RE: compat-drivers for older kernels

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Thanks for the clarification.  I'll certainly return patches if/when I am successful.

One more thing I'm betting you can answer relates to the LINUX_VERSION macros: How do you find in what kernel version a feature changed (e.g. tty->driver->write became tty->ops->write)?  I can't imagine it's by inspection,  yet a cross-reference site like lxr.linux.no, which should be able to spit that out, does not.

-David


-----Original Message-----
From: mcgrof@xxxxxxxxx [mailto:mcgrof@xxxxxxxxx] On Behalf Of Luis R. Rodriguez
Sent: Thursday, March 28, 2013 11:39 AM
To: Jonathan David (AA-DG/PAO-ENG1)
Cc: backports@xxxxxxxxxxxxxxx
Subject: Re: compat-drivers for older kernels

On Thu, Mar 28, 2013 at 10:36 AM, Jonathan David (AA-DG/PAO-ENG1) <David.Jonathan@xxxxxxxx> wrote:
> Hi Luis,
>
> In my search for answers I came up with this link, and therefore I’m 
> talking to you:
> https://github.com/mcgrof/compat-drivers/tree/master/enable-older-kernels.
> Feel free to point me elsewhere…
>
>
>
> I clicked through to do-not-panic.com, read about compat-drivers 
> testing, saw the list of 31 kernels, including 2.6.24 with OK, and I’m 
> wondering how this can be.

OK means the master branch from compat and compat-drivers has been compile tested on that vanilla upstream kernel. So compat.git would work compiling against 2.6.24. compat-drivers.git will also compile successfully on 2.6.24.

> I’m trying to build compat-drivers-3.8.1 (for ARM, but I don’t think 
> that is the issue) against a 2.6.24.4 kernel.

Great.

> I started by selecting just the
> Bluetooth drivers, and see structure incompatibilities.  One example:
> drivers/bluetooth/hci_ldisc.c assumes that struct tty_struct has an 
> ops field.  In 2.6.24.4 it does not.

Bluetooth is not enabled by default on the master branch for 2.6.24.
If you want to work on that, patches are welcomed.

> I’ve been #ifdef’ing my through a number of things like the above, but 
> thought I’d ask in case I’ve overlooked something obvious.

Patches welcomed for enabling drivers / subsystems on older kernels.
By default they are enabled only for certain kernels where the maintainer decides to start supporting something. This is currently expressed on config.mk. This will hopefully be a lot clearer now that we have kconfig support staring.

  Luis
��.n��������+%������w��{.n����i���l�)��jg��������ݢj����G�������j:+v���w�m������w�������h�����٥





[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux