Re: BIOS 1.05: conditions before jump to kernel image on EBSA285

[Russell King - ARM Linux Admin]

Gaixia Zhang writes:
>      Caches flushed,
>      MMU off
>       mode to supervisor mode
>       D-cache off, I-cache off
>       Interrupt disabled (?) (I saw sti() in BIOS 1.05 right before
> boot_kernel()) (?)
>       kernel image loaded at 0x8000
>       parameter page properly initialized
>       r0=0, r1=4
> 
> What else is needed?

Nothing else.  It's probably the PCI stuff which is hanging it.  I suggest
checking the hardware configuration of the DC21285, since it sounds like
any access to the PCI bus just hangs the processor.

> BTW, if I do all the above initialization in head-armv.S, the original
> code will be misplaced (shifted), should I load the produced image below
> 0x8000 so
> 
> that the original kernel image still begins at 0x8000?

Why do you want to integrate the above code into the kernel?  Some reasons
not to:

1. it will remove the flexibility of the loader/kernel, especially when
   things go wrong.
2. upgrades will be difficult - you will have to place this code in each
   and every kernel
3. the change you suggest will never be in the generic kernel - the code
   in head-armv.S is *designed* to be usable on any platform, rather than
   having 5000000 different instances of it for everyone's slightly
   different idea of the world.
   _____
  |_____| ------------------------------------------------- ---+---+-
  |   |        Russell King       [EMAIL PROTECTED]      --- ---
  | | | |  http://www.arm.linux.org.uk/~rmk/armlinux.html    /  /  |
  | +-+-+                                                     --- -+-
  /   |               THE developer of ARM Linux              |+| /|\
 /  | | |                                                     ---  |
    +-+-+ -------------------------------------------------  /\\\  |
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]