2.3.51 tulip broken
Fri Mar 17 03:55:37 2000
On Thu, 16 Mar 2000, Jim Morris wrote:
> Donald Becker wrote:
> > But this letter isn't about this detailed, narrow issue. It is instead
> > about the attitude that such interface changes are readily, and only lightly
> > considered. Each one of those minor interface change impacted multiple
> > drivers. Many of those drivers were being actively worked on, perhaps based
> > on a stable 2.2 kernel.
> Bingo. I think you've hit the nail on the head. It comes down to the
> difference between just junping in and doing a "quick hack" that seems
> to suit an immediate need, or actually "designing" the change. By
> design, I mean considering how the change might impact other peices of
> code, etc.
The PCI interface in 2.3 is quite well designed and covers a superset of
the features in Donald's pci-netif interface. Further, it receives/has
received much wider testing.
> > To you such an interface change is a minor tweak that can be with an
> > automated search-and-hack. What takes you only a minute per file to change
> > might take the developer hours over the next few months to deal with to
> > integrate with testing. This is especially true when Linus insists that all
> > traces of 2.2 support be expunged in the 2.3 drivers.
> Ouch! That's a sore spot for me... I have no intention of moving any
> production system to 2.3.xx, or even the 2.4 kernel until its been out
> for some time. For that matter, I've still got production systems
> running the 2.0.36 kernel, with uptimes of hundreds of days. If it
> ain't broke, don't fix it!
> The architectural changes to the drivers that Linus is insisting on mean
> that I have to have completely different drivers for 2.0.xx, 2.2.xx or
> the later 2.3/2.4 kernels.
It does not mean that at all. Look at current acenic or rrunner
drivers for example. It is ENTIRELY possible to create a 2.3 driver
which is fully backwards compatible with 2.0 and 2.2.
Donald chooses to create a 2.0 driver which happens to work in 2.2/2.3.
> 2.2.xx source, I think. I have yet to see a good solid explanation for
> breaking driver compatibility between 2.2 and 2.3.
Much better kernel infrastructure.
To unsubscribe send a message body containing "unsubscribe"