On 2020-09-29 11:32, Johannes Berg wrote:
On Tue, 2020-09-29 at 11:19 -0700, Thomas Pedersen wrote:
The sband->ht_cap was being processed before S1G Operation
element. Since the HT capability element should not be
present on the S1G band, avoid processing potential
garbage by moving the call to
ieee80211_apply_htcap_overrides() to after the S1G block.
Ah, heh. I hadn't even realized that.
What I meant though was something else: we have
if (s1g_oper && sband->band == NL80211_BAND_S1GHZ) {
...
goto out;
}
// process ht cap overrides (after this patch)
// check HT oper
// check VHT oper
// ...
So given that first condition, if you actually have an S1G AP *without*
S1G operation element (for some reason), you'd start processing HT,
VHT,
and whatever else capabilities, sending us down a misleading and likely
very confusing path, which seems like it should be avoided?
Ah ok, yes the !s1g_oper case. I'll take a look.
--
thomas
