[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [microblaze-uclinux] MicroBlaze/MMU - questions of confidence
To the list - this email may irritate a few people with its rather
overt commercial angle, and I would normally refrain from such a
strong plug for PetaLogix here as this is a technical list. However I
think it's valid in this instance given Dan's question. Please send
flames to me directly, not the list.
Dan,
You ask some good questions.
We have great confidence in our MMU offering (and noMMU, and PetaLinux
generally), from a number of perspectives:
1. Between PetaLogix and some of our key clients, tens of thousands of
dollars were invested last year in cleaning up the MMU support that
was prototyped by LynuxWorks (which was actually based largely on
linux-2.4.32 patches developed by Xilinx internally to validate the
MMU hardware).
2. We have commercial clients using this tree, and also deliver
training workshops on the MMU kernel with great results.
3. We are partnered now with Xilinx to merge the MicroBlaze
architecture (MMU and noMMU) upstream to kernel.org. This is in
progress on the kernel mailing list as we speak. Our target is the
2.6.30/31 merge window, but any bugs we do find, the fixes are
backported immediately to the 2.6.20 tree in the PetaLogix SVN repo.
(take a look at http://developer.petalogix.com/git/gitweb.cgi if you
want to see the work in progress)
4. As part of (3) we are running regular LTP (Linux Test Project)
regression tests against the new kernel. This is a great stress test
and validation. Again, any fixes identified feed directly back to the
2.6.20 SVN tree.
I am aware of the perspective that says "if you pay for it, it must be
better" - to that I have a few comments:
1. PetaLogix is a 100% Xilinx-focussed company, and MicroBlaze/Linux
is probably at least 90% of our business. The quality of the
MicroBlaze kernel is our lifeblood. You may wish to consider whether
the same can be said of other alternatives, commercial or otherwise.
2. You can pay us for it if you wish - PetaLogix's core business today
is service and support delivery. We would be delighted to support
your work with PetaLinux. Our model there is simple - let us do the
bits that we can do best (EDK subsystem design, board bringup, custom
IP development, kernel integration, ...), and let you focus on where
your key skills lie. Why spend weeks or months down in the hardware
and kernel when you could be prototyping your application code on the
desktop while we build the platform for you [1]?
Of course I hope you choose PetaLinux, regardless of whether you
engage with us commercially. But, if you do go with LynuxWorks,
please drop back in and tell us about your experiences. We're not
afraid of a little competition, it is a great motivator to improve!
OK, that's as close to spamming this list as I want to get, so please
feel free to follow up in a private email if you have any further
questions.
Best regards,
John
[1] The productivity gains of embedded Linux are no myth. Last year
we subcontracted a local developer to write a fairly complex
application which interacted closely with a custom kernel driver and
some VHDL IP that we were developing for a client. We spec'd the
driver interface and wrote a stubbed software layer, and the
subcontractors developed the whole application on a desktop PC while
we built the real hardware and did the kernel architecture. We had an
integration meeting, and had the application code running talking to
the hardware on the client's board in about 15 minutes.
On Fri, Mar 27, 2009 at 2:13 AM, Dan Miller <dan@xxxxxxxxxx> wrote:
>
> Hi all;
> We're beginning a new project here, that will use MicroBlaze with MMU, on a Virtex5 FPGA. We will be running Linux of some sort on it, because we wish to use existing applications from a stand-alone Linux system. My understanding is, we essentially have two Linux choices for this project; Petalinux (open source, at least for kernel) and Linuxworx (commercial). I would rather work with the open-source package, but others in my department have a tendency to expect that commercial software will be more stable and professional, so I want to ask some questions about the Petalinux implementation;
>
> In particular, I understand that the MMU implementation is very new, in fact the upcoming v0.40 distribution will be the first snapshot release to have all the MMU support in it. So how much usage has it actually gotten? If I tell my boss and the hardware engineer "just use Petalinux, the MMU support is solid and dependable, I have confidence in it.", how solid will my footing be?
>
> Mind ye, I don't know *anything* about the Linuxworx implementation either; once I saw Petalinux I just focused on that...
>
> Dan Miller
>
> ___________________________ microblaze-uclinux mailing list microblaze-uclinux@xxxxxxxxxxxxxx Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/
--
John Williams, PhD, B.Eng, B.IT
PetaLogix - Linux Solutions for a Reconfigurable World
w: www.petalogix.com p: +61-7-30090663 f: +61-7-30090663
___________________________
microblaze-uclinux mailing list
microblaze-uclinux@xxxxxxxxxxxxxx
Project Home Page : http://www.itee.uq.edu.au/~jwilliams/mblaze-uclinux
Mailing List Archive : http://www.itee.uq.edu.au/~listarch/microblaze-uclinux/