Greg, 54 Mbps would work for me. But I didn't see a LICENSE file or any text about it in the README file either in that repo. So I'm assuming this is not a driver that can be used. Or else it needs some work with Microchip to get them to clarify the license.
Re: using a BSD driver, I didn't get what you mean about bringing in 3rd party code– it's BSD licensed, my understanding is that it can be brought in. Is that wrong? I'll check out the driver you're enclosing... yes, I would need some other stuff besides the driver, but if I can get the userland stuff from BSD also... maybe that would work... -adam On Thu, Jun 11, 2020 at 2:54 PM Gregory Nutt <spudan...@gmail.com> wrote: > > That looks like a cool module. The issue for me is that I need high speed > Wifi connectivity (20 Mbps+, ideally 50Mbps+) for my upcoming projects. > SPI/UART modules like this one usually have much lower throughput. > > The WiFi interface is SDIO. So we would need your SD/MMC driver (plus a > couple of SDIO extensions). > > SPI and I2C are not used on the WiFi chip. The WiFi component is SDIO and > Bluetooth component is UART So I don't see any Wifi performance limitation > do to interface constraints. > > Radio performance is addressed in Table 4.4 here for various conditions > http://ww1.microchip.com/downloads/en/DeviceDoc/70005327A.pdf Under > certain conditions it can do 54Mbps > > This module does have a 50Mhz SDIO interface, so if it can use 4 bit mode, > it could work. However, the ATWIL3000 Linux driver on Github > <https://github.com/linux4wilc/driver> does not appear to have an open > license– all the code is marked "All rights reserved". Is this the driver > you were referring to? > > All of the NuttX files with BSD headers say "All rights reserved" too. > That is usually part of all copyright statements. You should look at the > LICENSE file if there is one. Bringing 3rd party code into NuttX is > basically impossible with the current state of things unless you can NXP to > provide you some written confirmation that they approve of you taking a > snapshot of the code. They will not sign an SGA or ICLA but if they are > like NXP, they will freely give consent otherwise > > And we can always download it on the fly at build time as we have done for > other 3rd party code. > > I am looking at porting a USB Wifi driver from FreeBSD... The thing about > porting a USB driver from FreeBSD that attracts me is that I could at least > have access to a whole family of fast Wifi hardware that way. And in the > best case scenario, if I can create a compatibility layer, then we could > use a variety of FreeBSD wifi drivers. > > There is a RTL8187 Wifi USB driver for NuttX that you can look at here > attached. This was ported from Linux years ago and is GPL so you can't > bring it into NuttX, but you can look at it. It is kind of useless with no > ieee802.11 SoftMAC stack anyway. > > > -- Adam Feuer <a...@starcat.io>
