diff mbox series

[v10,2/2] wifi: mwifiex: add host mlme for AP mode

Message ID 20240418060626.431202-3-yu-hao.lin@nxp.com
State New
Headers show
Series wifi: mwifiex: add code to support host mlme | expand

Commit Message

David Lin April 18, 2024, 6:06 a.m. UTC
Add host based MLME to enable WPA3 functionalities in AP mode.
This feature required a firmware with the corresponding V2 Key API
support. The feature (WPA3) is currently enabled and verified only
on IW416. Also, verified no regression with change when host MLME
is disabled.

Signed-off-by: David Lin <yu-hao.lin@nxp.com>
Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
---

v10:
   - none

v9:
   - remove unnecessary goto target.

v8:
   - first full and complete patch to support host based MLME for AP
     mode.

---
 .../net/wireless/marvell/mwifiex/cfg80211.c   |  79 +++++++-
 drivers/net/wireless/marvell/mwifiex/cmdevt.c |   2 +
 drivers/net/wireless/marvell/mwifiex/fw.h     |  21 +++
 drivers/net/wireless/marvell/mwifiex/ioctl.h  |   5 +
 .../wireless/marvell/mwifiex/sta_cmdresp.c    |   2 +
 .../net/wireless/marvell/mwifiex/uap_cmd.c    | 171 ++++++++++++++++++
 drivers/net/wireless/marvell/mwifiex/util.c   |  24 +++
 7 files changed, 301 insertions(+), 3 deletions(-)

Comments

Brian Norris May 23, 2024, 12:59 a.m. UTC | #1
Hi,

Hopefully-last comment for patch 2:

On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> --- a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> +++ b/drivers/net/wireless/marvell/mwifiex/cfg80211.c

> @@ -1712,7 +1735,7 @@ static const u32 mwifiex_cipher_suites[] = {
>  };
>  
>  /* Supported mgmt frame types to be advertised to cfg80211 */
> -static const struct ieee80211_txrx_stypes
> +static struct ieee80211_txrx_stypes
>  mwifiex_mgmt_stypes[NUM_NL80211_IFTYPES] = {
>  	[NL80211_IFTYPE_STATION] = {
>  		.tx = BIT(IEEE80211_STYPE_ACTION >> 4) |
...
> +	if (adapter->host_mlme_enabled) {
> +		mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].tx = 0xffff;
> +		mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].rx =
> +			BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
> +			BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
> +			BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
> +			BIT(IEEE80211_STYPE_DISASSOC >> 4) |
> +			BIT(IEEE80211_STYPE_AUTH >> 4) |
> +			BIT(IEEE80211_STYPE_DEAUTH >> 4) |
> +			BIT(IEEE80211_STYPE_ACTION >> 4);
> +	}

This doesn't look like a sound approach. I think you should keep
'mwifiex_mgmt_stypes' const, because if you're making modifications to
it, then you may break multi-adapter situations. Consider someone loads
this driver with host_mlme_enabled==true, and then later registers a
device with host_mlme_enabled==false. The second adapter will get the
wrong values.

I think 'mwifiex_mgmt_stypes' is small enough you might as well just
make a second copy with the MLME-enabled values, rather than fiddling
with modifying a single copy.

Aside from that:

Acked-by: Brian Norris <briannorris@chromium.org>

(Feel free to carry that to a v11, since my only remaining substantial
concerns are with patch 1 I think.)

Brian
David Lin May 23, 2024, 2:20 a.m. UTC | #2
> From: Brian Norris <briannorris@chromium.org>
> Sent: Thursday, May 23, 2024 8:59 AM
> To: David Lin <yu-hao.lin@nxp.com>
> Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> kvalo@kernel.org; francesco@dolcini.it; Pete Hsieh
> <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> <francesco.dolcini@toradex.com>
> Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP mode
> 
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report
> this email' button
> 
> 
> Hi,
> 
> Hopefully-last comment for patch 2:
> 
> On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > --- a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> > +++ b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> 
> > @@ -1712,7 +1735,7 @@ static const u32 mwifiex_cipher_suites[] = {  };
> >
> >  /* Supported mgmt frame types to be advertised to cfg80211 */ -static
> > const struct ieee80211_txrx_stypes
> > +static struct ieee80211_txrx_stypes
> >  mwifiex_mgmt_stypes[NUM_NL80211_IFTYPES] = {
> >       [NL80211_IFTYPE_STATION] = {
> >               .tx = BIT(IEEE80211_STYPE_ACTION >> 4) |
> ...
> > +     if (adapter->host_mlme_enabled) {
> > +             mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].tx = 0xffff;
> > +             mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].rx =
> > +                     BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
> > +                     BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
> > +                     BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
> > +                     BIT(IEEE80211_STYPE_DISASSOC >> 4) |
> > +                     BIT(IEEE80211_STYPE_AUTH >> 4) |
> > +                     BIT(IEEE80211_STYPE_DEAUTH >> 4) |
> > +                     BIT(IEEE80211_STYPE_ACTION >> 4);
> > +     }
> 
> This doesn't look like a sound approach. I think you should keep
> 'mwifiex_mgmt_stypes' const, because if you're making modifications to it,
> then you may break multi-adapter situations. Consider someone loads this
> driver with host_mlme_enabled==true, and then later registers a device with
> host_mlme_enabled==false. The second adapter will get the wrong values.
> 
> I think 'mwifiex_mgmt_stypes' is small enough you might as well just make a
> second copy with the MLME-enabled values, rather than fiddling with
> modifying a single copy.
> 
> Aside from that:
> 
> Acked-by: Brian Norris <briannorris@chromium.org>
> 
> (Feel free to carry that to a v11, since my only remaining substantial concerns
> are with patch 1 I think.)
> 
> Brian

I will modify mwifiex_mgmt_stypes for patch v11 and carry your "Acked-by" tag.

David
Sascha Hauer June 12, 2024, 1:11 p.m. UTC | #3
Hi David,

On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> Add host based MLME to enable WPA3 functionalities in AP mode.
> This feature required a firmware with the corresponding V2 Key API
> support. The feature (WPA3) is currently enabled and verified only
> on IW416. Also, verified no regression with change when host MLME
> is disabled.
> 
> Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> ---
> 

> diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> index 491e36611909..073c665183b3 100644
> --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct mwifiex_private *priv,
>  				bss_config->key_mgmt = KEY_MGMT_PSK;
>  			}
>  			break;
> +		case WLAN_AKM_SUITE_SAE:
> +			bss_config->protocol = PROTOCOL_WPA2;
> +			bss_config->key_mgmt = KEY_MGMT_SAE;
> +			break;

Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
Clearing the other flags when SAE is enabled looks wrong to me.

