diff mbox series

[v3,4/5] brcmfmac: Update SSID of hidden AP while informing its bss to cfg80211 layer

Message ID 20220927034138.20463-5-ian.lin@infineon.com
State New
Headers show
Series Fix connect/p2p issue series | expand

Commit Message

Ian Lin Sept. 27, 2022, 3:41 a.m. UTC 
From: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>

cfg80211 layer on DUT STA is disconnecting ongoing connection attempt after
receiving association response, because cfg80211 layer does not have valid
AP bss information. On association response event, brcmfmac communicates
the AP bss information to cfg80211 layer, but SSID seem to be empty in AP
bss information, and cfg80211 layer prints kernel warning and then
disconnects the ongoing connection attempt.

SSID is empty in SSID IE, but 'bi->SSID' contains a valid SSID, so
updating the SSID for hidden AP while informing its bss information
to cfg80211 layer.

Signed-off-by: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
Signed-off-by: Chung-Hsien Hsu <chung-hsien.hsu@infineon.com>
Signed-off-by: Chi-hsien Lin <chi-hsien.lin@infineon.com>
Signed-off-by: Ian Lin <ian.lin@infineon.com>
---
 .../net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c   | 8 ++++++++
 1 file changed, 8 insertions(+)

Comments

Kalle Valo Sept. 28, 2022, 6:38 a.m. UTC | #1 
Ian Lin <ian.lin@infineon.com> writes:

> From: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>
> cfg80211 layer on DUT STA is disconnecting ongoing connection attempt after
> receiving association response, because cfg80211 layer does not have valid
> AP bss information. On association response event, brcmfmac communicates
> the AP bss information to cfg80211 layer, but SSID seem to be empty in AP
> bss information, and cfg80211 layer prints kernel warning and then
> disconnects the ongoing connection attempt.
>
> SSID is empty in SSID IE, but 'bi->SSID' contains a valid SSID, so
> updating the SSID for hidden AP while informing its bss information
> to cfg80211 layer.
>
> Signed-off-by: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
> Signed-off-by: Chung-Hsien Hsu <chung-hsien.hsu@infineon.com>
> Signed-off-by: Chi-hsien Lin <chi-hsien.lin@infineon.com>
> Signed-off-by: Ian Lin <ian.lin@infineon.com>
> ---
>  .../net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c   | 8 ++++++++
>  1 file changed, 8 insertions(+)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> index 6c37da42e61b..3560afe0ccfe 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> @@ -3003,6 +3003,7 @@ static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
>  	u8 *notify_ie;
>  	size_t notify_ielen;
>  	struct cfg80211_inform_bss bss_data = {};
> +	struct brcmf_tlv *ssid;
>  
>  	if (le32_to_cpu(bi->length) > WL_BSS_INFO_MAX) {
>  		bphy_err(drvr, "Bss info is larger than buffer. Discarding\n");
> @@ -3032,6 +3033,13 @@ static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
>  	notify_ielen = le32_to_cpu(bi->ie_length);
>  	bss_data.signal = (s16)le16_to_cpu(bi->RSSI) * 100;
>  
> +	ssid = (struct brcmf_tlv *)
> +		brcmf_parse_tlvs(notify_ie, notify_ielen, WLAN_EID_SSID);

This still casts away the const. For some reason brcmf_parse_tlvs()
takes a const buffer:

static const struct brcmf_tlv *brcmf_parse_tlvs(const void *buf, int buflen, uint key)

