[PATCH v2 1/1] hostapd: Add support for OWE offload for STA/AP interface

[Vinayak Yadawad]

Hi Jouni,

Thanks for the review comments.

>How would this work for 4-way handshake since hostapd and wpa_supplicant
>would need to know the PMK? Or is this limited only to cases where the
>4-way handshake is offloaded? If so, that should be clearly mentioned
>here in the commit message. And for the AP mode, I guess that would
>imply dependency on the SAE offloading patch (or the part of it that I
>asked to be separated to an independent patch).
Yes, as suggested we will update the description to indicate
dependency on 4way HS offload.
We will also have separate patches for PSK offload, SAE offload and
the current OWE offload.

>Please use _STA postfix for the STA capability to match the _AP postfix
>style.
Ack.

>Why is this needed? Why would there be an NL80211_CMD_UPDATE_OWE_INFO
>event from a driver that uses OWE offloading?
Agree. We added this as a safety check, but will remove.

>No need to include this in hostap.git contributions; I update this file
>directly from wireless-next.git.
Ack

Regards,
Vinayak

On Tue, Nov 7, 2023 at 3:11 PM Jouni Malinen <j at w1.fi> wrote:
>
> On Tue, Oct 31, 2023 at 03:33:47PM +0530, Vinayak Yadawad wrote:
> > Driver/fw advertising OWE offload support would take care of
> > DH IE generation and processing part. Driver/FW would be
> > responsible for OWE PMK generation in this case.
> > This patch avoids the DH IE handling in wpa_supplicant/hostapd
> > for drivers advertising OWE offload support.
>
> How would this work for 4-way handshake since hostapd and wpa_supplicant
> would need to know the PMK? Or is this limited only to cases where the
> 4-way handshake is offloaded? If so, that should be clearly mentioned
> here in the commit message. And for the AP mode, I guess that would
> imply dependency on the SAE offloading patch (or the part of it that I
> asked to be separated to an independent patch).
>
> > diff --git a/src/drivers/driver.h b/src/drivers/driver.h
> > @@ -2253,6 +2253,10 @@ struct wpa_driver_capa {
> > +/** Driver supports OWE STA offload */
> > +#define WPA_DRIVER_FLAGS2_OWE_OFFLOAD        0x0000000000008000ULL
> > +/** Driver supports OWE AP offload */
> > +#define WPA_DRIVER_FLAGS2_OWE_OFFLOAD_AP     0x0000000000010000ULL
>
> Please use _STA postfix for the STA capability to match the _AP postfix
> style.
>
> > diff --git a/src/drivers/driver_nl80211_event.c b/src/drivers/driver_nl80211_event.c
> > @@ -1908,7 +1908,8 @@ static void mlme_event_dh_event(struct wpa_driver_nl80211_data *drv,
> >       u8 *addr, *link_addr = NULL;
> >       int assoc_link_id = -1;
> >
> > -     if (!is_ap_interface(drv->nlmode))
> > +     if (!is_ap_interface(drv->nlmode) ||
> > +         (drv->capa.flags2 & WPA_DRIVER_FLAGS2_OWE_OFFLOAD_AP))
> >               return;
>
> Why is this needed? Why would there be an NL80211_CMD_UPDATE_OWE_INFO
> event from a driver that uses OWE offloading?
>
> > diff --git a/src/drivers/nl80211_copy.h b/src/drivers/nl80211_copy.h
>
> No need to include this in hostap.git contributions; I update this file
> directly from wireless-next.git.
>
> --
> Jouni Malinen                                            PGP id EFC895FA
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4218 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.infradead.org/pipermail/hostap/attachments/20231108/4ce0e9fb/attachment-0001.p7s>