Sascha
David Lin June 14, 2024, 2:06 a.m. UTC | #4
> From: Sascha Hauer <s.hauer@pengutronix.de>
> Sent: Wednesday, June 12, 2024 9:12 PM
> To: David Lin <yu-hao.lin@nxp.com>
> Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it; Pete
> Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> <francesco.dolcini@toradex.com>
> Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP mode
> 
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report
> this email' button
> 
> 
> Hi David,
> 
> On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > Add host based MLME to enable WPA3 functionalities in AP mode.
> > This feature required a firmware with the corresponding V2 Key API
> > support. The feature (WPA3) is currently enabled and verified only on
> > IW416. Also, verified no regression with change when host MLME is
> > disabled.
> >
> > Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> > Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> > ---
> >
> 
> > diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > index 491e36611909..073c665183b3 100644
> > --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct
> mwifiex_private *priv,
> >                               bss_config->key_mgmt =
> KEY_MGMT_PSK;
> >                       }
> >                       break;
> > +             case WLAN_AKM_SUITE_SAE:
> > +                     bss_config->protocol = PROTOCOL_WPA2;
> > +                     bss_config->key_mgmt = KEY_MGMT_SAE;
> > +                     break;
> 
> Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
> Clearing the other flags when SAE is enabled looks wrong to me.
> 
> Sascha
> 

These fields are used for the configuration of FW, this is the correct setting.

David
Sascha Hauer June 14, 2024, 6:31 a.m. UTC | #5
On Fri, Jun 14, 2024 at 02:06:45AM +0000, David Lin wrote:
> > From: Sascha Hauer <s.hauer@pengutronix.de>
> > Sent: Wednesday, June 12, 2024 9:12 PM
> > To: David Lin <yu-hao.lin@nxp.com>
> > Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> > briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it; Pete
> > Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > <francesco.dolcini@toradex.com>
> > Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP mode
> > 
> > Caution: This is an external email. Please take care when clicking links or
> > opening attachments. When in doubt, report the message using the 'Report
> > this email' button
> > 
> > 
> > Hi David,
> > 
> > On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > > Add host based MLME to enable WPA3 functionalities in AP mode.
> > > This feature required a firmware with the corresponding V2 Key API
> > > support. The feature (WPA3) is currently enabled and verified only on
> > > IW416. Also, verified no regression with change when host MLME is
> > > disabled.
> > >
> > > Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> > > Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> > > ---
> > >
> > 
> > > diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > index 491e36611909..073c665183b3 100644
> > > --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct
> > mwifiex_private *priv,
> > >                               bss_config->key_mgmt =
> > KEY_MGMT_PSK;
> > >                       }
> > >                       break;
> > > +             case WLAN_AKM_SUITE_SAE:
> > > +                     bss_config->protocol = PROTOCOL_WPA2;
> > > +                     bss_config->key_mgmt = KEY_MGMT_SAE;
> > > +                     break;
> > 
> > Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
> > Clearing the other flags when SAE is enabled looks wrong to me.
> > 
> > Sascha
> > 
> 
> These fields are used for the configuration of FW, this is the correct setting.

This is done in a loop iterating over the different AKM suites, with your
patch this looks like this:

	for (i = 0; i < params->crypto.n_akm_suites; i++) {
		switch (params->crypto.akm_suites[i]) {
		case WLAN_AKM_SUITE_8021X:
			if (params->crypto.wpa_versions &
			    NL80211_WPA_VERSION_1) {
				bss_config->protocol = PROTOCOL_WPA;
				bss_config->key_mgmt = KEY_MGMT_EAP;
			}
			if (params->crypto.wpa_versions &
			    NL80211_WPA_VERSION_2) {
				bss_config->protocol |= PROTOCOL_WPA2;
				bss_config->key_mgmt = KEY_MGMT_EAP;
			}
			break;
		case WLAN_AKM_SUITE_PSK:
			if (params->crypto.wpa_versions &
			    NL80211_WPA_VERSION_1) {
				bss_config->protocol = PROTOCOL_WPA;
				bss_config->key_mgmt = KEY_MGMT_PSK;
			}
			if (params->crypto.wpa_versions &
			    NL80211_WPA_VERSION_2) {
				bss_config->protocol |= PROTOCOL_WPA2;
				bss_config->key_mgmt = KEY_MGMT_PSK;
			}
			break;
		case WLAN_AKM_SUITE_SAE:
			bss_config->protocol = PROTOCOL_WPA2;
			bss_config->key_mgmt = KEY_MGMT_SAE;
			break;

		default:
			break;
		}
	}

It looks wrong to overwrite bss_config->protocol and
bss_config->key_mgmt in each iteration of this loop. If that would be
correct, you wouldn't need a loop at all, but could instead configure
based on the last AKM suite entry.

In my understanding the bits in bss_config->key_mgmt should be ored
together depending on the possible AKM suites which is also what the
downstream driver does.

Sascha
David Lin June 17, 2024, 2:15 a.m. UTC | #6
> From: Sascha Hauer <s.hauer@pengutronix.de>
> Sent: Friday, June 14, 2024 2:32 PM
> To: David Lin <yu-hao.lin@nxp.com>
> Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it; Pete
> Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> <francesco.dolcini@toradex.com>
> Subject: Re: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP
> mode
>
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report
> this email' button
>
>
> On Fri, Jun 14, 2024 at 02:06:45AM +0000, David Lin wrote:
> > > From: Sascha Hauer <s.hauer@pengutronix.de>
> > > Sent: Wednesday, June 12, 2024 9:12 PM
> > > To: David Lin <yu-hao.lin@nxp.com>
> > > Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> > > briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it;
> > > Pete Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > > <francesco.dolcini@toradex.com>
> > > Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for
> > > AP mode
> > >
> > > Caution: This is an external email. Please take care when clicking
> > > links or opening attachments. When in doubt, report the message
> > > using the 'Report this email' button
> > >
> > >
> > > Hi David,
> > >
> > > On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > > > Add host based MLME to enable WPA3 functionalities in AP mode.
> > > > This feature required a firmware with the corresponding V2 Key API
> > > > support. The feature (WPA3) is currently enabled and verified only
> > > > on IW416. Also, verified no regression with change when host MLME
> > > > is disabled.
> > > >
> > > > Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> > > > Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> > > > ---
> > > >
> > >
> > > > diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > index 491e36611909..073c665183b3 100644
> > > > --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct
> > > mwifiex_private *priv,
> > > >                               bss_config->key_mgmt =
> > > KEY_MGMT_PSK;
> > > >                       }
> > > >                       break;
> > > > +             case WLAN_AKM_SUITE_SAE:
> > > > +                     bss_config->protocol = PROTOCOL_WPA2;
> > > > +                     bss_config->key_mgmt = KEY_MGMT_SAE;
> > > > +                     break;
> > >
> > > Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
> > > Clearing the other flags when SAE is enabled looks wrong to me.
> > >
> > > Sascha
> > >
> >
> > These fields are used for the configuration of FW, this is the correct setting.
>
> This is done in a loop iterating over the different AKM suites, with your patch
> this looks like this:
>
>         for (i = 0; i < params->crypto.n_akm_suites; i++) {
>                 switch (params->crypto.akm_suites[i]) {
>                 case WLAN_AKM_SUITE_8021X:
>                         if (params->crypto.wpa_versions &
>                             NL80211_WPA_VERSION_1) {
>                                 bss_config->protocol =
> PROTOCOL_WPA;
>                                 bss_config->key_mgmt =
> KEY_MGMT_EAP;
>                         }
>                         if (params->crypto.wpa_versions &
>                             NL80211_WPA_VERSION_2) {
>                                 bss_config->protocol |=
> PROTOCOL_WPA2;
>                                 bss_config->key_mgmt =
> KEY_MGMT_EAP;
>                         }
>                         break;
>                 case WLAN_AKM_SUITE_PSK:
>                         if (params->crypto.wpa_versions &
>                             NL80211_WPA_VERSION_1) {
>                                 bss_config->protocol =
> PROTOCOL_WPA;
>                                 bss_config->key_mgmt =
> KEY_MGMT_PSK;
>                         }
>                         if (params->crypto.wpa_versions &
>                             NL80211_WPA_VERSION_2) {
>                                 bss_config->protocol |=
> PROTOCOL_WPA2;
>                                 bss_config->key_mgmt =
> KEY_MGMT_PSK;
>                         }
>                         break;
>                 case WLAN_AKM_SUITE_SAE:
>                         bss_config->protocol = PROTOCOL_WPA2;
>                         bss_config->key_mgmt = KEY_MGMT_SAE;
>                         break;
>
>                 default:
>                         break;
>                 }
>         }
>
> It looks wrong to overwrite bss_config->protocol and bss_config->key_mgmt in
> each iteration of this loop. If that would be correct, you wouldn't need a loop
> at all, but could instead configure based on the last AKM suite entry.
>
> In my understanding the bits in bss_config->key_mgmt should be ored together
> depending on the possible AKM suites which is also what the downstream
> driver does.
>
> Sascha
>
> --
> Pengutronix e.K.                           |
> |
> Steuerwalder Str. 21                       |
> http://www.pen/
> gutronix.de%2F&data=05%7C02%7Cyu-hao.lin%40nxp.com%7C7c90883e11414
> 5a307fa08dc8c3ba753%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%
> 7C638539435049441786%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAw
> MDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&
> sdata=QSbX5IuEvP18tf0t6sMe8%2F4kuHo7IFEURWabdE4Cxcg%3D&reserved=0
> |
> 31137 Hildesheim, Germany                  | Phone:
> +49-5121-206917-0    |
> Amtsgericht Hildesheim, HRA 2686           | Fax:
> +49-5121-206917-5555 |

For the configuration of FW, ored only happens for the same AKM suite.

David
Sascha Hauer June 17, 2024, 6:29 a.m. UTC | #7
On Mon, Jun 17, 2024 at 02:15:41AM +0000, David Lin wrote:
> > From: Sascha Hauer <s.hauer@pengutronix.de>
> > Sent: Friday, June 14, 2024 2:32 PM
> > To: David Lin <yu-hao.lin@nxp.com>
> > Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> > briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it; Pete
> > Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > <francesco.dolcini@toradex.com>
> > Subject: Re: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP
> > mode
> >
> > Caution: This is an external email. Please take care when clicking links or
> > opening attachments. When in doubt, report the message using the 'Report
> > this email' button
> >
> >
> > On Fri, Jun 14, 2024 at 02:06:45AM +0000, David Lin wrote:
> > > > From: Sascha Hauer <s.hauer@pengutronix.de>
> > > > Sent: Wednesday, June 12, 2024 9:12 PM
> > > > To: David Lin <yu-hao.lin@nxp.com>
> > > > Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> > > > briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it;
> > > > Pete Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > > > <francesco.dolcini@toradex.com>
> > > > Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for
> > > > AP mode
> > > >
> > > > Caution: This is an external email. Please take care when clicking
> > > > links or opening attachments. When in doubt, report the message
> > > > using the 'Report this email' button
> > > >
> > > >
> > > > Hi David,
> > > >
> > > > On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > > > > Add host based MLME to enable WPA3 functionalities in AP mode.
> > > > > This feature required a firmware with the corresponding V2 Key API
> > > > > support. The feature (WPA3) is currently enabled and verified only
> > > > > on IW416. Also, verified no regression with change when host MLME
> > > > > is disabled.
> > > > >
> > > > > Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> > > > > Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> > > > > ---
> > > > >
> > > >
> > > > > diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > index 491e36611909..073c665183b3 100644
> > > > > --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct
> > > > mwifiex_private *priv,
> > > > >                               bss_config->key_mgmt =
> > > > KEY_MGMT_PSK;
> > > > >                       }
> > > > >                       break;
> > > > > +             case WLAN_AKM_SUITE_SAE:
> > > > > +                     bss_config->protocol = PROTOCOL_WPA2;
> > > > > +                     bss_config->key_mgmt = KEY_MGMT_SAE;
> > > > > +                     break;
> > > >
> > > > Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
> > > > Clearing the other flags when SAE is enabled looks wrong to me.
> > > >
> > > > Sascha
> > > >
> > >
> > > These fields are used for the configuration of FW, this is the correct setting.
> >
> > This is done in a loop iterating over the different AKM suites, with your patch
> > this looks like this:
> >
> >         for (i = 0; i < params->crypto.n_akm_suites; i++) {
> >                 switch (params->crypto.akm_suites[i]) {
> >                 case WLAN_AKM_SUITE_8021X:
> >                         if (params->crypto.wpa_versions &
> >                             NL80211_WPA_VERSION_1) {
> >                                 bss_config->protocol =
> > PROTOCOL_WPA;
> >                                 bss_config->key_mgmt =
> > KEY_MGMT_EAP;
> >                         }
> >                         if (params->crypto.wpa_versions &
> >                             NL80211_WPA_VERSION_2) {
> >                                 bss_config->protocol |=
> > PROTOCOL_WPA2;
> >                                 bss_config->key_mgmt =
> > KEY_MGMT_EAP;
> >                         }
> >                         break;
> >                 case WLAN_AKM_SUITE_PSK:
> >                         if (params->crypto.wpa_versions &
> >                             NL80211_WPA_VERSION_1) {
> >                                 bss_config->protocol =
> > PROTOCOL_WPA;
> >                                 bss_config->key_mgmt =
> > KEY_MGMT_PSK;
> >                         }
> >                         if (params->crypto.wpa_versions &
> >                             NL80211_WPA_VERSION_2) {
> >                                 bss_config->protocol |=
> > PROTOCOL_WPA2;
> >                                 bss_config->key_mgmt =
> > KEY_MGMT_PSK;
> >                         }
> >                         break;
> >                 case WLAN_AKM_SUITE_SAE:
> >                         bss_config->protocol = PROTOCOL_WPA2;
> >                         bss_config->key_mgmt = KEY_MGMT_SAE;
> >                         break;
> >
> >                 default:
> >                         break;
> >                 }
> >         }
> >
> > It looks wrong to overwrite bss_config->protocol and bss_config->key_mgmt in
> > each iteration of this loop. If that would be correct, you wouldn't need a loop
> > at all, but could instead configure based on the last AKM suite entry.
> >
> > In my understanding the bits in bss_config->key_mgmt should be ored together
> > depending on the possible AKM suites which is also what the downstream
> > driver does.
> >
> 
> For the configuration of FW, ored only happens for the same AKM suite.

