Mesh on BBBW

benjam...@gmail.com

unread,

Mar 1, 2017, 5:05:57 AM3/1/17

to BeagleBoard

I am trying to setup the BBB Wireless for mesh functionality, however i noticed that all the files for the wilink 8 required are not even built.

I am doing the tutorial from

http://www.ti.com/lit/an/swaa166/swaa166.pdf

and realise that the files are missing from

/usr/share/wl18xx/

can anyone please advise on how to proceed thanks.

i've tried to build the wilink8 using

http://processors.wiki.ti.com/index.php/WiLink8_Linux_Getting_Started_Guide

but it is really complex and there are tons of errors on the way, has anyone succesfully build it?

drhun...@gmail.com

unread,

Mar 1, 2017, 10:17:22 AM3/1/17

to BeagleBoard, benjam...@gmail.com

Hi Benjamin,

First some background before we go into the details.

There are two types of wilink8 driver deliveries

- TI driver release. R8.7 is the latest. This is the full feature driver release that includes all features including mesh. It is built via the build script at the link you mentioned. It uses backports to build the driver source (in a 4.4.8 kernel) against your kernel.

- mainline (K4.1 and above). This is a subset of the TI driver release including the standard features accepted by the community as non-proprietary. Mesh comes into this category which is why you don't see it. The mainline driver was mostly R8.6 - the full list of differences between mainline and R8.6 is at http://processors.wiki.ti.com/index.php/WiLink8_Release_Notes/R8.6#Change_log_from_mainline_versions . I'm not sure if there is any plan to push mesh to mainline.

And now to your question.

The build script is quite simple to use as long as the setup-env file points to your toolchain, kernel and filesystem.

The only problem you may be facing is that if your kernel is newer than 4.4.8 then the build may fail as backport will have nothing to do as it was designed to work back to older kernels.

If you paste your build error I may be able to identify if it is a setup-env problem which should be simple to fix.

Iain

Iain Hunter

unread,

Mar 1, 2017, 10:51:14 AM3/1/17

to beagl...@googlegroups.com

A follow up on this one.

- If you are just missing the user space script to setup the mesh you can get them from https://git.ti.com/wilink8-wlan/wl18xx-target-scripts/trees/R8.7_SP1 - they are installed as part of the build script, otherwise clone them independently

- The mesh patches have been pushed to mainline, the accepted ones should be around K4.9.

Iain

--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/gG0MQ3DzRSk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to beagleboard+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/cf46bdc4-8057-4fa1-bf46-6e0eb21f2cdc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Iain Hunter

unread,

Mar 8, 2017, 11:45:38 AM3/8/17

to Benjamin Hon, beagl...@googlegroups.com, Nicholas Wong

Yes I meant BBBW. You should use the open source kernel device tree as the basis.

If you are building the TI SDK then you can use the Makefile in the base directory of Processor SDK3 to build only the dtb files. You will need to add the dts file name you have added for the BBBW to the linux-dtbs rule in the Makefile. Then run:

make linux-dtbs

Iain

On 8 March 2017 at 14:32, Benjamin Hon <benjam...@gmail.com> wrote:

Hi Dr Hunter,

when you say "The only change you should need is to the device tree file to match BBGW use of mmc3"

do you mean BBBW instead of BBBGW? or if i get you correctly, we should create a new patch based on the BBGW device tree but just change the mmc2 to mmc3?

thanks

On Fri, Mar 3, 2017 at 12:05 AM, Iain Hunter <drhun...@gmail.com> wrote:
Hi Nicolas,
I'm assuming you are looking at http://processors.wiki.ti.com/index.php/Add-wilink8-to-processor-sdk-3.0 for instructions to build R8.7 on top of TI Processor SDK 3.0.
The simple answer is that there is no BBB_PLATFORM for BBBW support on this page as the BBBW was not available at the time it was written.

