Skip to contentSkip to conversation

rsi: Fix TX EAPOL packet handling against iwlwifi AP

PATCH In review Marek Vasut 3 participants 4 messages Oct 15, 2020 – May 27, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 rsi: Fix TX EAPOL packet handling against iwlwifi AP

In case RSI9116 SDIO WiFi operates in STA mode against Intel 9260 in AP mode, the association fails. The former is using wpa_supplicant during association, the later is set up using hostapd:

iwl$ cat hostapd.conf
interface=wlp1s0
ssid=test
country_code=DE
hw_mode=g
channel=1
wpa=2
wpa_passphrase=test
wpa_key_mgmt=WPA-PSK
iwl$ hostapd -d hostapd.conf

rsi$ wpa_supplicant -i wlan0 -c <(wpa_passphrase test test)

The problem is that the TX EAPOL data descriptor RSI_DESC_REQUIRE_CFM_TO_HOST flag and extended descriptor EAPOL4_CONFIRM frame type are not set in case the AP is iwlwifi, because in that case the TX EAPOL packet is 2 bytes shorter.

The downstream vendor driver has this change in place already [1], however there is no explanation for it, neither is there any commit history from which such explanation could be obtained.

[1] https://github.com/SiliconLabs/RS911X-nLink-OSD/blob/master/rsi/rsi_91x_hal.c#L238

Cc Angus Ainslie <[email protected]>
Cc David S. Miller <[email protected]>
Cc Jakub Kicinski <[email protected]>
Cc Kalle Valo <[email protected]>
Cc Lee Jones <[email protected]>
Cc Martin Kepplinger <[email protected]>
Cc Sebastian Krzyszkowiak <[email protected]>
Cc Siva Rebbagondla <[email protected]>
---
 drivers/net/wireless/rsi/rsi_91x_hal.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/wireless/rsi/rsi_91x_hal.c +2 -1
index 3f7e3cfb6f00..ce9892152f4d 100644
@@ -248,7 +248,8 @@ int rsi_prepare_data_desc(struct rsi_common *common, struct sk_buff *skb)
 			rsi_set_len_qno(&data_desc->len_qno,
 					(skb->len - FRAME_DESC_SZ),
 					RSI_WIFI_MGMT_Q);
-		if ((skb->len - header_size) == EAPOL4_PACKET_LEN) {
+		if (((skb->len - header_size) == EAPOL4_PACKET_LEN) ||
+		    ((skb->len - header_size) == EAPOL4_PACKET_LEN - 2)) {
 			data_desc->misc_flags |=
 				RSI_DESC_REQUIRE_CFM_TO_HOST;
 			xtend_desc->confirm_frame_type = EAPOL4_CONFIRM;
Signature · 1 line
2.28.0
#1 lore raw
Quoted 27 lines · Marek Vasut

In case RSI9116 SDIO WiFi operates in STA mode against Intel 9260 in AP mode, the association fails. The former is using wpa_supplicant during association, the later is set up using hostapd:

iwl$ cat hostapd.conf
interface=wlp1s0
ssid=test
country_code=DE
hw_mode=g
channel=1
wpa=2
wpa_passphrase=test
wpa_key_mgmt=WPA-PSK
iwl$ hostapd -d hostapd.conf

rsi$ wpa_supplicant -i wlan0 -c <(wpa_passphrase test test)

The problem is that the TX EAPOL data descriptor RSI_DESC_REQUIRE_CFM_TO_HOST flag and extended descriptor EAPOL4_CONFIRM frame type are not set in case the AP is iwlwifi, because in that case the TX EAPOL packet is 2 bytes shorter.

The downstream vendor driver has this change in place already [1], however there is no explanation for it, neither is there any commit history from which such explanation could be obtained.

