Manuals (Re: [Company] Weekly Operations Update 36/2009)

Wolfgang Spraul wolfgang at qi-hardware.com
Wed Sep 9 22:27:25 EDT 2009


Hi,

> > Ingenic gave us explicit permission to redistribute a number of their
> > manuals without NDA. They have also spent a lot of time updating their
> > documentation in recent months.
> That's great news.

Indeed. Thanks! You have no idea how many meetings were neccessary for that.

> I wait with bated breath...

I apologize for asking, but I need some new ways to prey them open. What
do you plan to do with the SIMD Manual? Do you want to do some hacking?
open source? Where is your project/source codes?
Would it be acceptable for you to sign a FOSS-friendly NDA with Ingenic and
get the SIMD documentation that way?
You know that a lot of the SIMD stuff is already open, I summarized it in
a recent mail on the list. With all that is open, I believe it's possible
to start SIMD hacking already. We have the names of all instructions, for
47 of them we have C equivalents, we have mplayer and jpeg .c source codes
using the instructions, and we have the assembler in source form.
I think the first step would be to integreate a SIMD-accelerated mplayer
into openwrt, using the openwrt-generated toolchain (needs to be enhanced
to support the SIMD instructions), etc.

> Does this mean there is no specific Jz4720 Programming Manual? In
> other words, is the Jz4740 Programming Manual the official reference
> for the jz4720?

I didn't ask but yes I would think so. Same for the 4725.
The 4720, 4725 and 4740 have the same die, just different packaging (and
thus different pin-out and potentially different end-user features).
But then the 4725b is the QFP variant of the 4750 (I think), but they have
2 programming manuals :-) Maybe I don't have full understanding yet...

> Is this related to the mysterious jz4757 chip ([1])? I don't think
> it's a simple updated jz4750 or it wouldn't have received its own tree
> in the Ingenic Linux kernel.

You never know. It's messy. I asked the CEO of Ingenic about the 4757 on
your behalf, and he was laughing. They did that chip for one big customer
who wanted to have their own chip. Even though it's the same die as the
4750 I believe. Chinese companies try all sorts of things to reduce the
chances of their products being copied. Even if they can throw off the
copy-cats for a few more weeks, that's already worth real money here.
So some manufacturers will meticulously scratch out (manually!) the labeling
of chips, or they will request the semiconductor to come out with a new
model name just for them.
Liu Qiang (the CEO) said they will not do this anymore in the future. Of
course that's just his wish, if a large enough customer comes they can
always ask for whatever they want.
I would not read much, if anything, into the Linux tree thing. That's also
very messy. In fact it may be that our (Qi Hardware's) Linux kernel will
become the official Ingenic kernel. We are working to setting up a tighter
collaboration. We just have to be careful that it doesn't mean that in the
end Ingenic puts less resources on Linux, what we want is that we
collaborate more efficiently.

I would say let's focus on the open devices in our hands, from my
perspective primarily the NanoNote of course, rather than to second
guess every chip variant Ingenic has made here or there.
What we need is a unified and recent Linux kernel and rootfs that runs
on as many Ingenic-based devices as possible.

Let's just start! Join the NanoNote hacking at our openwrt tree, or
dingux.com, or openinkpot.org! :-)
Wolfgang

On Wed, Sep 09, 2009 at 10:02:52PM +0200, W.G. van de Hulst wrote:
> 2009/9/9, Wolfgang Spraul <wolfgang at qi-hardware.com>:
> > Ingenic gave us explicit permission to redistribute a number of their
> > manuals without NDA. They have also spent a lot of time updating their
> > documentation in recent months.
> That's great news.
> 
> > Still, as of right now, we got the following updated manuals (SIMD is
> > still missing, working on it...):
> I wait with bated breath...
> 
> > Jz4720 Data Sheet, Rev 1, Jun 2008, 37 pages
> > Jz4740 Data Sheet, Jun 2007, 35 pages
> > Jz4740 Programming Manual, July 22, 2009, 539 pages
> Does this mean there is no specific Jz4720 Programming Manual? In
> other words, is the Jz4740 Programming Manual the official reference
> for the jz4720?
> 
> > Jz4750L Data Sheet, July 17, 2009, 37 pages
> > Jz4750L Programming Manual, July 14, 2009, 570 pages
> Is this related to the mysterious jz4757 chip ([1])? I don't think
> it's a simple updated jz4750 or it wouldn't have received its own tree
> in the Ingenic Linux kernel.
> 
> [1] http://forum.eepw.com.cn/thread/155995/1
> 
> _______________________________________________
> Qi Developer Mailing List
> Mail to list (members only): developer at lists.qi-hardware.com
> Subscribe or Unsubscribe: http://lists.qi-hardware.com/cgi-bin/mailman/listinfo/developer




More information about the discussion mailing list


interactive