On 1/30/24 15:47, Johannes Berg wrote:
On Tue, 2024-01-30 at 10:02 +0530, Aditya Kumar Singh wrote:
In order to support CSA with MLO, there is a need to handle the functions
ieee80211_set_csa_beacon() and ieee80211_set_after_csa_beacon() on per
link basis.
nit: "on a per link"
Will address in next version. Thanks.
Add changes for the same.
Is that some cultural thing?
I always find this phrasing with "for the same" very odd, and would
rather say something useful such as "Implement this by passing the
correct link data"... but I see this a lot, hence the question.
No idea :). Even I have seen these quite a few times and thought that
may be it is fine to use it that way. But I do agree that instead we
could put something useful instead. Thanks for pointing it out, I will
address this in next version.
@@ -3658,7 +3659,7 @@ static int __ieee80211_csa_finalize(struct ieee80211_link_data *link_data)
sdata->vif.bss_conf.csa_active = false;
- err = ieee80211_set_after_csa_beacon(sdata, &changed);
+ err = ieee80211_set_after_csa_beacon(&sdata->deflink, &changed);
weren't you just saying deflink shouldn't be used?
Correct. This patch's aim is to form the base - basically modify the
helper function to accept the link data argument. But at this point, CSA
is not started on per link basis hence in order to keep CSA working
still as it is before this patch, have used deflink here. Functionality
wise, this patch is not bringing any change yet.
@@ -3928,7 +3930,7 @@ __ieee80211_channel_switch(struct wiphy *wiphy, struct net_device *dev,
if (sdata->vif.bss_conf.color_change_active)
ieee80211_color_change_abort(sdata);
- err = ieee80211_set_csa_beacon(sdata, params, &changed);
+ err = ieee80211_set_csa_beacon(&sdata->deflink, params, &changed);
dito
Addressed above.