Sorry, I don't understand this. Could you elaborate what you mean here?

Sascha
David Lin June 17, 2024, 7:44 a.m. UTC | #8
> From: Sascha Hauer <s.hauer@pengutronix.de>
> Sent: Monday, June 17, 2024 2:30 PM
> To: David Lin <yu-hao.lin@nxp.com>
> Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it; Pete
> Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> <francesco.dolcini@toradex.com>
> Subject: Re: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme for AP
> mode
>
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report
> this email' button
>
>
> On Mon, Jun 17, 2024 at 02:15:41AM +0000, David Lin wrote:
> > > From: Sascha Hauer <s.hauer@pengutronix.de>
> > > Sent: Friday, June 14, 2024 2:32 PM
> > > To: David Lin <yu-hao.lin@nxp.com>
> > > Cc: linux-wireless@vger.kernel.org; linux-kernel@vger.kernel.org;
> > > briannorris@chromium.org; kvalo@kernel.org; francesco@dolcini.it;
> > > Pete Hsieh <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > > <francesco.dolcini@toradex.com>
> > > Subject: Re: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme
> > > for AP mode
> > >
> > > Caution: This is an external email. Please take care when clicking
> > > links or opening attachments. When in doubt, report the message
> > > using the 'Report this email' button
> > >
> > >
> > > On Fri, Jun 14, 2024 at 02:06:45AM +0000, David Lin wrote:
> > > > > From: Sascha Hauer <s.hauer@pengutronix.de>
> > > > > Sent: Wednesday, June 12, 2024 9:12 PM
> > > > > To: David Lin <yu-hao.lin@nxp.com>
> > > > > Cc: linux-wireless@vger.kernel.org;
> > > > > linux-kernel@vger.kernel.org; briannorris@chromium.org;
> > > > > kvalo@kernel.org; francesco@dolcini.it; Pete Hsieh
> > > > > <tsung-hsien.hsieh@nxp.com>; Francesco Dolcini
> > > > > <francesco.dolcini@toradex.com>
> > > > > Subject: [EXT] Re: [PATCH v10 2/2] wifi: mwifiex: add host mlme
> > > > > for AP mode
> > > > >
> > > > > Caution: This is an external email. Please take care when
> > > > > clicking links or opening attachments. When in doubt, report the
> > > > > message using the 'Report this email' button
> > > > >
> > > > >
> > > > > Hi David,
> > > > >
> > > > > On Thu, Apr 18, 2024 at 02:06:26PM +0800, David Lin wrote:
> > > > > > Add host based MLME to enable WPA3 functionalities in AP mode.
> > > > > > This feature required a firmware with the corresponding V2 Key
> > > > > > API support. The feature (WPA3) is currently enabled and
> > > > > > verified only on IW416. Also, verified no regression with
> > > > > > change when host MLME is disabled.
> > > > > >
> > > > > > Signed-off-by: David Lin <yu-hao.lin@nxp.com>
> > > > > > Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
> > > > > > ---
> > > > > >
> > > > >
> > > > > > diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > > b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > > index 491e36611909..073c665183b3 100644
> > > > > > --- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > > +++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
> > > > > > @@ -72,6 +72,10 @@ int mwifiex_set_secure_params(struct
> > > > > mwifiex_private *priv,
> > > > > >                               bss_config->key_mgmt =
> > > > > KEY_MGMT_PSK;
> > > > > >                       }
> > > > > >                       break;
> > > > > > +             case WLAN_AKM_SUITE_SAE:
> > > > > > +                     bss_config->protocol = PROTOCOL_WPA2;
> > > > > > +                     bss_config->key_mgmt =
> KEY_MGMT_SAE;
> > > > > > +                     break;
> > > > >
> > > > > Shouldn't this be |= PROTOCOL_WPA2 and |= KEY_MGMT_SAE?
> > > > > Clearing the other flags when SAE is enabled looks wrong to me.
> > > > >
> > > > > Sascha
> > > > >
> > > >
> > > > These fields are used for the configuration of FW, this is the correct
> setting.
> > >
> > > This is done in a loop iterating over the different AKM suites, with
> > > your patch this looks like this:
> > >
> > >         for (i = 0; i < params->crypto.n_akm_suites; i++) {
> > >                 switch (params->crypto.akm_suites[i]) {
> > >                 case WLAN_AKM_SUITE_8021X:
> > >                         if (params->crypto.wpa_versions &
> > >                             NL80211_WPA_VERSION_1) {
> > >                                 bss_config->protocol =
> PROTOCOL_WPA;
> > >                                 bss_config->key_mgmt =
> KEY_MGMT_EAP;
> > >                         }
> > >                         if (params->crypto.wpa_versions &
> > >                             NL80211_WPA_VERSION_2) {
> > >                                 bss_config->protocol |=
> > > PROTOCOL_WPA2;
> > >                                 bss_config->key_mgmt =
> KEY_MGMT_EAP;
> > >                         }
> > >                         break;
> > >                 case WLAN_AKM_SUITE_PSK:
> > >                         if (params->crypto.wpa_versions &
> > >                             NL80211_WPA_VERSION_1) {
> > >                                 bss_config->protocol =
> PROTOCOL_WPA;
> > >                                 bss_config->key_mgmt =
> KEY_MGMT_PSK;
> > >                         }
> > >                         if (params->crypto.wpa_versions &
> > >                             NL80211_WPA_VERSION_2) {
> > >                                 bss_config->protocol |=
> > > PROTOCOL_WPA2;
> > >                                 bss_config->key_mgmt =
> KEY_MGMT_PSK;
> > >                         }
> > >                         break;
> > >                 case WLAN_AKM_SUITE_SAE:
> > >                         bss_config->protocol = PROTOCOL_WPA2;
> > >                         bss_config->key_mgmt = KEY_MGMT_SAE;
> > >                         break;
> > >
> > >                 default:
> > >                         break;
> > >                 }
> > >         }
> > >
> > > It looks wrong to overwrite bss_config->protocol and
> > > bss_config->key_mgmt in each iteration of this loop. If that would
> > > be correct, you wouldn't need a loop at all, but could instead configure
> based on the last AKM suite entry.
> > >
> > > In my understanding the bits in bss_config->key_mgmt should be ored
> > > together depending on the possible AKM suites which is also what the
> > > downstream driver does.
> > >
> >
> > For the configuration of FW, ored only happens for the same AKM suite.
>
> Sorry, I don't understand this. Could you elaborate what you mean here?
>
> Sascha
>
> --
> Pengutronix e.K.                           |
> |
> Steuerwalder Str. 21                       |
> http://www.pen/
> gutronix.de%2F&data=05%7C02%7Cyu-hao.lin%40nxp.com%7Cc852ce81dd734
> 2c25e0408dc8e96ebb2%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%
> 7C638542026048434949%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAw
> MDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&
> sdata=6Y%2FM9gFWYhf0UfClFTBwwsuMwrasEimaHKzJ2y2bHjQ%3D&reserved
> =0  |
> 31137 Hildesheim, Germany                  | Phone:
> +49-5121-206917-0    |
> Amtsgericht Hildesheim, HRA 2686           | Fax:
> +49-5121-206917-5555 |

