Advertisement
Interested in hardware, diags, validation, Linux, C, ARM, Microcode and low level programming and blazing networks?
| |||||||||||||
|
| |||||||||||||
A Linux Driver Project status reportA Linux Driver Project status reportPosted Apr 8, 2008 6:11 UTC (Tue) by djabsolut (guest, #12799)Parent article: A Linux Driver Project status report
Perhaps one of the reasons for the perceived lack of drivers for Linux is that to many people Linux == a Linux distribution. Getting a driver into the kernel does not automatically make the hardware useful. No matter how much I alter my Fedora 8 setup, I still can't get the audio to work on my laptop. The kernel driver loads fine, but something in the elaborate wrapper mechanism for the driver (PulseAudio + etc) is screwing things up. Additionally, Xorg requires it's own drivers, and there is no denying that there is a lack of drivers for Xorg, e.g. recent ATI and NVIDIA chipsets. (As a sidenote, it would be useful to move this stuff out of Xorg and into the kernel).
(Log in to post comments)
A Linux Driver Project status report Posted Apr 8, 2008 6:53 UTC (Tue) by gregkh (subscriber, #8) [Link] If you have audio problems with Fedora, file a bug and work with the developers to solve the issue for you. As for xorg, some drivers are already migrating into the kernel. There really isn't that many different vendors for video chips these days, so there is not a large variety of drivers needed anymore.
A Linux Driver Project status report Posted Apr 8, 2008 13:40 UTC (Tue) by djabsolut (guest, #12799) [Link] If you have audio problems with Fedora, file a bug and work with the developers to solve the issue for you.The audio issue was an example. The point is that simply having a driver does not automatically mean the hardware is actually useful for the user. 'amartoq' and 'osma' describe this problem better in comments below.
A Linux Driver Project status report Posted Apr 8, 2008 8:29 UTC (Tue) by sylware (subscriber, #35259) [Link] >Additionally, Xorg requires it's own drivers, and there is no denying that there is a lack of drivers for Xorg, e.g. recent ATI and NVIDIA chipsets. (As a sidenote, it would be useful to move this stuff out of Xorg and into the kernel). Working on it... working on it... (nouveau reverse engineering and full AMD/Intel specs). A brand new graphic stack won't happen in a day, it's rather a long process when you look at those chips complexity.
A Linux Driver Project status report Posted Apr 8, 2008 19:58 UTC (Tue) by drosser (guest, #29597) [Link] Now, I am not an Xorg developer and the only platform I use is Linux, but why would Xorg want to move their drivers into the Linux kernel? Wouldn't that break compatibility with the BSD's, Solaris, .etc?
A Linux Driver Project status report Posted Apr 9, 2008 2:47 UTC (Wed) by djabsolut (guest, #12799) [Link] Now, I am not an Xorg developer and the only platform I use is Linux, but why would Xorg want to move their drivers into the Linux kernel? Wouldn't that break compatibility with the BSD's, Solaris, .etc? From a purely technical point of view, programs which are in "user space" should not be touching the hardware. That's the kernel's job, i.e. the kernel abstracts the hardware. Currently Xorg is half a user space program and half a graphics kernel. Having two kernels can and does lead to problems. As for taking care of BSDs/Solaris -- yes that's an issue. However, it may be possible to write low-level drivers (licensed under a BSD license) that are then integrated with Linux/BSD/Solaris kernels using a kernel-specific wrapper. (A ruder alternative is to look at the ratio of Linux to BSD users and concentrate only on the biggest slice).
|
|
||||||||||||