TLB programming in powerpc tree

[vb]

On Fri, Aug 22, 2008 at 6:24 AM, Josh Boyer <jwboyer at linux.vnet.ibm.com> wrote:
> On Thu, Aug 21, 2008 at 10:14:15PM -0700, vb wrote:
>>On Thu, Aug 21, 2008 at 7:29 PM, vb <vb at vsbe.com> wrote:
>>>
>>> But the main problem is that the kernel never sets up TLBs for neither
>>> the peripheral device, nor the onboard flash. I don't seem to be able
>>> to find the place where this is supposed to happen. I assumed that
>>> ioremap_nocache would take care of that, but this is not the case.
>>>
>>
>>well, in fact the TLB is set up as soon as an attempt to access the
>>peripheral is made. The problem apparently is the fact that the TLB
>>entry uses a wrong value in the nibble specifying the internal 460GT
>>block...
>
> I'm not entirely sure what you mean here.  Can you elaborate a bit more?
>

I was referring to the top 4 MSB of the 36 bit address.

So, I finally was able to access that device after massaging the
device tree definitions for it, but the next device (in a different
branch of the tree) still can not be accessed. What I see is that the
tlb record is created when the access is attempted, but the top
address nibble is 0 instead of 4, even though the device tree now
includes 36 bit addresses with the correct top bits.

I am trying to find code which configures the MMU and to understand
how this whole subsystem works and where the wrong top 4 bits come
from. Is there a document describing 44x powerpc memory management?

/vb

> josh
>