The only change you should need is to the device tree file to match BBGW use of mmc3 (in device tree, mmc2 in hardware) and gpios. Simply generate a patch and apply within a new BBB_PLATFORM option.
I do not have a BBBW to validate any changes to the device tree file. Looking at latest beagle bone debian build you can find the device tree it uses at https://github.com/RobertCNelson/bb-kernel/blob/am33x-v4.9/patches/soc/ti/bbbw/0001-ARM-dts-add-am335x-boneblack-wireless.dtb.patch

Hope that helps a little,
Iain

On 2 March 2017 at 06:49, Nicholas Wong <nichola...@hotmail.com> wrote:

Hi Iain,

I'm Nicholas, a colleague of Benjamin. May I ask if I'm building the SDK for BeagleBone Black Wireless, which BBB_PLATFORM should I build for when I run build_wilink.sh?

Is it 2 (EVM)? I don't see an option for BBGW in the Manual.

Please advise. Thanks.

Regards,
Nicholas

On 1 Mar 2017, 5:22 PM +0800, drhun...@gmail.com, wrote:

Hi Benjamin,

First some background before we go into the details.

There are two types of wilink8 driver deliveries
- TI driver release. R8.7 is the latest. This is the full feature driver release that includes all features including mesh. It is built via the build script at the link you mentioned. It uses backports to build the driver source (in a 4.4.8 kernel) against your kernel.

- mainline (K4.1 and above). This is a subset of the TI driver release including the standard features accepted by the community as non-proprietary. Mesh comes into this category which is why you don't see it. The mainline driver was mostly R8.6 - the full list of differences between mainline and R8.6 is at processors.wiki.ti.com/index.php/WiLink8_Release_Notes/R8.6#Change_log_from_mainline_versions . I'm not sure if there is any plan to push mesh to mainline.

jclem...@gmail.com

unread,

Oct 26, 2017, 4:57:00 PM10/26/17

to BeagleBoard

Hi Iain

I'm trying to configure ad-hoc mode on BBBW, and I am follow instructions http://processors.wiki.ti.com/index.php/WiLink8_Linux_Getting_Started_Guide and here http://www.ti.com/lit/an/swaa166/swaa166.pdf too.

I solved many problem on the way, but I reached a point that I get this error:

________________________________________________________________________________________

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c: In function ‘backport_tso_build_hdr’:

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:103:6: error: dereferencing pointer to incomplete type

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:106:40: error: dereferencing pointer to incomplete type

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:110:7: error: dereferencing pointer to incomplete type

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:111:7: error: dereferencing pointer to incomplete type

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:112:7: error: dereferencing pointer to incomplete type

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c: In function ‘backport_tso_start’:

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:140:2: error: implicit declaration of function ‘tcp_hdr’ [-Werror=implicit-function-declaration]

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:140:17: error: invalid type argument of ‘->’ (have ‘int’)

cc1: some warnings being treated as errors

make[6]: *** [/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.o] Error 1

make[5]: *** [/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat] Error 2

make[5]: *** Waiting for unfinished jobs....

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/nl80211.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/mlme.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/ibss.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/sme.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/chan.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/ethtool.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/mesh.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/ap.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/trace.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/ocb.o

CC [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/debugfs.o

LD [M] /home/sensorfield/wl8-build/build-utilites/src/compat_wireless/net/wireless/cfg80211.o

make[4]: *** [_module_/home/sensorfield/wl8-build/build-utilites/src/compat_wireless] Error 2

make[3]: *** [modules] Error 2

make[2]: *** [modules] Error 2

make[1]: *** [modules] Error 2

make: *** [default] Error 2

****** ERROR 0 *******

________________________________________________________________________________________

Can you please advise about it?

Thanks

Jose

To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.

Iain Hunter

unread,

Oct 27, 2017, 5:04:01 PM10/27/17

to beagl...@googlegroups.com

Hi Jose,

I’ve never used the build script with compat/backports on a 4.4 kernel. Only suggestion I can make is to work out what the actual build error is. Ie why is there a problem here

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c: In function ‘backport_tso_build_hdr’:

/home/sensorfield/wl8-build/build-utilites/src/compat_wireless/compat/backport-4.4.c:103:6: error: dereferencing pointer to incomplete type

It may be a change that occurred at some point in 4.4.x that the backports you are using is not providing a solution for.

Iain

Sent from Mail for Windows 10

To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/26bb4c85-6c8d-40f6-8dc7-b9de5b5bd465%40googlegroups.com.

Reply all

Reply to author

Forward