On Fri, Sep 10, 2010 at 11:32 AM, J. L. <firstname.lastname@example.org> wrote:
use and is not reliable. The best luck I have had so far is using the
driver the OLPC has developed for the marvell chip. But its not fully
designed for our setup so there are other issues that I have no clue
BTW, a number of people here on the Gumstix list have contributed patches to improve the driver. The latest linux kernel is at:
This is an up-to-date (as of last Friday anyway) wireless-testing kernel clone. Note that of course this isn't a production kernel, but wireless-testing + my changes for the libertas_tf driver.
Note that the standard firmware isn't the correct one either, you need the thinfirmware firmware from OLPC:
The libertas_tf driver won't work with the standard firmware, but the libertas driver _will_ still work with the thin-firmware.
To me the adapter does not perform up to par but thats me, as I have no issues with any other USB adapter I have run,
It would probably be more useful to everyone if people gave quantified measurements. Specific throughput measurements from iperf or other similar quantified measurements with stated specific test configurations would be helpful. Information as to exactly what is happening, preferably supported with wireshark traces. I assure you those on the libertas list, or Marvell or other people would be more likely to fix things if there were specific reports.