v4:
1. rebased to top commit id 9e263e193af7 kernel/git/jberg/mac80211-next.git
2. add NULL check for "mac80211: use ieee802_11_parse_elems() to find ies instead of ieee80211_bss_get_ie()"
3. remove the 3 patches which already upstream:
"mac80211: parse transmit power envelope element"
"ieee80211: add definition for transmit power envelope element"
"ieee80211: add definition of regulatory info in 6 GHz operation information"
v3: change per comments of Johannes.
1. add patch "mac80211: use ieee802_11_parse_elems() to find ies instead of ieee80211_bss_get_ie()"
2. move nl80211_ap_reg_power to ieee80211
3. change some comments, length check, stack big size variable...
v2: change per comments of johannes.
including code style, code logic, patch merge, commit log...
It introduced some new concept:
power type of AP(STANDARD_POWER_AP, INDOOR_AP, VERY_LOW_POWER_AP)
power type of STATION(DEFAULT_CLIENT, SUBORDINATE_CLIENT)
power spectral density(psd)
This patchset for cfg80211/mac80211 is to add the definition of new
concept of 6G and add basic parse of IE(transmit power envelope
element) in beacon and save power spectral density(psd) reported
by lower-driver for 6G channel, the info will be passed to lower
driver when connecting to 6G AP.
Wen Gong (6):
mac80211: add power type definition for 6 GHz
mac80211: add parse regulatory info in 6 GHz operation information
cfg80211: add definition for 6 GHz power spectral density(psd)
cfg80211: save power spectral density(psd) of regulatory rule
mac80211: use ieee802_11_parse_elems() to find ies instead of
ieee80211_bss_get_ie()
mac80211: save transmit power envelope element and power constraint
include/linux/ieee80211.h | 34 +++++++++++++++
include/net/cfg80211.h | 5 +++
include/net/mac80211.h | 8 ++++
include/net/regulatory.h | 1 +
include/uapi/linux/nl80211.h | 2 +
net/mac80211/mlme.c | 82 +++++++++++++++++++++++++++++-------
net/mac80211/util.c | 13 ++++++
net/wireless/reg.c | 17 ++++++++
8 files changed, 146 insertions(+), 16 deletions(-)
base-commit: 9e263e193af73d2509dc3102a680a11130f44e20
--
2.31.1
The power spectral density(psd) of regulatory rule should be take
effect to the channels. This patch is to save the values to the
channel which has psd value.
Signed-off-by: Wen Gong <[email protected]>
---
net/wireless/reg.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index df87c7f3a049..8f765befb9bc 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -1590,6 +1590,8 @@ static u32 map_regdom_flags(u32 rd_flags)
channel_flags |= IEEE80211_CHAN_NO_160MHZ;
if (rd_flags & NL80211_RRF_NO_HE)
channel_flags |= IEEE80211_CHAN_NO_HE;
+ if (rd_flags & NL80211_RRF_PSD)
+ channel_flags |= IEEE80211_CHAN_PSD;
return channel_flags;
}
@@ -1794,6 +1796,9 @@ static void handle_channel_single_rule(struct wiphy *wiphy,
chan->dfs_cac_ms = reg_rule->dfs_cac_ms;
}
+ if (chan->flags & IEEE80211_CHAN_PSD)
+ chan->psd = reg_rule->psd;
+
return;
}
@@ -1814,6 +1819,9 @@ static void handle_channel_single_rule(struct wiphy *wiphy,
chan->dfs_cac_ms = IEEE80211_DFS_MIN_CAC_TIME_MS;
}
+ if (chan->flags & IEEE80211_CHAN_PSD)
+ chan->psd = reg_rule->psd;
+
if (chan->orig_mpwr) {
/*
* Devices that use REGULATORY_COUNTRY_IE_FOLLOW_POWER
@@ -1883,6 +1891,12 @@ static void handle_channel_adjacent_rules(struct wiphy *wiphy,
rrule2->dfs_cac_ms);
}
+ if ((rrule1->flags & NL80211_RRF_PSD) &&
+ (rrule2->flags & NL80211_RRF_PSD))
+ chan->psd = min_t(s8, rrule1->psd, rrule2->psd);
+ else
+ chan->flags &= ~NL80211_RRF_PSD;
+
return;
}
@@ -2540,6 +2554,9 @@ static void handle_channel_custom(struct wiphy *wiphy,
chan->dfs_cac_ms = IEEE80211_DFS_MIN_CAC_TIME_MS;
}
+ if (chan->flags & IEEE80211_CHAN_PSD)
+ chan->psd = reg_rule->psd;
+
chan->max_power = chan->max_reg_power;
}
--
2.31.1
This is to save the transmit power envelope element and power
constraint in struct ieee80211_bss_conf for 6 GHz. Lower driver
will use this info to calculate the power limit.
