I don't see a lot of value in splitting UL/DL attributes when
> they're basically mostly the same - we can probably unify that into one
> space and maybe have a few that are one direction or the other only?
> 

Sure, Johannes. I will combine these two into one namespace with separate
labels for those that differ in type.

> johannes