I'm not familiar with brcmfmac internal so I don't know why it does
that, but that means the buffer cannot be modified. If you need to
modify the ssid should you make a copy of it?
Ian Lin Sept. 28, 2022, 10:07 a.m. UTC | #2 
On 9/28/2022 2:38 PM, Kalle Valo wrote:
> Caution: This e-mail originated outside Infineon Technologies. Do not click on links or open attachments unless you validate it is safe<https://intranet-content.infineon.com/explore/aboutinfineon/rules/informationsecurity/ug/SocialEngineering/Pages/SocialEngineeringElements_en.aspx>.
>
>
>
> Ian Lin <ian.lin@infineon.com> writes:
>
>> From: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>>
>> cfg80211 layer on DUT STA is disconnecting ongoing connection attempt after
>> receiving association response, because cfg80211 layer does not have valid
>> AP bss information. On association response event, brcmfmac communicates
>> the AP bss information to cfg80211 layer, but SSID seem to be empty in AP
>> bss information, and cfg80211 layer prints kernel warning and then
>> disconnects the ongoing connection attempt.
>>
>> SSID is empty in SSID IE, but 'bi->SSID' contains a valid SSID, so
>> updating the SSID for hidden AP while informing its bss information
>> to cfg80211 layer.
>>
>> Signed-off-by: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>> Signed-off-by: Chung-Hsien Hsu <chung-hsien.hsu@infineon.com>
>> Signed-off-by: Chi-hsien Lin <chi-hsien.lin@infineon.com>
>> Signed-off-by: Ian Lin <ian.lin@infineon.com>
>> ---
>>   .../net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c   | 8 ++++++++
>>   1 file changed, 8 insertions(+)
>>
>> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>> index 6c37da42e61b..3560afe0ccfe 100644
>> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>> @@ -3003,6 +3003,7 @@ static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
>>        u8 *notify_ie;
>>        size_t notify_ielen;
>>        struct cfg80211_inform_bss bss_data = {};
>> +     struct brcmf_tlv *ssid;
>>
>>        if (le32_to_cpu(bi->length) > WL_BSS_INFO_MAX) {
>>                bphy_err(drvr, "Bss info is larger than buffer. Discarding\n");
>> @@ -3032,6 +3033,13 @@ static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
>>        notify_ielen = le32_to_cpu(bi->ie_length);
>>        bss_data.signal = (s16)le16_to_cpu(bi->RSSI) * 100;
>>
>> +     ssid = (struct brcmf_tlv *)
>> +             brcmf_parse_tlvs(notify_ie, notify_ielen, WLAN_EID_SSID);
> This still casts away the const. For some reason brcmf_parse_tlvs()
> takes a const buffer:
>
> static const struct brcmf_tlv *brcmf_parse_tlvs(const void *buf, int buflen, uint key)
>
> I'm not familiar with brcmfmac internal so I don't know why it does
> that, but that means the buffer cannot be modified. If you need to
> modify the ssid should you make a copy of it?

In brcmf_parse_tlvs(const void *buf, int buflen, uint key),
it will find the key in buf and return the address of this key, as the 
return pointer.
This function don't intend caller to modify content of buf in most 
cases, so it defines a const return pointer.
But in this case, it just need to do it, so I need the typecast here.

