F7 nvidia-96xx driver problem and [rant]

[Robin Laing]

Gene Heskett wrote:
> On Tuesday 24 July 2007, Lonni J Friedman wrote:
>> On 7/24/07, Frank Cox <theatre at sasktel.net> wrote:
>>> On Tue, 24 Jul 2007 15:47:38 -0700
>>>
>>> Lonni J Friedman <netllama at gmail.com> wrote:
>>>>> If you didn't wish to receive an answer, don't ask.
>>>> You didn't answer the question, you merely started spouting FUD.
>>> When you stop providing mis-information, I'll stop correcting it.
>> What exactly would you be correcting?  Please educate us all on your
>> background using NVIDIA products.  Or are you just speaking on a topic
>> of which you have no first hand knowledge?
> 
> Lonnie, I too use the NVIDIA driver.  That code is closed source and NO ONE 
> HERE has access to it should it be defective.
> 
> Now, if you can duplicate the problem using the nv driver, which we as a 
> group, do have the src code for, then someone might be able to help you.
> 
> But, since the NVIDIA driver is closed, these people have no way of knowing if 
> it is correct, or contains a miss initialized pointer that can scribble all 
> over the rest of the systems memory, causing all sorts of hate, discontent 
> and problems.  So as long as you and I use that driver, there is absolutely 
> zero chance that a kernel bug report filed by either of us will get more 
> attention than cleared with a notabug, kernel is tainted note in bugzilla.
> 
> Those are the revelant facts, and if they don't happen to fit your opinion, 
> that is unfortunate but it doesn't alter the facts in the real world, so why 
> don't you either duplicate the problem while using the nv driver, or toddle 
> off and install windows something or other on your box and leave these good 
> people to do what they do best, which is to write better code than Bills 
> best.
> 
> You might also take your problem to nvidia, where legit bugs are usually fixed 
> and new drivers issued in just a few days.  If its indeed their problem.
> 

I for one am having problems with the Nvidia driver but I don't think 
the issue is with the driver as I have issues with the nv driver as 
well.  If I change from the Nvidia driver to the nv driver, I get the 
same trace info from xorg that I get with nvidia when xorg decides it 
wants to use a large amount of processor power.

I ran the tests last night after upgrading to the latest kernel.  The 
latest kernel works better than the previous kernel.


<rant>
On the note of Fedora not wanting to look at closed source driver 
issues, I understand but that affects the ability to be adapted to the 
desktop and the desktop user.  Many users like myself have a hard time 
tracing problems and then do the best we can to submit bugs.  But when 
the finger pointing starts it gets harder.  The people that know how to 
trace and diagnose these types of problems are the same people that 
develop and produce the software.  They understand the code and the 
procedures to trace the problem.

In this case, changing a kernel changes the problem.  In one post on the 
nvidia site, it was mentioned cfq, in the kernel may be causing the 
problem.  In a previous post in this thread I mentioned CFS but I was 
wrong.  Now if the problem is the Linux Kernel and cfq, who is then 
responsible?  Is it Nvidia or the kernel team?  Who should be submitting 
the bugs?  Where does this leave the users that doesn't know how to do 
the tracing and submit a proper bug?

FWIW, I am reading the same type of issues with Xorg going to almost 
100% CPU and becoming almost unresponsive for all types of video cards.

As one article I read recently states.  Linux development is geared 
towards the server market and is ignoring issues with the desktop.  From 
the response to this issue, I am leaning towards believing this.

For many desktop users, 3D is a necessity.  For home users it is almost 
a requirement, especially with children who want to play games such as 
PPRacer or Super Tux Kart.

</rant>
-- 
Due to the move to Exchange Server,
    anything that is a priority, please phone.
Robin Laing