Firmware can only support one of WLAN_AKM_SUITE_8021X, WLAN_AKM_SUITE_PSK, or WLAN_AKM_SUITE_SAE.
For WLAN_AKM_SUITE_8021X and WLAN_AKM_SUITE_PSK, the protocol can be WPA, WPA2 or both.

David
diff mbox series

Patch

diff --git a/drivers/net/wireless/marvell/mwifiex/cfg80211.c b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
index 53eeda388802..e122cc686dad 100644
--- a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
+++ b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
@@ -221,6 +221,26 @@  mwifiex_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 		return 0;
 	}
 
+	if (GET_BSS_ROLE(priv) == MWIFIEX_BSS_ROLE_UAP) {
+		if (ieee80211_is_auth(mgmt->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "auth: send auth to %pM\n", mgmt->da);
+		if (ieee80211_is_deauth(mgmt->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "auth: send deauth to %pM\n", mgmt->da);
+		if (ieee80211_is_disassoc(mgmt->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: send disassoc to %pM\n", mgmt->da);
+		if (ieee80211_is_assoc_resp(mgmt->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: send assoc resp to %pM\n",
+				    mgmt->da);
+		if (ieee80211_is_reassoc_resp(mgmt->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: send reassoc resp to %pM\n",
+				    mgmt->da);
+	}
+
 	pkt_len = len + ETH_ALEN;
 	skb = dev_alloc_skb(MWIFIEX_MIN_DATA_HEADER_LEN +
 			    MWIFIEX_MGMT_FRAME_HEADER_SIZE +
@@ -505,6 +525,9 @@  mwifiex_cfg80211_set_default_mgmt_key(struct wiphy *wiphy,
 
 	wiphy_dbg(wiphy, "set default mgmt key, key index=%d\n", key_index);
 
+	if (priv->adapter->host_mlme_enabled)
+		return 0;
+
 	memset(&encrypt_key, 0, sizeof(struct mwifiex_ds_encrypt_key));
 	encrypt_key.key_len = WLAN_KEY_LEN_CCMP;
 	encrypt_key.key_index = key_index;
@@ -1712,7 +1735,7 @@  static const u32 mwifiex_cipher_suites[] = {
 };
 
 /* Supported mgmt frame types to be advertised to cfg80211 */
-static const struct ieee80211_txrx_stypes
+static struct ieee80211_txrx_stypes
 mwifiex_mgmt_stypes[NUM_NL80211_IFTYPES] = {
 	[NL80211_IFTYPE_STATION] = {
 		.tx = BIT(IEEE80211_STYPE_ACTION >> 4) |
@@ -3951,12 +3974,43 @@  mwifiex_cfg80211_tdls_cancel_chan_switch(struct wiphy *wiphy,
 	}
 }
 
+static int
+mwifiex_cfg80211_uap_add_station(struct mwifiex_private *priv, const u8 *mac,
+				 struct station_parameters *params)
+{
+	struct mwifiex_sta_info add_sta;
+	int ret;
+
+	memcpy(add_sta.peer_mac, mac, ETH_ALEN);
+	add_sta.params = params;
+
+	ret = mwifiex_send_cmd(priv, HostCmd_CMD_ADD_NEW_STATION,
+			       HostCmd_ACT_ADD_STA, 0, (void *)&add_sta, true);
+
+	if (!ret) {
+		struct station_info *sinfo;
+
+		sinfo = kzalloc(sizeof(*sinfo), GFP_KERNEL);
+		if (!sinfo)
+			return -ENOMEM;
+
+		cfg80211_new_sta(priv->netdev, mac, sinfo, GFP_KERNEL);
+		kfree(sinfo);
+	}
+
+	return ret;
+}
+
 static int
 mwifiex_cfg80211_add_station(struct wiphy *wiphy, struct net_device *dev,
 			     const u8 *mac, struct station_parameters *params)
 {
 	struct mwifiex_private *priv = mwifiex_netdev_get_priv(dev);
 
+	if (priv->adapter->host_mlme_enabled &&
+	    (GET_BSS_ROLE(priv) == MWIFIEX_BSS_ROLE_UAP))
+		return mwifiex_cfg80211_uap_add_station(priv, mac, params);
+
 	if (!(params->sta_flags_set & BIT(NL80211_STA_FLAG_TDLS_PEER)))
 		return -EOPNOTSUPP;
 
@@ -4194,6 +4248,10 @@  mwifiex_cfg80211_change_station(struct wiphy *wiphy, struct net_device *dev,
 	int ret;
 	struct mwifiex_private *priv = mwifiex_netdev_get_priv(dev);
 
