ppc manual paging question

[Wang, Baojun]

On Monday 22 October 2007 14:01:33, Benjamin Herrenschmidt wrote：
> > I'm porting an adeos nano kernel named xtratum (http://www.xtratum.org)
> > from x86 to ppc, I think I'm near the ending except the above problem.
> > xtratum is doing things like xen but it's much simpler (it's aimed for
> > realtime), it need provides memory space sperations for it's domains, so
> > I need manually paging. Each domain is loaded by a userspace program
> > (instead of the root domain as a kernel module), the loader will load the
> > domain's (ELF staticly excutable) PT_LOAD section into memory, and then
> > raise a properly system call (passing the structurized loaded data as
> > arguments) to load the domain via load_domain_sys(), and at the last step
> > of loading the domain, xtratum will jump to the entry code of the new
> > domain(asm wrappered start() routine) and then everything should be fine.
> > The problem now is as follow:
> >
> > under my ppc (440GR/440EP) platform, start() is always at 0x100000a0, but
> > I guess there is something wrong with my mm code so after the domain is
> > loaded, the virt addres 0x100000a0 just point to garbage instead of the
> > right start() routine. So how can I setup paging properly so that the
> > virtual memory could be translated to proper data?
>
> Are you aware that the 440 MMU doesn't actually know what a page table
> is and doesn't load PTEs from memory ?
Sorry I didn't realized that. I've finished looking pem64b.pdf, but it's 
almost nothing like in ppc 4xx/44x.
> It's a software loaded TLB, you'll have to put translations in the TLB
> yourself. You'll need to design your own data structures for that, tho
> you can use a page table for tracking, like we do in linux, and then
> have your own TLB miss handler to fill the TLB from that.
OK, Shall I look for the DataTLBError code in head.S? I realized that I've got 
DataTLBError sometimes via BDI2000/kgdb (the fault address is within 
DataTLBError), Shall I should also look for arch/ppc/mm/fault.c?
> Your initial code probably need to bolt a TLB entry for the kernel
> itself.
I don't understand this clearly, how can I do this?
> Ben.

  Regards,
Wang


-- 
Wang, Baojun                                        Lanzhou University
Distributed & Embedded System Lab              http://dslab.lzu.edu.cn
School of Information Science and Engeneering        wangbj at lzu.edu.cn
Tianshui South Road 222. Lanzhou 730000                     .P.R.China
Tel:+86-931-8912025                                Fax:+86-931-8912022