RFI doesn't branch off?

[Amit Shah]

Amit Shah wrote:

>
> Hi all,
>
> I have a strange issue with a custom board running a 750GX on a
> Marvell 64360 bridge. I transferred control to the uImage via u-boot
> (which I had to modify for my board-specific UART and SDRAM inits,
> also disabling PCI).
>
> Using the UART that's initialized by u-boot, I put some putchar()s in
> head.S to figure out where the kernel was freezing. It seems 'rfi' in
> mmu_off doesn't jump off to '__after_mmu_off' as expected. I also put
> some pattern into SDRAM before an after mmu_off. I just see the
> pattern (via a PCI interface) put before mmu_off.
>
> The contents of SRR0 correctly point to __after_mmu_off just before
> the 'rfi'. However, the 'rfi' documentation says the value of SRR0 is
> concatenated with '0x0b00'. I don't think this should be done.
> Anyways, since this code is supposed to work, I'm not sure why it
> isn't working.

If I keep the DR enabled in MSR, the RFI goes back properly... disabling DR
causes some randomness.

>
> Can any u-boot initializations be causing this? I guess not, I've not
> touched any CPU-specific settings in u-boot.
>
> Can anyone give me some pointers?

--
Amit Shah
http://amitshah.nav.to/


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/