Marvell

[Bernardo Innocenti]

(cc dwmw2)

David Woodhouse and myself will attend at LCA.  It would
be nice if we could meet and discuss the technical details.

Moreover, I'm being kicked from the US due to visa issues,
and I'll be staying in Australia until Feb 15.  OLPC is
trying to get me a new visa, but it is unclear when I'll
be able to return to Boston.

I was planning to go sightseeing around, but maybe we
could use this opportunity to spin this project up.

Note that I've never been an OLPC employee, so I'm
not bond to any NDA they may have signed with Marvell.


Alex Gibson wrote:

> Trying to reply to both your emails in one.
> 
> Quoting Edward Cherlin <echerlin at gmail.com>:
> 
>> OK, my hand is up.
>>
>> I started a Wiki page, Marvell microkernel, and pulled together
>> various references to what we know and don't know, including a link to
>> the 88W8388 page, which says that it is the only public documentation
>> in the world on that chip.
>>
>> First thing, we need to know who has OLPC's copy of the Marvell
>> developer's kit, and what's in it.
>>
>> Do we have to sign an NDA to use the kit?
> 
> I don't know about the kit.
> 
> To get access to the chip docs(datasheet) and even then an NDA is no guarantee
> of receiving the docs as in our(UTS's) case.
> 
>> We apparently need more detail on the particular version of the
>> ARM946E-S core on the chip. There are a number of hardware options.
> 
> We need the datasheet , peripherals and memory/register map.
> 
> Marvell does not give out this information.
> They do not provide source code or FULL
> register information to anyone outside of marvell.
> And for those of us who have signed an NDA with Marvell, reverse engineering is
> forbidden.
> 
>> The code evidently mixes 32-bit ARM instructions and 16-bit thumb
>> instructions with data. We need a fairly smart disassembler. And a
>> legal opinion about reverse engineering.
> 
> Need internetworking support (mixing of arm and thumb code).
> 
> We hit a brickwall in terms of the docs.
> 
> We signed an NDA with Marvell but the last time I checked
> had not received access to any docs for chip/s used by olpc
> just for some other product families.
> 
> Marvell was trying to encourage us to participate in open802.11s
> started by Javier Cardona
> 
> http://open80211s.org/
> 
> 
> For toolchains there are a few options for opensource arm gcc builds.
> Can expect pretty much none of them will directly support marvells 
> extended ARM946E-S (by extended I mean the extra registers).
> 
> Note you want arm-elf-gcc not arm-elf-linux-gcc .
> A few I know of include 
> 
> code sourcery lite http://www.codesourcery.com/gnu_toolchains/arm
> has support for arm946es. Usually have the most up to date 
> in terms of chip support and optimisation as they work closely with arm.
> Note https://support.codesourcery.com/GNUToolchain/kbentry51
> 
> gnu arm   www.gnuarm.com - has support for arm946es
> What I usually use for embedded work as it provides a known working package for
> linux , mac osx and windows.
> 
> build yourself from source , must make sure you end up with support for 
> arm946es (if that is what Marvell have used).
> No easy way to verify that yur self built tool chain is working properly on
> arm946es
> See http://www.gnuarm.com/support.html on build process
> 
> commercial tools  
> http://www.rowley.co.uk/arm/index.htm  cross works for arm
> arm gcc + libraries + ide
> 
> full version of code sourcery g++ http://www.codesourcery.com
> 
> You have to be careful as not all arm gcc varieties floating around on the web
> have support or have been tested on a variety of arm9 cores.
> 
> 
> Next question is which micro kernal/rtos to use ?
> 
> I don't see the need to rewrite one from scratch, there are a few opensource
> rtos available that have been previously ported to arm9, not arm946es.
> 
> ecos http://ecos.sourceware.org/
> free rtos  www.freertos.org
> tinyos - xscale not arm9
> http://nuttx.sourceforge.net/  
> ???
> 
> One other question is how much of the spectrum management is in software ?
> 
> re:Microkernel documentation from owner
> Some of the threadx docs can be downloaded from www.rtos.com once
> you have registered.
> 
> The big question is can the OLPC use and or distribute firmware that has been
> produced by reverse engineering, considering their agreement/s/NDA/s ??
> with Marvell, which from memory(don't have our NDA at hand) prohibit such
> activities.
> 
> One option that may be workable is Marvell provide binaries or objects for the
> parts of the chip they will not give out the docs for.
> This was raised previously I think by Dan Williams ?
> 
> In my opinion there is no point in reverse engineering or developing firmware if
> olpc can't legally distribute and use it.
> 
> Alex Gibson
> 
> 
> 
> --
> UTS CRICOS Provider Code:  00099F
> DISCLAIMER: This email message and any accompanying attachments may contain
> confidential information.  If you are not the intended recipient, do not
> read, use, disseminate, distribute or copy this message or attachments.  If
> you have received this message in error, please notify the sender
> immediately and delete this message. Any views expressed in this message
> are those of the individual sender, except where the sender expressly, and
> with authority, states them to be the views the University of Technology,
> Sydney. Before opening any attachments, please check them for viruses and
> defects.


-- 
 \___/
 |___|   Bernardo Innocenti - http://www.codewiz.org/
  \___\  One Laptop Per Child - http://www.laptop.org/