WPA2

[Tom`]

Re: WPA

This thread should be merged with this one, in which the situation is explained a bit further.

Done.

Don't know whether "Hardware" is the place for this, though, since it has software implications as well...

 

[TJSomething]

But will it have a good GUI for WPA2-Enterprise?

 

[javaJake]

Re: WPA

Don't know whether "Hardware" is the place for this, though, since it has software implications as well...

Then I'd stash it away in the good ol' General Discussion fall-back.

 

[Capn_Fish]

There is good and bad news with the wifi/BT chip. The bad news is that the driver is not open, and never will be. In fact, even the make and model are under NDA until the Pandora ships. There were very few options for wifi modules that fit the specs we needed, and none of the others were any better.

The good news is that the manufacturer is willing to (likely already has) provide the driver source to the devs in exchange for keeping it confidential. So that while we don't have access to it, Craig and Michael do. WPA encryption may well be possible with existing software (they may not be allowed to say yes or no until it ships), but even if it isn't, they have the capability to make it happen.

So, in short, we're stuck with whatever kernel versions Craig and Michael decide to build and release? That's mildly disappointing, though I suppose they'll do a good job of staying at least somewhat up-to-date.

 

[Chip]

So, in short, we're stuck with whatever kernel versions Craig and Michael decide to build and release? That's mildly disappointing, though I suppose they'll do a good job of staying at least somewhat up-to-date.

Not at all. As is common with closed Linux drivers, the driver itself is supplied as a binary blob, but there is an open source shim which acts as a go-between for the kernel and the blob. This shim can be recompiled with any kernel version you like.

 

[Capn_Fish]

So, in short, we're stuck with whatever kernel versions Craig and Michael decide to build and release? That's mildly disappointing, though I suppose they'll do a good job of staying at least somewhat up-to-date.

Not at all. As is common with closed Linux drivers, the driver itself is supplied as a binary blob, but there is an open source shim which acts as a go-between for the kernel and the blob. This shim can be recompiled with any kernel version you like.

Awesome! You just made my day approx. 10x brighter.

 

[parag0n]

The wifi chip that is going to be used is the WL1251 from TI.

It currently ships with a binary blob driver, but there is apparently a community growing around providing a free software alternative driver.

 

[Tor]

It'll be very interesting to see how sensitive it is. The N8x0 has incredibly high sensitivity and very low power consumption. According to TI's docu that 1251 has low power consumption. It remains to see about the sensitivity.. ah, November! Waiting for you.

 

[TaG]

The wifi chip that is going to be used is the WL1251 from TI.

Is that official? If so I must have missed it, I couldn't find it on the main site, either forum or wiki. - do you have a source?

 

[parag0n]

Its not official, but it was mentioned in IRC a while ago, it may just be a rumour.

But it does fit all the information we have been given (supports b, g e, i, d, k, bluetooth built in, ultra low power, there are versions with bluetooth), and is a TI product that they are pushing for use with the OMAP3.

Plus it would likely be covered by the same NDA as the main Chipset!

 

[nitsnipe]

I run full Debian (armel) on my Nokia N800 which has an OMAP2 chi[ and I have absolutely no problems wirelessly connecting to WEP, WPA, WPA2. Heck, on enterprise wifi networks(like in offices and some universities) I can even authenticate on RADIUS server, use openvpn, and even log in on domains. Don't know if Microsofts IPSec would work though.

Anyways, don't worry about WPA2, it shouldn't be a problem at all. If the older OMAP2 can do it then OMAP3 shouldn't have any problems hardware wise. The drivers shouldn't be a problem, I don't know if we will get just binaries or fully open-source drivers but i couldn't care less as long as it works.

 

[nitsnipe]

It'll be very interesting to see how sensitive it is. The N8x0 has incredibly high sensitivity and very low power consumption. According to TI's docu that 1251 has low power consumption. It remains to see about the sensitivity.. ah, November! Waiting for you.

That is so true, I wish omap3 carries the same wireless chip as the one in the omap2 N8x0. Lol on low power the wireless antenna consumes JUST 10mW when in use. If I up it to 100mW and turn off the powersavings, I can pick up like 20 access points even though I'm in the middle of a park in the suburbs and the nearest house is at least 50m away.

 

[Kanji]

Well the G1 uses the WL1251 so if that source is as easily confused as I am then he may have misremembered information. I would have to buy one of these if it could h.264 high 5.1 480p and packet injection. That would be some toy.

http://android.git.kernel.org/?p=platfo ... ;a=summary