Re: Which ARM CPU to use? (Level of support, etc.)

Top Page
Attachments:
Message as email
+ (text/plain)
Delete this message
Reply to this message
Author: Russell King - ARM Linux
Date:  
To: Xavier DEBREUIL
CC: Shane Nay, linux-arm-kernel
Subject: Re: Which ARM CPU to use? (Level of support, etc.)
On Tue, Jun 19, 2001 at 07:55:04PM +0200, Xavier DEBREUIL wrote:
> cf emails last month and the one I've written yesterday on that very
> subject ; people you'll meet from Cirrus will always tell you that the
> support from them is the best and blahblahblah ; and what you will see
> when crucial problems arise is that it is always the same : you are
> alone to solve it...


You are not the only person who is alone in this. I wish you could
solve your problems, or someone else could.

One of the great attractions of Linux is this "open source" nature of
it, where you're not alone in your problems, and the community helps
to solve them. Well, yes, you probably aren't alone. But if the
people who provide Linux to you aren't providing support, and aren't
_working with_ the community, then everyone who is getting involved
with that port isn't going to be successful.

For the purposes of the following discussion, I'll call the body
producing the code "body A".

In this particular case, I have seen very little evidence of body A
merging their code back into the community. What's more is that they
appear to be providing virtually zero support for their code to people
trying to use it.

This tells me that they're not particularly interested in their product
lasting a long time, or anyone developing anything from it. When people
then come around here expecting us to support it. Obviously its not our
job to support body A's port, especially when they aren't interested in
merging it back into the main source, and letting the community benefit
from their work, and body A _and their users_ benefit from our work.

Since that is the case, when a user of body A's code complains that
something doesn't work, we will of course not be able to provide any
help with it. We don't have visibility of everything in body A's code.
As far as I personally am concerned, if body A is not willing to make
the effort to get their code merged, then body A's code is effectively
closed source.

The Linux community response to supporting closed source stuff is
"run it without the closed source code or no support". Why? We don't
know what effects that code is having on the rest of the kernel because
it hasn't been reviewed.

I hope that people who are producing their own trees with little or no
intention of merging it read this to get a grasp on the best way of
not making money, which seems to be in common practice today.

To be frank, anyone who is selling a product and providing Linux on
it without any intention of merging or supporting it deserves to
loose.

PS, If someone wants to get support, I am more than willing to provide
support on one condition: that they guarantee that they will provide
patches for their kernel to be merged into my tree, and provide
compensation for preventing me from doing what I'd like to be doing.
_____
|_____| ------------------------------------------------- ---+---+-
  |   |        Russell King             --- ---
  | | | |            http://www.arm.linux.org.uk/            /  /  |
  | +-+-+                                                     --- -+-

  /   |               THE developer of ARM Linux              |+| /|\
 /  | | |                                                     ---  |
    +-+-+ -------------------------------------------------  /\\\  |