[U-Boot] Latest u-boot-atmel git hangs on AVR32 board

[U-Boot] Latest u-boot-atmel git hangs on AVR32 board

[Mark Jackson]

Hi

Our MIMC200 board has been using 2009.08 for some time, so I thought I would take a look at the
latest git version to keep our boards more up-to-date.

I have checked out the u-boot-atmel branch, and done:-

$ make mimc200_config
$ make

Everything compiles, but when I try the image, I only get:-

U-Boot 2011.06-rc2-02816-g5f2e176 (Jun 13 2011 - 12:50:29)

U-Boot code: 00000000 -> 00012668  data: 0001ae38 -> 00050840
LCD: Frame buffer allocated at preset 0x10600000
VMR table @ 0x11fca7f4

I have traced through the code to see where things stop, and it seems to be in:-

arch/avr32/cpu/at32ap700x/mmu.c

void mmu_init_r(unsigned long dest_addr)
{
	...
	printf("VMR table @ 0x%08x\n", vmr_table_addr);

 	/* Enable paging */
puts("a");
	sysreg_write(MMUCR, SYSREG_BF(DRP, 1) | SYSREG_BF(DLA, 1)
			| SYSREG_BIT(MMUCR_S) | SYSREG_BIT(M) | SYSREG_BIT(E));
puts("b");
}

I get the "VMR table ..." string, as well as "a", but I never get "b".

Can anyone help ?

Regards
Mark

[U-Boot] Latest u-boot-atmel git hangs on AVR32 board

[Andreas Bießmann]

Dear Mark Jackson,

Am 13.06.2011 um 14:26 schrieb Mark Jackson:

> Hi

<snip mimc upgrate to 2011.06, something wrong with mmu_init_r()>

this is a quite new function, but used to work on stk1000 and our not mainline board.

> Can anyone help ?

My last try with avr32 boards was around 2011.03 release ... will have a look for it tomorrow.

regards

Andreas Bie?mann

[U-Boot] Latest u-boot-atmel git hangs on AVR32 board

[Andreas Bießmann]

Dear Mark Jackson,

Am 13.06.2011 15:28, schrieb Andreas Bie?mann:
> Dear Mark Jackson,
> 
> Am 13.06.2011 um 14:26 schrieb Mark Jackson:

>> Can anyone help ?
> 
> My last try with avr32 boards was around 2011.03 release ... will have a look for it tomorrow.

ATSTK1002 and our (not mainline) board do still work with
2011.06-rc2-00012-g9571865.

Maybe it is something with the 'simple tlb' -> your board is the only
one with 3 entries in struct mmu_vm_range
mmu_vmr_table[CONFIG_SYS_NR_VM_REGIONS]. This feature was heavily tested
on boards with 2 entries, therefore i guess you need to look for that
table mainly. Maybe there is also something wrong with
mmu_handle_tlb_miss() ... can you debug this?

It would be great to get this fixed for 2011.06 if possible.

regards

Andreas Bie?mann