+	if (priv->adapter->host_mlme_enabled &&
+	    (GET_BSS_ROLE(priv) == MWIFIEX_BSS_ROLE_UAP))
+		return 0;
+
 	/* we support change_station handler only for TDLS peers*/
 	if (!(params->sta_flags_set & BIT(NL80211_STA_FLAG_TDLS_PEER)))
 		return -EOPNOTSUPP;
@@ -4662,6 +4720,17 @@  int mwifiex_register_cfg80211(struct mwifiex_adapter *adapter)
 	}
 	wiphy->max_scan_ssids = MWIFIEX_MAX_SSID_LIST_LENGTH;
 	wiphy->max_scan_ie_len = MWIFIEX_MAX_VSIE_LEN;
+	if (adapter->host_mlme_enabled) {
+		mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].tx = 0xffff;
+		mwifiex_mgmt_stypes[NL80211_IFTYPE_AP].rx =
+			BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
+			BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
+			BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
+			BIT(IEEE80211_STYPE_DISASSOC >> 4) |
+			BIT(IEEE80211_STYPE_AUTH >> 4) |
+			BIT(IEEE80211_STYPE_DEAUTH >> 4) |
+			BIT(IEEE80211_STYPE_ACTION >> 4);
+	}
 	wiphy->mgmt_stypes = mwifiex_mgmt_stypes;
 	wiphy->max_remain_on_channel_duration = 5000;
 	wiphy->interface_modes = BIT(NL80211_IFTYPE_STATION) |
@@ -4704,14 +4773,18 @@  int mwifiex_register_cfg80211(struct mwifiex_adapter *adapter)
 
 	ether_addr_copy(wiphy->perm_addr, adapter->perm_addr);
 	wiphy->signal_type = CFG80211_SIGNAL_TYPE_MBM;
-	wiphy->flags |= WIPHY_FLAG_HAVE_AP_SME |
-			WIPHY_FLAG_AP_PROBE_RESP_OFFLOAD |
+	wiphy->flags |= WIPHY_FLAG_AP_PROBE_RESP_OFFLOAD |
 			WIPHY_FLAG_AP_UAPSD |
 			WIPHY_FLAG_HAS_REMAIN_ON_CHANNEL |
 			WIPHY_FLAG_HAS_CHANNEL_SWITCH |
 			WIPHY_FLAG_NETNS_OK |
 			WIPHY_FLAG_PS_ON_BY_DEFAULT;
 
+	if (adapter->host_mlme_enabled)
+		wiphy->flags |= WIPHY_FLAG_REPORTS_OBSS;
+	else
+		wiphy->flags |= WIPHY_FLAG_HAVE_AP_SME;
+
 	if (ISSUPP_TDLS_ENABLED(adapter->fw_cap_info))
 		wiphy->flags |= WIPHY_FLAG_SUPPORTS_TDLS |
 				WIPHY_FLAG_TDLS_EXTERNAL_SETUP;
diff --git a/drivers/net/wireless/marvell/mwifiex/cmdevt.c b/drivers/net/wireless/marvell/mwifiex/cmdevt.c
index da983e27023c..ea6ebc9c23ef 100644
--- a/drivers/net/wireless/marvell/mwifiex/cmdevt.c
+++ b/drivers/net/wireless/marvell/mwifiex/cmdevt.c
@@ -635,6 +635,8 @@  int mwifiex_send_cmd(struct mwifiex_private *priv, u16 cmd_no,
 		case HostCmd_CMD_UAP_STA_DEAUTH:
 		case HOST_CMD_APCMD_SYS_RESET:
 		case HOST_CMD_APCMD_STA_LIST:
+		case HostCmd_CMD_CHAN_REPORT_REQUEST:
+		case HostCmd_CMD_ADD_NEW_STATION:
 			ret = mwifiex_uap_prepare_cmd(priv, cmd_no, cmd_action,
 						      cmd_oid, data_buf,
 						      cmd_ptr);
diff --git a/drivers/net/wireless/marvell/mwifiex/fw.h b/drivers/net/wireless/marvell/mwifiex/fw.h
index 0f89b86aa527..65799ae3bc72 100644
--- a/drivers/net/wireless/marvell/mwifiex/fw.h
+++ b/drivers/net/wireless/marvell/mwifiex/fw.h
@@ -211,6 +211,7 @@  enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define TLV_TYPE_CHAN_ATTR_CFG      (PROPRIETARY_TLV_BASE_ID + 237)
 #define TLV_TYPE_MAX_CONN           (PROPRIETARY_TLV_BASE_ID + 279)
 #define TLV_TYPE_HOST_MLME          (PROPRIETARY_TLV_BASE_ID + 307)
+#define TLV_TYPE_UAP_STA_FLAGS      (PROPRIETARY_TLV_BASE_ID + 313)
 #define TLV_TYPE_SAE_PWE_MODE       (PROPRIETARY_TLV_BASE_ID + 339)
 
 #define MWIFIEX_TX_DATA_BUF_SIZE_2K        2048
@@ -407,6 +408,7 @@  enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define HostCmd_CMD_STA_CONFIGURE		      0x023f
 #define HostCmd_CMD_CHAN_REGION_CFG		      0x0242
 #define HostCmd_CMD_PACKET_AGGR_CTRL		      0x0251
+#define HostCmd_CMD_ADD_NEW_STATION		      0x025f
 
 #define PROTOCOL_NO_SECURITY        0x01
 #define PROTOCOL_STATIC_WEP         0x02
@@ -417,6 +419,7 @@  enum MWIFIEX_802_11_PRIVACY_FILTER {
 #define KEY_MGMT_NONE               0x04
 #define KEY_MGMT_PSK                0x02
 #define KEY_MGMT_EAP                0x01
+#define KEY_MGMT_SAE                0x400
 #define CIPHER_TKIP                 0x04
 #define CIPHER_AES_CCMP             0x08
 #define VALID_CIPHER_BITMAP         0x0c
@@ -502,6 +505,9 @@  enum mwifiex_channel_flags {
 #define HostCmd_ACT_GET_TX              0x0008
 #define HostCmd_ACT_GET_BOTH            0x000c
 
+#define HostCmd_ACT_REMOVE_STA          0x0
+#define HostCmd_ACT_ADD_STA             0x1
+
 #define RF_ANTENNA_AUTO                 0xFFFF
 
 #define HostCmd_SET_SEQ_NO_BSS_INFO(seq, num, type) \
@@ -2331,6 +2337,20 @@  struct host_cmd_ds_sta_configure {
 	u8 tlv_buffer[];
 } __packed;
 
