Content-Type: multipart/mixed; boundary="===============4280655170715453421==" MIME-Version: 1.0 From: Ariel D'Alessandro Subject: Re: [connman v1] wifi: Set ieee80211w as required for WPA3-Personal-only mode Date: Wed, 04 Aug 2021 18:17:20 -0300 Message-ID: <39fccc8f-4f53-ccff-97ba-194258f07923@collabora.com> In-Reply-To: <20210803074104.dawii3v3bknovg2e@beryllium.lan> List-Id: To: iwd@lists.01.org --===============4280655170715453421== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Daniel, On 8/3/21 4:41 AM, Daniel Wagner wrote: > Hi Ariel, > = > Thanks for tracking this! Just a stupid question: Thanks for the quick feedback :-) see below. > = > On Mon, Aug 02, 2021 at 12:10:06PM -0300, Ariel D'Alessandro wrote: >> From wpa_supplicant docs [0] section "ieee80211w": >> >> "WPA3-Personal-only mode: ieee80211w=3D2 and key_mgmt=3DSAE". >> >> This means that the parameter ieee80211w has to be set to 2 (required) >> if SAE is used. Fix it. >> >> [0] https://w1.fi/cgit/hostap/plain/wpa_supplicant/wpa_supplicant.conf >> --- >> plugins/wifi.c | 7 ++++++- >> 1 file changed, 6 insertions(+), 1 deletion(-) >> >> diff --git a/plugins/wifi.c b/plugins/wifi.c >> index 578b4be7..02603e85 100644 >> --- a/plugins/wifi.c >> +++ b/plugins/wifi.c >> @@ -2174,7 +2174,12 @@ static void ssid_init(GSupplicantSSID *ssid, stru= ct connman_network *network) >> security =3D connman_network_get_string(network, "WiFi.Security"); >> ssid->security =3D network_security(security); >> ssid->keymgmt =3D network_data->keymgmt; >> - ssid->ieee80211w =3D G_SUPPLICANT_MFP_OPTIONAL; >> + /* >> + * From wpa_supplicant docs: >> + * "WPA3-Personal-only mode: ieee80211w=3D2 and key_mgmt=3DSAE". >> + * Set ieee80211w as required, which is sent if SAE key is used. >> + */ >> + ssid->ieee80211w =3D G_SUPPLICANT_MFP_REQUIRED; > = > Should we check the keymgmt is SAE and only then set the ieee80211w > value? Thanks for asking that, two answers: * `ssid->ieee80211w` only gets sent if key management is SAE, so that'd be redundant. See this code block [0]. So, there's no real need to check the keymgmt is SAE here, and `required` could be the default value. * On the other hand, my concern now is about supporting SAE transition mode. Let's *drop* this patchset and I'll send an RFC patch implementing WPA3-Personal transition mode. Thanks! Ariel [0] https://git.kernel.org/pub/scm/network/connman/connman.git/tree/gsupplicant= /supplicant.c#n4908 --===============4280655170715453421==--