Hi,
charliem_1216 <> [20080304 15:51:14 -0000]:
>
> --- In Alexander Clouter <> wrote:
>
> > So I git yanked myself a copy of the orion Linux tree and yesterday after
> > finding that it had stalled at 2.6.25-rc1 I noticed that it is in the
> > mainline tree...WIN!
>
> Great! I was hoping someone would step in and get this platform code
> in upstream kernel (eventually). This orion chip has a lot of cool
> features to play with ...
>
yeah, digging around the Marvell 'opensource' handbook I was shocked to see
a crypto and XOR device on board too :)
> > For the past week[3] or so I have been slapping around support for the
> > TS-7800 which I found surprisingly easy. Serial, SATA and the NIC seem
> > to be working just fine. My patch and .config file are available here:
> >
>
I noticed the EHCI USB support is there, I'll insert the module tonight and
see if it detects everything. If that works then the NAND becomes less
important as I can continue booting off an external harddisk.
> > Anyone know 'where' it is? Whilst I'm digging for bits, any idea if the
> > RTC is on the I2C bus and where on it? Just, if someone knows off the
> > top of their head[4] that would be great.
>
> AFAIK the battery backed RTC clock module is the same one as used on
> the ep9302 boards: M48T86PC1 (equivalent to DS-12887). So it's not on
> I2C, but on the TS-7250 the rtc-m48t86.c works OK. The source for the
> rtc7800 utility might also help you out; it's is at:
> ftp://ftp.embeddedarm.com/ts-arm-sbc/ts-7800-linux/samples/
>
I'm sure I had trawled the whole FTP site, guessing I missed that, probably
as it was called 'samples'. :-/ That will learn me.
> SD on TS boards will never go upstream without TS re-evaluating the SD
> NDA they signed.
>
>From my understanding they coded up an FPGA to do the dirty work, did I spy
two FPGA's on the board whilst I was digging around for chip ID's? I'm
probably naive on this (I have read the previous postings on this topic
though) but I cannot see how this is much different to nvidia's binary blob
drivers? If it is FPGA code being uploaded then I'm finding it difficult to
understand why the wrapper and device access code could not be GPL'ed? I'm
guessing this is all covered in the details of the NDA.
Of course the solution to all this is UIO[1] (the stable userspace IO API)
added a little while back. Would get all us off TS's back and keep us all
generally happy too. Worth TS moving to a slightly newer kernel for this
reason alone I'm guessing? Would also mean I get my Mobile IPv6 support I'm
doing all this for too :)
> BTW, what do you plan for your second ethernet?
>
As a network monkey I use VLAN's[2], to the point of excess usually. VLAN
aware switches are easy to come across on fleabay and even a number of home
xDSL routers seem to support them too.
What I'm still pondering about is what to do with ten serial ports :)
Cheers
Alex
[1] http://lwn.net/Articles/232575/
[2] http://en.wikipedia.org/wiki/Virtual_LAN
--
___________________________________
/ Better hope you get what you want \
\ before you stop wanting it. /
-----------------------------------
\ ^__^
\ (oo)\_______
(__)\ )\/\
||----w |
|| ||
signature.asc
Description: Digital signature
|