> --
> https://patchwork.kernel.org/project/linux-wireless/list/
>
> https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
Ian Lin Oct. 4, 2022, 11:01 a.m. UTC | #3 
On 9/28/2022 6:07 PM, Lin Ian (CSSITB CSS ICW SW WFS / EE) wrote:
>
>
> On 9/28/2022 2:38 PM, Kalle Valo wrote:
>> Caution: This e-mail originated outside Infineon Technologies. Do not 
>> click on links or open attachments unless you validate it is 
>> safe<https://intranet-content.infineon.com/explore/aboutinfineon/rules/informationsecurity/ug/SocialEngineering/Pages/SocialEngineeringElements_en.aspx>.
>>
>>
>>
>> Ian Lin <ian.lin@infineon.com> writes:
>>
>>> From: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>>>
>>> cfg80211 layer on DUT STA is disconnecting ongoing connection 
>>> attempt after
>>> receiving association response, because cfg80211 layer does not have 
>>> valid
>>> AP bss information. On association response event, brcmfmac 
>>> communicates
>>> the AP bss information to cfg80211 layer, but SSID seem to be empty 
>>> in AP
>>> bss information, and cfg80211 layer prints kernel warning and then
>>> disconnects the ongoing connection attempt.
>>>
>>> SSID is empty in SSID IE, but 'bi->SSID' contains a valid SSID, so
>>> updating the SSID for hidden AP while informing its bss information
>>> to cfg80211 layer.
>>>
>>> Signed-off-by: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>>> Signed-off-by: Chung-Hsien Hsu <chung-hsien.hsu@infineon.com>
>>> Signed-off-by: Chi-hsien Lin <chi-hsien.lin@infineon.com>
>>> Signed-off-by: Ian Lin <ian.lin@infineon.com>
>>> ---
>>>   .../net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c   | 8 
>>> ++++++++
>>>   1 file changed, 8 insertions(+)
>>>
>>> diff --git 
>>> a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c 
>>> b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>> index 6c37da42e61b..3560afe0ccfe 100644
>>> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>> @@ -3003,6 +3003,7 @@ static s32 brcmf_inform_single_bss(struct 
>>> brcmf_cfg80211_info *cfg,
>>>        u8 *notify_ie;
>>>        size_t notify_ielen;
>>>        struct cfg80211_inform_bss bss_data = {};
>>> +     struct brcmf_tlv *ssid;
>>>
>>>        if (le32_to_cpu(bi->length) > WL_BSS_INFO_MAX) {
>>>                bphy_err(drvr, "Bss info is larger than buffer. 
>>> Discarding\n");
>>> @@ -3032,6 +3033,13 @@ static s32 brcmf_inform_single_bss(struct 
>>> brcmf_cfg80211_info *cfg,
>>>        notify_ielen = le32_to_cpu(bi->ie_length);
>>>        bss_data.signal = (s16)le16_to_cpu(bi->RSSI) * 100;
>>>
>>> +     ssid = (struct brcmf_tlv *)
>>> +             brcmf_parse_tlvs(notify_ie, notify_ielen, WLAN_EID_SSID);
>> This still casts away the const. For some reason brcmf_parse_tlvs()
>> takes a const buffer:
>>
>> static const struct brcmf_tlv *brcmf_parse_tlvs(const void *buf, int 
>> buflen, uint key)
>>
>> I'm not familiar with brcmfmac internal so I don't know why it does
>> that, but that means the buffer cannot be modified. If you need to
>> modify the ssid should you make a copy of it?
>
> In brcmf_parse_tlvs(const void *buf, int buflen, uint key),
> it will find the key in buf and return the address of this key, as the 
> return pointer.
> This function don't intend caller to modify content of buf in most 
> cases, so it defines a const return pointer.
> But in this case, it just need to do it, so I need the typecast here.
>
Do you accept the typecast here?
Thank you.

