[prev in list] [next in list] [prev in thread] [next in thread] 

List:       redhat-devel
Subject:    kernel modules
From:       Panu Matilainen pmatilai () pp ! htv ! fi
Date:       2000-09-27 20:50:03
[Download RAW message or body]

On Thu, 28 Sep 2000, John Summerfield wrote:
> > 
> > When you change the kernel configuration even while staying on the same
> > version it's possible (or likely, I think) that internal kernel
> > datastructures change. Some of them are more critical than others. 
> > 
> > If you want more detailed explanation you'll have to search the net (read
> > linux-kernel archives) - this has been discussed about on linux-kernel
> > several times over the years.
> >
> 
> Come come. It's up to YOU to provide evidence to support your assertion.
> 
> It's because the documentation says that 'n' is a safe reply that I've been 
> doing it that way for years. If, as you say, it's been discussed on l-k and 
> found to be unsafe, then that documentation would surely have been changed 
> (and maybe the option removed).

.. already regretting opening my mouth about this.. I probably should've
said: *as far as memory serves* there has been quite a few discussions
around this topic on l-k. Mostly it's got to do with binary modules and
problems related to them as linux interfaces sometimes change even in the
stable kernel series. Module versioning is there to prevent these invalid
modules from being loaded and turning it off has the *potential* of being
unsafe.  

That said, I'm not going to spend my night hunting for information I
personally don't have much use for. Some mails around the issue can be
found at
http://www.uwsg.iu.edu/hypermail/linux/kernel/9712.2/index.html#start
It's old but AFAIK the basic problem is still relevant. (and no that's not
really the discussion I was referring to, just not interested staying
awake because of this)

Also I'm going to shut up now about this topic since I'm 100% sure there
are people on this list far more qualified to answer this than me :)

	- Panu -

> 
> To be sure, were I a vendor distributing a binary-only module to support my 
> Flash New Device, then I would seriously consider turning it ON so that if 
> there were a problem, it arose out of the customers' silliness, not mine.
> 
> 
> 
> 
> 
> 
> _______________________________________________
> Redhat-devel-list mailing list
> Redhat-devel-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/redhat-devel-list
> 

[prev in list] [next in list] [prev in thread] [next in thread]