Hi,
init_mm is no longer exported in 2.6.25, because there are no in tree
modules that use it.
But the closed sources nvidia drivers are using it.
Is it possible to reexport this symbol to let the driver work with this kernel?
Please CC me when repyling.
On Wed, Feb 27, 2008 at 6:29 PM, drago01 <[email protected]> wrote:
> Hi,
> init_mm is no longer exported in 2.6.25, because there are no in tree
> modules that use it.
> But the closed sources nvidia drivers are using it.
> Is it possible to reexport this symbol to let the driver work with this kernel?
>
Yes we'll get to it right away...
Oh wait...
Dave.
drago01 wrote:
> init_mm is no longer exported in 2.6.25, because there are no in tree
> modules that use it.
> But the closed sources nvidia drivers are using it.
> Is it possible to reexport this symbol to let the driver work with this kernel?
The fact that there are no in-tree modules that use init_mm is rather
compelling evidence that it's not a necessary part of the kernel module
API. Nvidia needs to fix their code. If this is a burden, perhaps they
should publish their code under a GPLv2-compatible license so we can
show them how to do it.
-- Chris
On Thu, Feb 28, 2008 at 9:23 AM, Chris Snook <[email protected]> wrote:
> drago01 wrote:
> > init_mm is no longer exported in 2.6.25, because there are no in tree
> > modules that use it.
> > But the closed sources nvidia drivers are using it.
> > Is it possible to reexport this symbol to let the driver work with this kernel?
>
> The fact that there are no in-tree modules that use init_mm is rather
> compelling evidence that it's not a necessary part of the kernel module
> API. Nvidia needs to fix their code. If this is a burden, perhaps they
> should publish their code under a GPLv2-compatible license so we can
> show them how to do it.
Publish it and get it merged so you'd have no choice but to keep the
symbol exported in the first place?
Ioan Ionita wrote:
> On Thu, Feb 28, 2008 at 9:23 AM, Chris Snook <[email protected]> wrote:
>> drago01 wrote:
>> > init_mm is no longer exported in 2.6.25, because there are no in tree
>> > modules that use it.
>> > But the closed sources nvidia drivers are using it.
>> > Is it possible to reexport this symbol to let the driver work with this kernel?
>>
>> The fact that there are no in-tree modules that use init_mm is rather
>> compelling evidence that it's not a necessary part of the kernel module
>> API. Nvidia needs to fix their code. If this is a burden, perhaps they
>> should publish their code under a GPLv2-compatible license so we can
>> show them how to do it.
>
> Publish it and get it merged so you'd have no choice but to keep the
> symbol exported in the first place?
High-performance graphics developers are pretty much the last people on
earth I want to see writing code that calls init_mm. These people
patent their bugs, rather than fixing them, when they turn out to make
things go a little faster and generate "mostly correct" results. I
think we have a lot to teach them about kernel driver development,
including how to get by without init_mm being exported.
-- Chris
> High-performance graphics developers are pretty much the last people on
> earth I want to see writing code that calls init_mm. These people
> patent their bugs, rather than fixing them, when they turn out to make
> things go a little faster and generate "mostly correct" results. I
> think we have a lot to teach them about kernel driver development,
> including how to get by without init_mm being exported.
Interesting how the less knowledgeable seem to be the ones who sign the NDAs
Ioan Ionita <opslynx <at> gmail.com> writes:
>
> > High-performance graphics developers are pretty much the last people on
> > earth I want to see writing code that calls init_mm. These people
> > patent their bugs, rather than fixing them, when they turn out to make
> > things go a little faster and generate "mostly correct" results. I
> > think we have a lot to teach them about kernel driver development,
> > including how to get by without init_mm being exported.
>
> Interesting how the less knowledgeable seem to be the ones who sign the NDAs
>
I ma sorry, but new linux users (mostly ubuntu) want their system to work out of
the box. Ruining, at least for the time being, the only decent nvidia drivers
isn't going to help. nd I don't think gamers, such as myself, will be very happy
if you proceed with this change. Enemy territory quake wars or other games won't
work anymore because their 600 $ Geforce 8 cards won't work correctly.
If I wouldn't be able to play game anymore, with ATI having bad performance on
linux, I would switch back to windows and I'm not the only one. It's as simple
as that.
Great Idea!!
On Mon, Mar 03, 2008 at 03:54:48PM +0000, Eragon wrote:
>
> I ma sorry, but new linux users (mostly ubuntu) want their system to work out of
> the box. Ruining, at least for the time being, the only decent nvidia drivers
> isn't going to help. nd I don't think gamers, such as myself, will be very happy
> if you proceed with this change. Enemy territory quake wars or other games won't
> work anymore because their 600 $ Geforce 8 cards won't work correctly.
> If I wouldn't be able to play game anymore, with ATI having bad performance on
> linux, I would switch back to windows and I'm not the only one. It's as simple
> as that.
>
> Great Idea!!
Independent of this specific issue and even ignoring the question
whether binary-only Linux kernel drivers are legal at all there are a
few points to note:
First of all, Linux is great for many people and in many cases. But
buying a computer with a $600 graphics card for gaming under Linux
doesn't make sense. Simply because having the latest desktop hardware
perfectly supported and with perfect performance is not yet realistic
under Linux (and there's more than just the graphics card in a computer).
And binary-only kernel drivers have the problems that noone except for
the vendor can debug them. Some binary-only graphics drivers are
well-known for having caused quite a few kernel kernel crashes under
Linux - and this harms the reputation of Linux as a stable operating
system.
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed