Re: [patch 4/4] Markers Support for Proprierary Modules

[Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>]

* Ingo Molnar (mingo@elte.hu) wrote:
> 
> * Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> wrote:
> 
> > There seems to be good arguments for markers to support proprierary 
> > modules. So I am throwing this one-liner in and let's see how people 
> > react. [...]
> 
> ugh, this is unbelievably stupid move technically - so a very strong 
> NACK. Allowing marker use in unfixable modules (today it's placing 
> markers into unfixable modules, tomorrow it's marker use by such 
> modules) has only one clear and predictable effect: it turns marker 
> calls into essential ABIs because when faced with any breakage in an 
> unfixable module that makes use of a marker in some kernel subsystem 
> then all the pressure is on those who _can_ fix their code - meaning the 
> kernel subsystem maintainers that use markers.
> 
> unfixable modules should only be allowed access to easy things they can 
> access anyway, or to such fundamental things which we wont realistically 
> change anyway. Markers are neither.
> 
> (i also find it puzzling why you go out on a limb helping a piece of 
> _irrelevant_ technology that has been the unparalleled source of pain 
> and anguish to both kernel users and kernel developers.)
> 
> 	Ingo

Please note that this patch has a single purpose : to let proprietary
modules define markers to *export* information. The opposite (connect
callbacks to markers) is not allowed since the rest of the markers API
is EXPORT_SYMBOL_GPL'd.

I would also be strongly against letting proprietary modules access the
information provided by the markers. However, I think it's only useful
for the end user to let proprietary modules open up a bit, considering
that proprierary module writers can use the markers as they want
in-house, but would have to leave them disabled on shipped kernels.

As far as I am concerned, I want to help the end user, not the
technology itself.

Unless I have a proof that markers in proprietary modules (information
*providers* only) would be a pain to maintain, I won't object against
supporting proprietary modules.

Mathieu

-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68