Signed-off-by: Wen Gong <[email protected]>
---
include/net/mac80211.h | 6 ++++++
net/mac80211/mlme.c | 30 ++++++++++++++++++++++++++++++
2 files changed, 36 insertions(+)
diff --git a/include/net/mac80211.h b/include/net/mac80211.h
index 8923a9fc4126..16a965262a4f 100644
--- a/include/net/mac80211.h
+++ b/include/net/mac80211.h
@@ -633,6 +633,9 @@ struct ieee80211_fils_discovery {
* @beacon_tx_rate: The configured beacon transmit rate that needs to be passed
* to driver when rate control is offloaded to firmware.
* @power_type: power type of BSS for 6 GHz
+ * @tx_pwr_env: transmit power envelope array of BSS.
+ * @tx_pwr_env_num: number of @tx_pwr_env.
+ * @pwr_reduction: power constraint of BSS.
*/
struct ieee80211_bss_conf {
const u8 *bssid;
@@ -704,6 +707,9 @@ struct ieee80211_bss_conf {
bool s1g;
struct cfg80211_bitrate_mask beacon_tx_rate;
enum ieee80211_ap_reg_power power_type;
+ struct ieee80211_tx_pwr_env tx_pwr_env[IEEE80211_TPE_MAX_IE_COUNT];
+ u8 tx_pwr_env_num;
+ u8 pwr_reduction;
};
/**
diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
index 6e8e90ee8b6c..b72fa1f6116a 100644
--- a/net/mac80211/mlme.c
+++ b/net/mac80211/mlme.c
@@ -2258,6 +2258,7 @@ static void ieee80211_set_disassoc(struct ieee80211_sub_if_data *sdata,
{
struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
struct ieee80211_local *local = sdata->local;
+ struct ieee80211_bss_conf *bss_conf;
u32 changed = 0;
struct ieee80211_prep_tx_info info = {
.subtype = stype,
@@ -2407,6 +2408,11 @@ static void ieee80211_set_disassoc(struct ieee80211_sub_if_data *sdata,
cancel_delayed_work_sync(&ifmgd->tx_tspec_wk);
sdata->encrypt_headroom = IEEE80211_ENCRYPT_HEADROOM;
+
+ bss_conf = &sdata->vif.bss_conf;
+ bss_conf->pwr_reduction = 0;
+ bss_conf->tx_pwr_env_num = 0;
+ memset(bss_conf->tx_pwr_env, 0, sizeof(bss_conf->tx_pwr_env));
}
static void ieee80211_reset_ap_probe(struct ieee80211_sub_if_data *sdata)
@@ -5095,6 +5101,30 @@ static int ieee80211_prep_channel(struct ieee80211_sub_if_data *sdata,
else
he_oper = NULL;
+ if (is_6ghz) {
+ struct ieee80211_bss_conf *bss_conf;
+ u8 i, j = 0;
+
+ bss_conf = &sdata->vif.bss_conf;
+
+ if (elems->pwr_constr_elem)
+ bss_conf->pwr_reduction = *elems->pwr_constr_elem;
+
+ BUILD_BUG_ON(ARRAY_SIZE(bss_conf->tx_pwr_env) !=
+ ARRAY_SIZE(elems->tx_pwr_env));
+
+ for (i = 0; i < elems->tx_pwr_env_num; i++) {
+ if (elems->tx_pwr_env_len[i] >
+ sizeof(bss_conf->tx_pwr_env[j]))
+ continue;
+
+ bss_conf->tx_pwr_env_num++;
+ memcpy(&bss_conf->tx_pwr_env[j], elems->tx_pwr_env[i],
+ elems->tx_pwr_env_len[i]);
+ j++;
+ }
+ }
+
if (!ieee80211_verify_sta_he_mcs_support(sdata, sband, he_oper))
ifmgd->flags |= IEEE80211_STA_DISABLE_HE;
}
--
2.31.1
6 GHz regulatory domains introduces power spectral density(psd). This
patch is define the flags for psd.
Signed-off-by: Wen Gong <[email protected]>
---
include/net/cfg80211.h | 5 +++++
include/net/regulatory.h | 1 +
include/uapi/linux/nl80211.h | 2 ++
3 files changed, 8 insertions(+)
diff --git a/include/net/cfg80211.h b/include/net/cfg80211.h
index 62dd8422e0dc..f8e0cc19e0ce 100644
--- a/include/net/cfg80211.h
+++ b/include/net/cfg80211.h
@@ -109,6 +109,8 @@ struct wiphy;
* on this channel.
* @IEEE80211_CHAN_16MHZ: 16 MHz bandwidth is permitted
* on this channel.
+ * @IEEE80211_CHAN_PSD: power spectral density (in dBm)
+ * on this channel.
*
*/
enum ieee80211_channel_flags {
@@ -131,6 +133,7 @@ enum ieee80211_channel_flags {
IEEE80211_CHAN_4MHZ = 1<<16,
IEEE80211_CHAN_8MHZ = 1<<17,
IEEE80211_CHAN_16MHZ = 1<<18,
+ IEEE80211_CHAN_PSD = 1<<19,
};
#define IEEE80211_CHAN_NO_HT40 \
@@ -164,6 +167,7 @@ enum ieee80211_channel_flags {
* on this channel.
* @dfs_state_entered: timestamp (jiffies) when the dfs state was entered.
* @dfs_cac_ms: DFS CAC time in milliseconds, this is valid for DFS channels.
+ * @psd: power spectral density (in dBm)
*/
struct ieee80211_channel {
enum nl80211_band band;
@@ -180,6 +184,7 @@ struct ieee80211_channel {
enum nl80211_dfs_state dfs_state;
unsigned long dfs_state_entered;
unsigned int dfs_cac_ms;
+ s8 psd;
};
/**
diff --git a/include/net/regulatory.h b/include/net/regulatory.h
index 47f06f6f5a67..ed20004fb6a9 100644
--- a/include/net/regulatory.h
+++ b/include/net/regulatory.h
@@ -221,6 +221,7 @@ struct ieee80211_reg_rule {
u32 flags;
u32 dfs_cac_ms;
bool has_wmm;
+ s8 psd;
};
struct ieee80211_regdomain {
diff --git a/include/uapi/linux/nl80211.h b/include/uapi/linux/nl80211.h
index c2efea98e060..2ca8fa06f536 100644
--- a/include/uapi/linux/nl80211.h
+++ b/include/uapi/linux/nl80211.h
@@ -4085,6 +4085,7 @@ enum nl80211_sched_scan_match_attr {
* @NL80211_RRF_NO_80MHZ: 80MHz operation not allowed
* @NL80211_RRF_NO_160MHZ: 160MHz operation not allowed
* @NL80211_RRF_NO_HE: HE operation not allowed
+ * @NL80211_RRF_PSD: channels has power spectral density value
*/
enum nl80211_reg_rule_flags {
NL80211_RRF_NO_OFDM = 1<<0,
@@ -4103,6 +4104,7 @@ enum nl80211_reg_rule_flags {
NL80211_RRF_NO_80MHZ = 1<<15,
NL80211_RRF_NO_160MHZ = 1<<16,
NL80211_RRF_NO_HE = 1<<17,
+ NL80211_RRF_PSD = 1<<18,
};
#define NL80211_RRF_PASSIVE_SCAN NL80211_RRF_NO_IR
--
2.31.1
On Fri, 2021-09-24 at 06:00 -0400, Wen Gong wrote:
> The power spectral density(psd) of regulatory rule should be take
> effect to the channels. This patch is to save the values to the
> channel which has psd value.
