Loud ramblings of a Software Artisan

Wednesday 19 January 2005

Show that you need Open Source drivers for Wifi.

Given my recent problem with a Netgear Wifi adapter and that the SMC worked, I thought it would be a good idea to send feedback to Netgear so that they understand that they lose customers. I encourage anyone that have returned a product that did not work with an Open Source driver to do the same to let these people know.

Manufacturer of devices, those that buy thhe chips, really have the power to make this change and require from the chip supplier to provide the source code under an Open Source license.

Here is a verbatim copy of my e-mail sent Monday January 17th to Netgear Customer Service:



Hi,

On January 5th 2005, I bought a WG511 Wireless 802.11g PCMCIA adapter to run on my Apple PowerBook laptop running Linux. The choice was made because this card was reported to be compatible with the prism54 open source driver for Linux (http://www.prism54.org/) . I unpacked the card, inserted it, and after a few seconds discovered that it was not compatible. The reason is that, without changing the product name, you changed the chipset, making it incompatible.

For that single reason, I returned the product to Future Shop in Montreal, where I bought it, for immediate refund, and went to buy a SMC card, that works perfectly well. SMC got my money, you didn't. In fact, since it was supported, I bought 2 of them, one being for a coworker in the same situation.

I just wanted to bring you attention that I won't buy a Netgear product that does not work with an Open Source driver on Linux, and that I'm not the only one being in that case.

Sincerely,

Hub

I haven't had a reply so far.

Recent work from OpenBSD team to open source drivers seems to be efficient as they got some rights to redistribute firmwares as part of OpenBSD and that they got completely reverse engineered the Atheros HAL code, unlike the Madwifi project which seems to be willing to keep it closed (making the Atheros driver not available in the Linux kernel, for a good reason). I hope someone ports the OpenBSD HAL code to Linux, as Atheros based card are quite common.

I'm myself a strong opponent to binary only drivers on Linux, and I believe that the ndiswrapper project harm Linux more than it make good: if people can use ndiswrapper to run Windows drivers on their Linux/x86, then they'll buy cards are not supported on Linux, and the manufacturer won't be inclined to provide an Open Source driver or even a Linux driver, claiming that this is not needed since they can run the Windows driver. FreeBSD seems to have the NDIS approach too... Driver support is just money talks. Show manufacturers that you give money to those that supports your OS. There are chipset for 802.11g that works great with Open Source drivers: old Prism "G" (newer does not, hence my troubles with the Netgear) using prism54, Ralink, Intel, etc. See A survey on Linux and Wifi for comprehensive information.

There is still a long run until that kind of problem get solved, and given that chip manufacturer make their product more and more dumb, requiring more and more complex drivers (like in the new Prism Frisbee where the 802.11 is no longer in the chip), and more and more firmware (because it cost less to store the firmware on the hard drive than in a ROM/Flash on the chip or device).

Wednesday 5 January 2005

Found a Wifi card

I managed to find a SMC2835W wifi 802.11g card for CDN69. Works well with Linux using the prism54 driver, that comes with the kernel. At least they did not change their chipset like Netgear did by manufacturing in China instead of Taiwan. This one use a Prism GT.

Update (01/09): I recently discovered that there are at least 2 versions of the card and that I got lucky to get v1. Be careful. The other one apparently do not work with Linux (unless you want to use ndiswrapper which is not a good idea).

Wifi adapter, warning

Today I bought a Wifi 802.11g PCMCIA adapter for CDN80 + tx. It is a Netgear WG511 and I checked that is was supported by the prism54 driver.

Big mistake. It does not work. After downloading the firmware, it ends up with errors. The side label says WG511CN and Made in China, and under the card (after opening the box), it says v3. Yes, that does matter. I'm gonna have to find a Made in Taiwan also know as v2. The friends on #prism54 did confirm this.

For reference, the FCC ID is PY3WG511-F. You can lookup FCC ID, this is really informative.

Saturday 1 January 2005

GPL Violation

Greg KH pointed out on 18 Nov 2004 in the Linux Kernel mailing list a GPL Violation by Silicon Laboratories. See also his blog entry

In talking with the company, they insist that they will not release the source code to this module, and they claim that they are not infringing on any rights by not doing so. I claim that this is not true, as to write a usb to serial driver for Linux you have to use the drivers/usb/serial/usb-serial.h header file which is specifically licensed under the GPL v2. This file contains inline functions and structures that all usb-serial drivers need to use in order to work properly.

In short, there's no way you can write a Linux usb-serial driver, that uses the usbserial interface, without it being a derived work of other, GPL only code.

So, they are in violation, so what. Well, I can't do much about this (due to my employer's rules about suing companies). But I can do my best to spread the word that the CP2102 device is not supported on Linux, and should be avoided at all costs by anyone considering such a device in a future design. I encourage everyone else to help spread this information too.

Word spread. I had to talk about that as well. I don't like blatant license violations.