[1] https://github.com/SiliconLabs/RS911X-nLink-OSD/blob/master/rsi/rsi_91x_hal.c#L238

Quoted 11 lines
Cc Angus Ainslie <[email protected]>
Cc David S. Miller <[email protected]>
Cc Jakub Kicinski <[email protected]>
Cc Kalle Valo <[email protected]>
Cc Lee Jones <[email protected]>
Cc Martin Kepplinger <[email protected]>
Cc Sebastian Krzyszkowiak <[email protected]>
Cc Siva Rebbagondla <[email protected]>

Patch applied to wireless-drivers-next.git, thanks.

65277100caa2 rsi: Fix TX EAPOL packet handling against iwlwifi AP

Signature · 3 lines
https://patchwork.kernel.org/project/linux-wireless/patch/[email protected]/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
#2 lore raw
In reply to Marek Vasut

Hi Marek,

I've just run into the same problem (on -5.4) and found your (now merged) patch

Quoted 17 lines · Marek Vasut

In case RSI9116 SDIO WiFi operates in STA mode against Intel 9260 in AP mode, the association fails. The former is using wpa_supplicant during association, the later is set up using hostapd:

iwl$ cat hostapd.conf
interface=wlp1s0
ssid=test
country_code=DE
hw_mode=g
channel=1
wpa=2
wpa_passphrase=test
wpa_key_mgmt=WPA-PSK
iwl$ hostapd -d hostapd.conf

rsi$ wpa_supplicant -i wlan0 -c <(wpa_passphrase test test)

Quoted 8 lines

The problem is that the TX EAPOL data descriptor RSI_DESC_REQUIRE_CFM_TO_HOST flag and extended descriptor EAPOL4_CONFIRM frame type are not set in case the AP is iwlwifi, because in that case the TX EAPOL packet is 2 bytes shorter.

The downstream vendor driver has this change in place already [1], however there is no explanation for it, neither is there any commit history from which such explanation could be obtained.

I get this using 2 RSI9116 s, for both AP and STA using hostapd.
Comparing packet captures in the working and non working (without your patch) case shows that the working case has a 802.11 QOS header whereas the non working case does not, hence the 2 byte difference. The size of the EAPOL data is the same, it's the previous header that causes the problem...

This whole use the message size to determine the messages to ACK seems very fragile...

Regards,

Martin

#3 lore raw
Quoted 1 line · Martin Fuzzey

Hi Marek,

Hi,

Quoted 1 line

I've just run into the same problem (on -5.4) and found your (now merged) patch

The patch should already be part of 5.4.y, no ?

Quoted 22 lines · Marek Vasut
Quoted 21 lines

In case RSI9116 SDIO WiFi operates in STA mode against Intel 9260 in AP mode, the association fails. The former is using wpa_supplicant during association, the later is set up using hostapd:

iwl$ cat hostapd.conf
interface=wlp1s0
ssid=test
country_code=DE
hw_mode=g
channel=1
wpa=2
wpa_passphrase=test
wpa_key_mgmt=WPA-PSK
iwl$ hostapd -d hostapd.conf

rsi$ wpa_supplicant -i wlan0 -c <(wpa_passphrase test test)

The problem is that the TX EAPOL data descriptor RSI_DESC_REQUIRE_CFM_TO_HOST flag and extended descriptor EAPOL4_CONFIRM frame type are not set in case the AP is iwlwifi, because in that case the TX EAPOL packet is 2 bytes shorter.

Quoted 6 lines
Quoted 4 lines

The downstream vendor driver has this change in place already [1], however there is no explanation for it, neither is there any commit history from which such explanation could be obtained.

I get this using 2 RSI9116 s, for both AP and STA using hostapd.

Do I understand it correctly that two RSI9116 did not even work against one another as STA and AP respectively ? Sigh ...

Quoted 5 lines

Comparing packet captures in the working and non working (without your patch) case shows that the working case has a 802.11 QOS header whereas the non working case does not, hence the 2 byte difference. The size of the EAPOL data is the same, it's the previous header that causes the problem...

This whole use the message size to determine the messages to ACK seems very fragile...

I'm not surprised, the quality of this driver is low and the documentation is lacking. Thanks for clarifying.

Do you think you can write and submit a patch which would fix this in a better way?