+struct mwifiex_ie_types_sta_flag {
+	struct mwifiex_ie_types_header header;
+	__le32 sta_flags;
+} __packed;
+
+struct host_cmd_ds_add_station {
+	__le16 action;
+	__le16 aid;
+	u8 peer_mac[ETH_ALEN];
+	__le32 listen_interval;
+	__le16 cap_info;
+	u8 tlv[];
+} __packed;
+
 struct host_cmd_ds_command {
 	__le16 command;
 	__le16 size;
@@ -2409,6 +2429,7 @@  struct host_cmd_ds_command {
 		struct host_cmd_ds_chan_region_cfg reg_cfg;
 		struct host_cmd_ds_pkt_aggr_ctrl pkt_aggr_ctrl;
 		struct host_cmd_ds_sta_configure sta_cfg;
+		struct host_cmd_ds_add_station sta_info;
 	} params;
 } __packed;
 
diff --git a/drivers/net/wireless/marvell/mwifiex/ioctl.h b/drivers/net/wireless/marvell/mwifiex/ioctl.h
index e8825f302de8..516159b721d3 100644
--- a/drivers/net/wireless/marvell/mwifiex/ioctl.h
+++ b/drivers/net/wireless/marvell/mwifiex/ioctl.h
@@ -158,6 +158,11 @@  struct mwifiex_bss_info {
 	u8 bssid[ETH_ALEN];
 };
 
+struct mwifiex_sta_info {
+	u8 peer_mac[ETH_ALEN];
+	struct station_parameters *params;
+};
+
 #define MAX_NUM_TID     8
 
 #define MAX_RX_WINSIZE  64
diff --git a/drivers/net/wireless/marvell/mwifiex/sta_cmdresp.c b/drivers/net/wireless/marvell/mwifiex/sta_cmdresp.c
index 7b69d27e0c0e..9c53825f222d 100644
--- a/drivers/net/wireless/marvell/mwifiex/sta_cmdresp.c
+++ b/drivers/net/wireless/marvell/mwifiex/sta_cmdresp.c
@@ -1398,6 +1398,8 @@  int mwifiex_process_sta_cmdresp(struct mwifiex_private *priv, u16 cmdresp_no,
 		break;
 	case HostCmd_CMD_UAP_STA_DEAUTH:
 		break;
+	case HostCmd_CMD_ADD_NEW_STATION:
+		break;
 	case HOST_CMD_APCMD_SYS_RESET:
 		break;
 	case HostCmd_CMD_MEF_CFG:
diff --git a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
index 491e36611909..073c665183b3 100644
--- a/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
+++ b/drivers/net/wireless/marvell/mwifiex/uap_cmd.c
@@ -72,6 +72,10 @@  int mwifiex_set_secure_params(struct mwifiex_private *priv,
 				bss_config->key_mgmt = KEY_MGMT_PSK;
 			}
 			break;
+		case WLAN_AKM_SUITE_SAE:
+			bss_config->protocol = PROTOCOL_WPA2;
+			bss_config->key_mgmt = KEY_MGMT_SAE;
+			break;
 		default:
 			break;
 		}
@@ -751,6 +755,28 @@  mwifiex_cmd_uap_sys_config(struct host_cmd_ds_command *cmd, u16 cmd_action,
 	return 0;
 }
 
+/* This function prepares AP start up command with or without host MLME
+ */
+static void mwifiex_cmd_uap_bss_start(struct mwifiex_private *priv,
+				     struct host_cmd_ds_command *cmd)
+{
+	struct mwifiex_ie_types_host_mlme *tlv;
+	int size;
+
+	cmd->command = cpu_to_le16(HostCmd_CMD_UAP_BSS_START);
+	size = S_DS_GEN;
+
+	if (priv->adapter->host_mlme_enabled) {
+		tlv = (struct mwifiex_ie_types_host_mlme *)((u8 *)cmd + size);
+		tlv->header.type = cpu_to_le16(TLV_TYPE_HOST_MLME);
+		tlv->header.len = cpu_to_le16(sizeof(tlv->host_mlme));
+		tlv->host_mlme = 1;
+		size += sizeof(struct mwifiex_ie_types_host_mlme);
+	}
+
+	cmd->size = cpu_to_le16(size);
+}
+
 /* This function prepares AP specific deauth command with mac supplied in
  * function parameter.
  */
@@ -768,6 +794,144 @@  static int mwifiex_cmd_uap_sta_deauth(struct mwifiex_private *priv,
 	return 0;
 }
 
