This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM3517 and WL1271L WiFi chip issue

Other Parts Discussed in Thread: WL1271, AM3517

The issue is a driver in a system that incorporates the WL1271L WiFi chip.  We have been unable to get it working on our kernel which is Linux 3.4 with real-time patches (-rt31).  We get the same results on an earlier kernel 2.6.33.20 (-rt31). with real time extensions. We get repeated kernel hangs.  The device is configured as a an Access Point. 

There is a patch on the Real Time mailing list posted by Grégoire Gentil  which worked on an OMAP4 ARM9.  The patch fixed a problem that manifested itself the same way ours did.  However, when we incorporated the patch it did not fix the problem.

Hardware is a Logic PD module, part number SOMAM3517-10-1780RJCR.
WiFi module is from RFM, part number WLS1271L.  The documentation page states the above uses the WL1271L SoC from TI.

If you have a version of the WL1271 driver that runs with real time extensions, please send us the Linux kernel configuration file.  Are there any options that must be enabled or disabled for this driver to work.  We could send you our Linux kernel config file.

  • Hy Paul,

    have you tried to build the new kernel based on the 3.4 without the real time extensions?

    Maybe there is a bad interacting with the network kernel driver.

    Greetings Bert.

  • Hi Paul,

    I wanted to reflect the comments I received from the Sitara business unit here regarding their support policy for issues like this.  From the BU:  Outside of TI support will be required for RT-patched linux kernel issues.  Additionally, TI has not supported the WL1271 on the AM3517 platform.

    Sitara team:

    If a different RTOS was utilized (QNX, for example) would you be in a position to provide more support on this issue?

     

    Thanks everyone,

    Walter

     

  • Thanks Walter,

    I am just curious about why the RT extensions have caused TI to take an official position.  Usually, this would come based on experience, or rather one or more negative experiences.  I am looking down the line for this program, as well as future development projects.

    More mainstream OS choices might be VxWorks, QNX, or the Mentor Embedded products.  None of these products have an off the shelf solution for me either, unfortunately.  The AM3517 is supported however in at least two of the above (including the Logic PD board).

    Regards - Paul

  • To my knowledge, the team did not try this without the RT extensions.  Once it was determined that this was needed, we did not look back.  The maker of the module noted that this was working with an earlier version of Linux, non real-time of course, however I did not try this build.

    Thanks for posting, though, it is a good thought.

    Paul

  • Paul,

    I've gotten some comments back from the Sitara group on why RT-extended Linux is unsupported.

    The current Linux drivers were not designed and structured to handle the real-time environment.  While in most circumstances the kernel will be stable with RT patches, problems tend to occur when the system becomes stressed.  Based upon past experiences, this may not be a stress that is seen until the system is deployed to the field.  This is the reason for the official no-support position.

    With regard to an alternative RTOS such as QNX, support for QNX would come from QNX and not from TI.  If this is a route you are interested in, I can put you in touch with our third party liason who works with the companies that own support for their respective operating system.

    Regards,
    Walter