WL1271L AP + BT coexistence regression with R5.SP7.01

Jesse Gilles

Other Parts Discussed in Thread: WL1271, WL1837

Hi all,

We are using the Murata LBEE5ZSTNC-523 module that is WL1271-based. It is on a custom board running Linux 2.6.39 with an Atmel SAM9G25.

We recently upgraded the wl12xx driver and firmware to the ol_R5.SP7.01 tag release on github.com/Ti-OpenLink. We were initially using the R4.SP2 release obtained from the TI wiki.

After the upgrade, we have noticed that any rfcomm bluetooth activity severely disrupts Wifi when running as an access point with hostapd. We are using a vanilla version of hostapd 1.0 and bluez 4.98.

Basic steps to reproduce:

* run hostapd on wlan0

* connect a wifi client to AP

* initiate a repeating ping from wifi client to AP

* run rfcomm and bind a bluetooth device to /dev/rfcomm0

* open /dev/rfcomm0 and send/receive data from BT device

After opening /dev/rfcomm0, Wifi traffic on the AP is noticeably disrupted. Pings start timing out and there is a lot of packet loss. New clients also can't seem to connect to the AP.

If /dev/rfcomm0 is closed and the BT connection is terminated, wifi traffic resumes normal performance after a few seconds and new clients can connect again.

The R5.SP7.01 wl12xx drivers and firmware have been much more stable than the R4.SP2 release, so we would really like to upgrade, but we can't get BT and AP to work at the same time. With the R4.SP2 firmware/drivers, we could have a BT client connection through rfcomm and it would not disrupt wifi AP traffic.

I have also tried the latest BTS file from github.com/TI-ECS. This didn't seem to make any difference. Is there a specific version of the Bluetooth firmware that is meant to be used in conjunction with the R5.SP7.01 release? I can't find any release notes at all on the various revisions to the firmware. The latest commit mentions service pack 2.14 for TIInit_7.6.15.bts -- how do I know what has changed? The lack of information and release notes is frustrating.

Has anyone else noticed this issue of AP + BT problems with the latest firmware/drivers? Since the same scenario worked fine with the R4.SP2 firmware/drivers, I know it is possible.

Suggestions?

Thanks,

Jesse

over 11 years ago

0 Joseph Gigi over 11 years ago

TI__Guru* 85400 points

Hi Jesse,

I will check this and get back to you...

Regards,
Gigi Joseph.

0 Jesse Gilles over 11 years ago in reply to Joseph Gigi

Prodigy 50 points

Hi, any update?

Thanks,

Jesse

0 Jesse Gilles over 11 years ago in reply to Jesse Gilles

Prodigy 50 points

I performed the same test with R5.SP4.01 and that release seems to work fine. I see no conflict between BT and Wifi AP. Certainly there is some regression between R5.SP4.01 and R5.SP7.01.

Jesse

0 Joseph Gigi over 11 years ago in reply to Jesse Gilles

TI__Guru* 85400 points

Hi Jesse,

There has been about 9 drops between R4 SP2 to R5 SP7. Have you happened to test any of the intermediate drops? Its very strange that an older version worked, and a newer version did not...

Also, you mentioned that you are using hostapd v "1.0" vanilla version. This is not the recommended version. Any particular reason for this?
Can you move to TI's hostapd version and check? You can find TI's version at: https://github.com/TI-OpenLink/hostap.git (git version tag: "ol_R5.SP7.01"). This is "2.0-devel" version.

Regards,
Gigi Joseph.

0 Jesse Gilles over 11 years ago in reply to Joseph Gigi

Prodigy 50 points

I think maybe you misread my last post -- we have tried version R5.SP4 and it is working much better than SP7. So it would seem something changed between SP4 and SP7.

We are using v1.0 of hostapd because when development started I was not aware that TI had a custom version. We so far have not had any issues that seem related to hostapd itself, so we have not changed it. Last I checked, I could not find any release notes/summary of what TI has changed/fixed in their custom hostapd. Can you provide any such info? Also, if TI is making fixes, are they getting merged back into the main hostapd project?

Thanks,

Jesse

0 Joseph Gigi over 11 years ago in reply to Jesse Gilles

TI__Guru* 85400 points

Hi Jesse,

Some of the changes that TI makes to hostapd are custom (like smart config). So it won't get merged back. I suggest that you once test with TI's hostapd version and let me know the result because this is the version that we officially use, and test with...

Regards,
Gigi Joseph.

0 Jesse Gilles over 11 years ago in reply to Joseph Gigi

Prodigy 50 points

Ok, I did the same test again using R5.SP7.01 wl12xx and wl1271 firmware, but this time with TI's hostapd on github (tag ol_R5.SP7.01) and I still see the same behavior as with hostapd 1.0.

Thanks,

Jesse

0 Ryan Erickson over 11 years ago in reply to Jesse Gilles

Intellectual 625 points

Has this issue been resolved? I have a customer who is having the same issues and is looking for a resolution.

0 Elwood Friesen over 9 years ago in reply to Ryan Erickson

Prodigy 130 points

Was this issue ever addressed and fixed? Did TI have a solution that wasn't published here? We are currently experiencing Bluetooth coexistence issues with the WL1837.

e2e.ti.com/.../517538

0 Joseph Gigi over 9 years ago in reply to Elwood Friesen

TI__Guru* 85400 points

Hi,

Yes, this issue was resolved with the R5.SP8 release. Your problem is different and it's on WL8 .
It will be handled in the other thread.

Regards,
Gigi Jospeh.

Wi-Fi

Wi-Fi forum

WL1271L AP + BT coexistence regression with R5.SP7.01