+/* This function prepares AP specific add station command.
+ */
+static int mwifiex_cmd_uap_add_station(struct mwifiex_private *priv,
+				       struct host_cmd_ds_command *cmd,
+				       u16 cmd_action, void *data_buf)
+{
+	struct host_cmd_ds_add_station *new_sta = &cmd->params.sta_info;
+	struct mwifiex_sta_info *add_sta = (struct mwifiex_sta_info *)data_buf;
+	struct station_parameters *params = add_sta->params;
+	struct mwifiex_sta_node *sta_ptr;
+	u8 *pos;
+	u8 qos_capa;
+	u16 header_len = sizeof(struct mwifiex_ie_types_header);
+	u16 tlv_len;
+	int size;
+	struct mwifiex_ie_types_data *tlv;
+	struct mwifiex_ie_types_sta_flag *sta_flag;
+	int i;
+
+	cmd->command = cpu_to_le16(HostCmd_CMD_ADD_NEW_STATION);
+	new_sta->action = cpu_to_le16(cmd_action);
+	size = sizeof(struct host_cmd_ds_add_station) + S_DS_GEN;
+
+	if (cmd_action == HostCmd_ACT_ADD_STA)
+		sta_ptr = mwifiex_add_sta_entry(priv, add_sta->peer_mac);
+	else
+		sta_ptr = mwifiex_get_sta_entry(priv, add_sta->peer_mac);
+
+	if (!sta_ptr)
+		return -1;
+
+	memcpy(new_sta->peer_mac, add_sta->peer_mac, ETH_ALEN);
+
+	if (cmd_action == HostCmd_ACT_REMOVE_STA) {
+		cmd->size = cpu_to_le16(size);
+		return 0;
+	}
+
+	new_sta->aid = cpu_to_le16(params->aid);
+	new_sta->listen_interval = cpu_to_le32(params->listen_interval);
+	new_sta->cap_info = cpu_to_le16(params->capability);
+
+	pos = new_sta->tlv;
+
+	if (params->sta_flags_set & NL80211_STA_FLAG_WME)
+		sta_ptr->is_wmm_enabled = 1;
+	sta_flag = (struct mwifiex_ie_types_sta_flag *)pos;
+	sta_flag->header.type = cpu_to_le16(TLV_TYPE_UAP_STA_FLAGS);
+	sta_flag->header.len = cpu_to_le16(sizeof(__le32));
+	sta_flag->sta_flags = cpu_to_le32(params->sta_flags_set);
+	pos += sizeof(struct mwifiex_ie_types_sta_flag);
+	size += sizeof(struct mwifiex_ie_types_sta_flag);
+
+	if (params->ext_capab_len) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_EXT_CAPABILITY);
+		tlv_len = params->ext_capab_len;
+		tlv->header.len = cpu_to_le16(tlv_len);
+		memcpy(tlv->data, params->ext_capab, tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+	}
+
+	if (params->link_sta_params.supported_rates_len) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_SUPP_RATES);
+		tlv_len = params->link_sta_params.supported_rates_len;
+		tlv->header.len = cpu_to_le16(tlv_len);
+		memcpy(tlv->data,
+		       params->link_sta_params.supported_rates, tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+	}
+
+	if (params->uapsd_queues || params->max_sp) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_QOS_CAPA);
+		tlv_len = sizeof(qos_capa);
+		tlv->header.len = cpu_to_le16(tlv_len);
+		qos_capa = params->uapsd_queues | (params->max_sp << 5);
+		memcpy(tlv->data, &qos_capa, tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+		sta_ptr->is_wmm_enabled = 1;
+	}
+
+	if (params->link_sta_params.ht_capa) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_HT_CAPABILITY);
+		tlv_len = sizeof(struct ieee80211_ht_cap);
+		tlv->header.len = cpu_to_le16(tlv_len);
+		memcpy(tlv->data, params->link_sta_params.ht_capa, tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+		sta_ptr->is_11n_enabled = 1;
+		sta_ptr->max_amsdu =
+			le16_to_cpu(params->link_sta_params.ht_capa->cap_info) &
+			IEEE80211_HT_CAP_MAX_AMSDU ?
+			MWIFIEX_TX_DATA_BUF_SIZE_8K :
+			MWIFIEX_TX_DATA_BUF_SIZE_4K;
+	}
+
+	if (params->link_sta_params.vht_capa) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_VHT_CAPABILITY);
+		tlv_len = sizeof(struct ieee80211_vht_cap);
+		tlv->header.len = cpu_to_le16(tlv_len);
+		memcpy(tlv->data, params->link_sta_params.vht_capa, tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+		sta_ptr->is_11ac_enabled = 1;
+	}
+
+	if (params->link_sta_params.opmode_notif_used) {
+		tlv = (struct mwifiex_ie_types_data *)pos;
+		tlv->header.type = cpu_to_le16(WLAN_EID_OPMODE_NOTIF);
+		tlv_len = sizeof(u8);
+		tlv->header.len = cpu_to_le16(tlv_len);
+		memcpy(tlv->data, &params->link_sta_params.opmode_notif,
+		       tlv_len);
+		pos += (header_len + tlv_len);
+		size += (header_len + tlv_len);
+	}
+
+	for (i = 0; i < MAX_NUM_TID; i++) {
+		if (sta_ptr->is_11n_enabled)
+			sta_ptr->ampdu_sta[i] =
+				      priv->aggr_prio_tbl[i].ampdu_user;
+		else
+			sta_ptr->ampdu_sta[i] = BA_STREAM_NOT_ALLOWED;
+	}
+
+	memset(sta_ptr->rx_seq, 0xff, sizeof(sta_ptr->rx_seq));
+	cmd->size = cpu_to_le16(size);
+
+	return 0;
+}
+
 /* This function prepares the AP specific commands before sending them
  * to the firmware.
  * This is a generic function which calls specific command preparation
@@ -785,6 +949,8 @@  int mwifiex_uap_prepare_cmd(struct mwifiex_private *priv, u16 cmd_no,
 			return -1;
 		break;
 	case HostCmd_CMD_UAP_BSS_START:
+		mwifiex_cmd_uap_bss_start(priv, cmd);
+		break;
 	case HostCmd_CMD_UAP_BSS_STOP:
 	case HOST_CMD_APCMD_SYS_RESET:
 	case HOST_CMD_APCMD_STA_LIST:
@@ -800,6 +966,11 @@  int mwifiex_uap_prepare_cmd(struct mwifiex_private *priv, u16 cmd_no,
 							  data_buf))
 			return -1;
 		break;
+	case HostCmd_CMD_ADD_NEW_STATION:
+		if (mwifiex_cmd_uap_add_station(priv, cmd, cmd_action,
+						data_buf))
+			return -1;
+		break;
 	default:
 		mwifiex_dbg(priv->adapter, ERROR,
 			    "PREP_CMD: unknown cmd %#x\n", cmd_no);
diff --git a/drivers/net/wireless/marvell/mwifiex/util.c b/drivers/net/wireless/marvell/mwifiex/util.c
index 3817c08a1507..42c04bf858da 100644
--- a/drivers/net/wireless/marvell/mwifiex/util.c
+++ b/drivers/net/wireless/marvell/mwifiex/util.c
@@ -497,6 +497,30 @@  mwifiex_process_mgmt_packet(struct mwifiex_private *priv,
 		cfg80211_rx_mlme_mgmt(priv->netdev, skb->data, pkt_len);
 	}
 
+	if (priv->adapter->host_mlme_enabled &&
+	    (GET_BSS_ROLE(priv) == MWIFIEX_BSS_ROLE_UAP)) {
+		if (ieee80211_is_auth(ieee_hdr->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "auth: receive auth from %pM\n",
+				    ieee_hdr->addr2);
+		if (ieee80211_is_deauth(ieee_hdr->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "auth: receive deauth from %pM\n",
+				    ieee_hdr->addr2);
+		if (ieee80211_is_disassoc(ieee_hdr->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: receive disassoc from %pM\n",
+				    ieee_hdr->addr2);
+		if (ieee80211_is_assoc_req(ieee_hdr->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: receive assoc req from %pM\n",
+				    ieee_hdr->addr2);
+		if (ieee80211_is_reassoc_req(ieee_hdr->frame_control))
+			mwifiex_dbg(priv->adapter, MSG,
+				    "assoc: receive reassoc req from %pM\n",
+				    ieee_hdr->addr2);
+	}
+
 	cfg80211_rx_mgmt(&priv->wdev, priv->roc_cfg.chan.center_freq,
 			 CAL_RSSI(rx_pd->snr, rx_pd->nf), skb->data, pkt_len,
 			 0);