charliem_1216 wrote:
> Hi --
>
> [...]
> You can shorten this to:
> mem= mem= mem= mem= Not much shorter, I
> know, but less likely to make a typo. There is another thread
> discussing this issue of the upper 32M (see "only 32M detected by
> redboot").
Thanks for the pointer.
>>>> There is ep93xx_defconfig, which you could use as base. You would
>
> Also, there is an updated ep93xx_defconfig in 2.6.20-rc2. You may
> want to look over those diffs.
Ok. 2.6.20-rc3 is already out there, so maybe I'll try it, too.
>>> also need
>>>> some of ep93xx_* patches available from Lennert's svn repository at
>>>> http://svn.wantstofly.org/kernel/
>
> Some patches may be needed, but Lennert has done a great job sending
> his ep93xx patches upstream. There is definitely less-stable
> development stuff in his repository though (look at the series file).
> I'll have to try a vanilla kernel build, and see how it goes.
Will do.
>>>[...]
>>> I tried booting it nevertheless, but the kernel just hung after
>>> "Decompressing linux......... done" (or something like that).
>
> Printascii should help here.
It did, indeed :)
>>> [...]
>>> This really depends on how different the 2 boards are. The first
>>> option
>
> Yes, identifying the differences is key. I haven't worked with the
> 7400 though, so I'm not much help here. But TS seems quite ready to
> answer questions (and I would think especially so when helping out
> community support for 2.6 kernels).
From I can see, the kernel almost booted (except for not finding out a
NFS root :P) and it found the flash partitions, serial ports, etc., so
it seems that the boards are pretty identical.
I got the impression that most of the peripherals that differ between
the 2 boards are directly controlled from userspace anyway, and don't
need kernel support.
> IMHO, it would be best to work with the MACH_TYPE_TS72XX definition if
> possible. The idea is that the bootloader passes the machineid, and
> then the kernel 'does the right thing'based on that. If you define a
> new machine type, that would mean a different redboot image for each
> machine type. (But hey, I guess that's the case already, so maybe
> it's no big deal).
Yes, if the boards are identical from a kernel POV, then it makes
perfect sense to use the same machine ID.
> (Redboot currently _doesn't_ pass the right machine code, hence the
> need for the head.S hack in Lennerts' "local.diff", but this may be
> changing).
The new bootloader that is used on the ts7400 needs to grow a module
parameter or an ioctl that allows userspace to provide the machine ID to
boot the kernel.
>>> [...] I really wanted to do this The Right Way(tm) eventually leading to
>>> the
>>> board being supported by official vanilla kernels.
>
> I suggest you get direction from Lennert Buytenhek (maintainer of the
> ep93xx stuff in the kernel). AFAIK, he would be the one to feed the
> patches upstream, so it would be very good if he agreed with your
> approach :) You might post on the linux-cirrus ML for feedback.[2]
For now, I'm trying to find out what are the exact differences between
the 2 boards. I might just realize that a kernel for the ts7250 Just
Works on the ts7400.
Thanks for the help,
--
Paulo Marques - www.grupopie.com
"The face of a child can say it all, especially the
mouth part of the face."
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/ts-7000/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/ts-7000/join
(Yahoo! ID required)
<*> To change settings via email:
<*> To unsubscribe from this group, send an email to:
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|