>> -- 
>> https://patchwork.kernel.org/project/linux-wireless/list/
>>
>> https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches 
>>
>
Arend Van Spriel Oct. 10, 2022, 9:23 a.m. UTC | #4 
On 10/6/2022 10:34 AM, Lin Ian (CSSITB CSS ICW SW WFS / EE) wrote:
> 
> 
> On 10/5/2022 6:22 PM, Arend Van Spriel wrote:
>> Caution: This e-mail originated outside Infineon Technologies. Do not 
>> click on links or open attachments unless you validate it is 
>> safe<https://intranet-content.infineon.com/explore/aboutinfineon/rules/informationsecurity/ug/SocialEngineering/Pages/SocialEngineeringElements_en.aspx>. 
>>
>>
>>
>>
>> On 10/5/2022 9:58 AM, Kalle Valo wrote:
>>> "Lin Ian (CSSITB CSS ICW SW WFS / EE)" <ian.lin@infineon.com> writes:
>>>
>>>> On 9/28/2022 6:07 PM, Lin Ian (CSSITB CSS ICW SW WFS / EE) wrote:
>>>>>
>>>>>
>>>>> On 9/28/2022 2:38 PM, Kalle Valo wrote:
>>>>>> Caution: This e-mail originated outside Infineon Technologies. Do
>>>>>> not click on links or open attachments unless you validate it is
>>>>>> safe<https://intranet-content.infineon.com/explore/aboutinfineon/rules/informationsecurity/ug/SocialEngineering/Pages/SocialEngineeringElements_en.aspx>. 
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Ian Lin <ian.lin@infineon.com> writes:
>>>>>>
>>>>>>> From: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>>>>>>>
>>>>>>> cfg80211 layer on DUT STA is disconnecting ongoing connection
>>>>>>> attempt after
>>>>>>> receiving association response, because cfg80211 layer does not
>>>>>>> have valid
>>>>>>> AP bss information. On association response event, brcmfmac
>>>>>>> communicates
>>>>>>> the AP bss information to cfg80211 layer, but SSID seem to be
>>>>>>> empty in AP
>>>>>>> bss information, and cfg80211 layer prints kernel warning and then
>>>>>>> disconnects the ongoing connection attempt.
>>>>>>>
>>>>>>> SSID is empty in SSID IE, but 'bi->SSID' contains a valid SSID, so
>>>>>>> updating the SSID for hidden AP while informing its bss information
>>>>>>> to cfg80211 layer.
>>>>>>>
>>>>>>> Signed-off-by: Syed Rafiuddeen <syed.rafiuddeen@cypress.com>
>>>>>>> Signed-off-by: Chung-Hsien Hsu <chung-hsien.hsu@infineon.com>
>>>>>>> Signed-off-by: Chi-hsien Lin <chi-hsien.lin@infineon.com>
>>>>>>> Signed-off-by: Ian Lin <ian.lin@infineon.com>
>>>>>>> ---
>>>>>>> .../net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c | 8
>>>>>>> ++++++++
>>>>>>>    1 file changed, 8 insertions(+)
>>>>>>>
>>>>>>> diff --git
>>>>>>> a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>>>>>> b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>>>>>> index 6c37da42e61b..3560afe0ccfe 100644
>>>>>>> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>>>>>> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
>>>>>>> @@ -3003,6 +3003,7 @@ static s32 brcmf_inform_single_bss(struct
>>>>>>> brcmf_cfg80211_info *cfg,
>>>>>>>         u8 *notify_ie;
>>>>>>>         size_t notify_ielen;
>>>>>>>         struct cfg80211_inform_bss bss_data = {};
>>>>>>> +     struct brcmf_tlv *ssid;
>>>>>>>
>>>>>>>         if (le32_to_cpu(bi->length) > WL_BSS_INFO_MAX) {
>>>>>>>                 bphy_err(drvr, "Bss info is larger than buffer.
>>>>>>> Discarding\n");
>>>>>>> @@ -3032,6 +3033,13 @@ static s32 brcmf_inform_single_bss(struct
>>>>>>> brcmf_cfg80211_info *cfg,
>>>>>>>         notify_ielen = le32_to_cpu(bi->ie_length);
>>>>>>>         bss_data.signal = (s16)le16_to_cpu(bi->RSSI) * 100;
>>>>>>>
>>>>>>> +     ssid = (struct brcmf_tlv *)
>>>>>>> +             brcmf_parse_tlvs(notify_ie, notify_ielen, 
>>>>>>> WLAN_EID_SSID);
>>>>>> This still casts away the const. For some reason brcmf_parse_tlvs()
>>>>>> takes a const buffer:
>>>>>>
>>>>>> static const struct brcmf_tlv *brcmf_parse_tlvs(const void *buf,
>>>>>> int buflen, uint key)
>>>>>>
>>>>>> I'm not familiar with brcmfmac internal so I don't know why it does
>>>>>> that, but that means the buffer cannot be modified. If you need to
>>>>>> modify the ssid should you make a copy of it?
>>>>>
>>>>> In brcmf_parse_tlvs(const void *buf, int buflen, uint key),
>>>>> it will find the key in buf and return the address of this key, as
>>>>> the return pointer.
>>>>> This function don't intend caller to modify content of buf in most
>>>>> cases, so it defines a const return pointer.
>>>>> But in this case, it just need to do it, so I need the typecast here.
>>>>
>>>> Do you accept the typecast here?
>>>
>>> To me writing a const data is wrong. IIRC it was something like six
>>> months ago that rtw89 was also writing a const variable and it caused a
>>> crash because the memory was in a read-only area (or something like
>>> that).
>>
>> So how should this be solved. The pointer returned by the
>> brcmf_parse_tlvs() function is pointing inside the buffer passed as
>> first argument, ie. notify_ie which is non-const. So it is perfectly
>> safe to do the cast as suggested here. We could do a pointer-arithmetic
>> dance here to avoid the cast, but that only make things more obscure.
>>
> I may calculate the offset and cast on notify_ie.
> Than the code will be like this, is that ok?
> 
> u8 *notify_ie;
> int ssid_offset;
> ssid_offset = brcmf_parse_tlvs(notify_ie, notify_ielen, WLAN_EID_SSID) - 
> (struct brcmf_tlv *)notify_ie;
> memcpy(notify_ie + ssid_offset + offsetof(struct brcmf_tlv, data), 
> bi->SSID, bi->SSID_len);

