| 1 |
Am Dienstag 27 September 2011, 04:05:31 schrieb Grant Edwards: |
| 2 |
> On 2011-09-27, Michael Orlitzky <michael@××××××××.com> wrote: |
| 3 |
> > On 09/26/11 16:13, Grant Edwards wrote: |
| 4 |
> >> That's hilarious. |
| 5 |
> >> |
| 6 |
> >> The Linux developers are _constantly_ changing APIs in ways that break |
| 7 |
> >> existing device driver code. There are repeatedly wholesale |
| 8 |
> >> re-designs of some APIs that happen between minor versions of a |
| 9 |
> >> supposedly "stable" kernel. |
| 10 |
> >> |
| 11 |
> >> We have to touch our NetBSD and FreeBSD drivers maybe once every 3-4 |
| 12 |
> >> years. Often our Linux drivers have to be updated every 3-4 _months_ |
| 13 |
> >> to keep up with changes in the kernel that break things. |
| 14 |
> >> |
| 15 |
> >> I suppose one could try to claim that people who ship Linux drivers |
| 16 |
> >> for their hardware aren't "users" of the kernel, and therefore our |
| 17 |
> >> dealing with such breakage isn't a "user experience". |
| 18 |
> > |
| 19 |
> > Contribute your drivers upstream. When the devs change an API, they'll |
| 20 |
> > update your code for you. |
| 21 |
> |
| 22 |
> That sounds good, but in practice it doesn't work. |
| 23 |
> |
| 24 |
> 1) The kernel developers don't support any existing customers. Bugs |
| 25 |
> are only fixed for customers who are willing to run the next |
| 26 |
> kernel verison. I've got customers that are still running 2.4 |
| 27 |
> kernels. 2.6.18 is still widely used. Will the kernel developers |
| 28 |
> add new features, support for new hardware, or fix bugs for those |
| 29 |
> customers. Not a chance. |
| 30 |
|
| 31 |
so what? There are long term stable kernels with no api changes. Hmm... |
| 32 |
|
| 33 |
> |
| 34 |
> 2) The kernel developers only make sure that drivers compile. They |
| 35 |
> don't have the hardware or knowlege required to actually test |
| 36 |
> them. One of our drivers _is_ in the kernel. Sure, it builds, |
| 37 |
> but AFAIK, it hasn't actually worked for at least 10 years. |
| 38 |
|
| 39 |
and nobody complains on lkml about it - seems that nobody uses your hardware. |
| 40 |
|
| 41 |
If something stops working (called a 'regression' btw) it has to be fixed. |
| 42 |
Linus is very clear about that. |
| 43 |
|
| 44 |
> |
| 45 |
> Trying to maintain two drivers (one in-kernel and one out-of-kernel) |
| 46 |
> just creates twice as much work for no gain. |
| 47 |
|
| 48 |
then don't be outside the kernel. |
| 49 |
|
| 50 |
-- |
| 51 |
#163933 |
Archives