>
> Signed-off-by: Wen Gong <[email protected]>
> ---
> net/wireless/reg.c | 17 +++++++++++++++++
> 1 file changed, 17 insertions(+)
>
> diff --git a/net/wireless/reg.c b/net/wireless/reg.c
> index df87c7f3a049..8f765befb9bc 100644
> --- a/net/wireless/reg.c
> +++ b/net/wireless/reg.c
> @@ -1590,6 +1590,8 @@ static u32 map_regdom_flags(u32 rd_flags)
> channel_flags |= IEEE80211_CHAN_NO_160MHZ;
> if (rd_flags & NL80211_RRF_NO_HE)
> channel_flags |= IEEE80211_CHAN_NO_HE;
> + if (rd_flags & NL80211_RRF_PSD)
> + channel_flags |= IEEE80211_CHAN_PSD;
I went to go squash this with patch 3 and took a closer look, and then I
realized you're doing this weird.
Please when you resend also squash this - it's a bit weird to read in
two patches.
However, I think this is missing a lot of things - we already talked
about the regulatory database, and while that'd be nice, I guess I
conceded that you don't really have to do it now.
However, for visibility reasons, I *really* think you need to add
nl80211 attributes for all of this data - when the regdomain is dumped
in nl80211_put_regdom() you would have the flag now (it dumps the value
of reg_rule->flags in NL80211_ATTR_REG_RULE_FLAGS), but you didn't add
the value of reg_rule->psd which you've added.
Similarly, you're not adding the PSD flag nor the PSD value for the
*channel* in nl80211_msg_put_channel(), both of which I think you should
have for visibility into what's going on in the kernel/driver.
I've applied all the other patches, so please just resend 3 and 4,
squashed into a single patch, with the fixes.
johannes
On Fri, 2021-09-24 at 06:00 -0400, Wen Gong wrote:
>
> NL80211_RRF_NO_80MHZ = 1<<15,
> NL80211_RRF_NO_160MHZ = 1<<16,
> NL80211_RRF_NO_HE = 1<<17,
> + NL80211_RRF_PSD = 1<<18,
Indentation seems to be missing a tab here.
johannes
On 2021-09-27 19:27, Johannes Berg wrote:
> On Fri, 2021-09-24 at 06:00 -0400, Wen Gong wrote:
>> The power spectral density(psd) of regulatory rule should be take
>> effect to the channels. This patch is to save the values to the
>> channel which has psd value.
>>
>> Signed-off-by: Wen Gong <[email protected]>
>> ---
>> net/wireless/reg.c | 17 +++++++++++++++++
>> 1 file changed, 17 insertions(+)
>>
>> diff --git a/net/wireless/reg.c b/net/wireless/reg.c
>> index df87c7f3a049..8f765befb9bc 100644
>> --- a/net/wireless/reg.c
>> +++ b/net/wireless/reg.c
>> @@ -1590,6 +1590,8 @@ static u32 map_regdom_flags(u32 rd_flags)
>> channel_flags |= IEEE80211_CHAN_NO_160MHZ;
>> if (rd_flags & NL80211_RRF_NO_HE)
>> channel_flags |= IEEE80211_CHAN_NO_HE;
>> + if (rd_flags & NL80211_RRF_PSD)
>> + channel_flags |= IEEE80211_CHAN_PSD;
>
> I went to go squash this with patch 3 and took a closer look, and then
> I
> realized you're doing this weird.
>
> Please when you resend also squash this - it's a bit weird to read in
> two patches.
>
> However, I think this is missing a lot of things - we already talked
> about the regulatory database, and while that'd be nice, I guess I
> conceded that you don't really have to do it now.
>
> However, for visibility reasons, I *really* think you need to add
> nl80211 attributes for all of this data - when the regdomain is dumped
> in nl80211_put_regdom() you would have the flag now (it dumps the value
> of reg_rule->flags in NL80211_ATTR_REG_RULE_FLAGS), but you didn't add
> the value of reg_rule->psd which you've added.
>
> Similarly, you're not adding the PSD flag nor the PSD value for the
> *channel* in nl80211_msg_put_channel(), both of which I think you
> should
> have for visibility into what's going on in the kernel/driver.
>
> I've applied all the other patches, so please just resend 3 and 4,
> squashed into a single patch, with the fixes.
>
Thanks
I have sent "[PATCH v5] cfg80211: save power spectral density(psd) of
regulatory rule"
for this.
> johannes