Hi Ian,

I am strarting to doubt the entire patch now. The notify_ie contains 
beacon/proberesp elements in TLV format. So how can this work by simply 
copying an SSID into the notify_ie buffer. The patch says: "SSID is 
empty in SSID IE" so I would conclude the element to be two bytes, ie. T 
= WLAN_EID_SSID and L=0 (no V). If this is true, it means the SSID will 
overwrite/corrupt the TLVs located after the SSID. Also the length is 
not corrected. Maybe I am mistaken and the SSID element for the 
Hidden-SSID scenarion is different. Will check the 802.11 spec.

Regards,
Arend
Johannes Berg Oct. 10, 2022, 9:41 a.m. UTC | #5 
On Mon, 2022-10-10 at 11:23 +0200, Arend Van Spriel wrote:
>  Maybe I am mistaken and the SSID element for the 
> Hidden-SSID scenarion is different. Will check the 802.11 spec.
> 

hidden SSID isn't in the spec, but these days is typically implemented
by a zero-size SSID element:

 00 00

In the past, sometimes it would also be done by some APs with a zeroed
SSID element, e.g. for the SSID "myAP" you could see

 00 04 00 00 00 00

instead.

johannes
Arend Van Spriel Oct. 10, 2022, 11:18 a.m. UTC | #6 
On 10/10/2022 11:41 AM, Johannes Berg wrote:
> On Mon, 2022-10-10 at 11:23 +0200, Arend Van Spriel wrote:
>>   Maybe I am mistaken and the SSID element for the
>> Hidden-SSID scenarion is different. Will check the 802.11 spec.
>>
> 
> hidden SSID isn't in the spec, but these days is typically implemented
> by a zero-size SSID element:
> 
>   00 00
> 
> In the past, sometimes it would also be done by some APs with a zeroed
> SSID element, e.g. for the SSID "myAP" you could see
> 
>   00 04 00 00 00 00
> 
> instead.

Thanks, Johannes

If both species are out there than it still means blindly copying things 
into the TLV buffer is not the proper thing to do here.

Regards,
Arend

> johannes
diff mbox series

Patch

diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
index 6c37da42e61b..3560afe0ccfe 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
@@ -3003,6 +3003,7 @@  static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
 	u8 *notify_ie;
 	size_t notify_ielen;
 	struct cfg80211_inform_bss bss_data = {};
+	struct brcmf_tlv *ssid;
 
 	if (le32_to_cpu(bi->length) > WL_BSS_INFO_MAX) {
 		bphy_err(drvr, "Bss info is larger than buffer. Discarding\n");
@@ -3032,6 +3033,13 @@  static s32 brcmf_inform_single_bss(struct brcmf_cfg80211_info *cfg,
 	notify_ielen = le32_to_cpu(bi->ie_length);
 	bss_data.signal = (s16)le16_to_cpu(bi->RSSI) * 100;
 
+	ssid = (struct brcmf_tlv *)
+		brcmf_parse_tlvs(notify_ie, notify_ielen, WLAN_EID_SSID);
+	if (ssid && ssid->data[0] == '\0' && ssid->len == bi->SSID_len) {
+		/* Update SSID for hidden AP */
+		memcpy(ssid->data, bi->SSID, bi->SSID_len);
+	}
+
 	brcmf_dbg(CONN, "bssid: %pM\n", bi->BSSID);
 	brcmf_dbg(CONN, "Channel: %d(%d)\n", channel, freq);
 	brcmf_dbg(CONN, "Capability: %X\n", notify_capability);