The return of the Philips driver? [LWN.net]

When the Philips webcam driver maintainer requested that driver's removal, the kernel developers complied. The fact remains, however, that the code for the core driver was released under the GPL; it remains out there for those who wish to make use of it. The proprietary "pwcx" decompression code is another story; it has been withdrawn and is unlikely to return. But the GPL code could, perhaps, come back.

The original maintainer questions the value of the GPL-only code. Without the decompression module, the camera can only be used in a very low-resolution mode. There are a couple of reasons for wanting that code back, however. One of the more interesting ones was posted by a member of the LavaRnd project. It seems that a Philips webcam, with the lens cap in place, is a good source of entropy for random number generators. In fact, the low-resolution stream is even better than the full-resolution version for this application. The LavaRnd folks would like to see the GPL driver back - and they have even volunteered to maintain it.

The other use for the GPL driver would be as a starting point while the compression protocol is reverse engineered and a completely free driver is created. There has been some speculation that this reverse engineering would be relatively easy - but it will remain speculation until somebody produces some code.

In any case, the PWC driver is likely to come back in some form; USB maintainer Greg Kroah-Hartman has stated that a conversation is in progress with Nemosoft (the original author) and that a patch is forthcoming. Getting a driver which only supports the low-resolution mode is unlikely to please many PWC owners, but it is a start. If the end result of all this is, eventually, a 100% free driver supporting full functionality, everybody will be better off.

(Log in to post comments)

Alan Cox agreed to maintain it

Posted Sep 2, 2004 4:02 UTC (Thu) by subhasroy (guest, #325) [Link]

I saw a kerneltrap article which mentions a post from AC that he agreed to maintain the driver when Linus challenged him.

The return of the Philips driver?

Posted Sep 2, 2004 7:46 UTC (Thu) by rankincj (subscriber, #4865) [Link]

Good-oh! I use the pwc driver quite happily without the pwcx module. GnomeMeeting's "small" and "large" video formats don't need compression to work, and this particular machine (120 MHz Pentium) can't spare the CPU power anyway. Without pwc, the webcam's nothing more than a USB microphone.

The return of the Philips driver?

Posted Sep 2, 2004 18:26 UTC (Thu) by bronson (subscriber, #4806) [Link]

Why can't this driver just be ported to libusb? All proprietary stuff would then be in userspace. Problem solved. Since this seems like such an obvious solution, there's got to be something that I'm missing...?

The return of the Philips driver?

Posted Sep 3, 2004 17:50 UTC (Fri) by horen (subscriber, #2514) [Link]

It is important to view the pwc/pwcx driver/decompressor in the context of all webcams which rely on them. In addition to the Philips family of webcams (11 models), the following make/models also use Philips chips, necessitating use of the pwc/pwcx driver/decompressor:

Askey:         VC010
Creative Labs: Webcam 5, Webcam Pro Ex
Logitech:      QuickCam 3000 Pro, QuickCam 4000 Pro, QuickCam Notebook 
               Pro, QuickCam Zoom, QuickCam Orbit/Sphere
Samsung:       MPC-C10, MPC-C30
Sotec:         Afina Eye
AME:           Afina Eye
Visionite:     VCS UM100, VCS UC300

This, as we say, is not "chopped liver"; neither are the multitude of Linux users who presently use these products (or Windows users who contemplate upgrading to Linux). Moreover, the manufacturers of these products are not going to stop using Philips chips.

At the same time that the ideological-purity of the GPL must be maintained, so, too, must pwc/pwcx-dependent hardware users not be disenfranchised.

My $0.02, YMMV.

The return of the Philips driver?

Posted Sep 5, 2004 1:51 UTC (Sun) by piman (subscriber, #8957) [Link]

Those people were disenfranchised when they bought proprietary hardware; they were further disenfranchised when the original author demanded the driver's removal from the kernel. This is not a fight between ideological purity and users' rights -- it's a fight between software hoarders and users' rights, as it almost always is.

The return of the Philips driver?

Posted Sep 4, 2004 22:47 UTC (Sat) by jmshh (guest, #8257) [Link]

As the original auther mentioned that the NDA has expired and he just didn't want to use that fact out of courtesy for Philips: Has anybody tried to ask Philips, what their actual idea about all this is? Maybe they would give the info without an NDA now, maybe not.