Abatron Bdi2000 and 405GP

Abatron Bdi2000 and 405GP

[ibm405gp]

Hello,

How can i read back (or write) for example, the FPGA registers from the
walnut board (at @OxFO300000) with abatron.
I can read this memory area just after the ioremap operation (in
walnut.c) when the entry for 0XF0300000 is present
in the TLB without any problem (0xF0300000 -> 0xF0300000). With the
abatron debugger, i should be able to read any virtual adress, even if
this adress is not present in the TLB. Am i right ? Any suggestion ?
Thanks.

I'am using the latest version of bdiGdb (1.07), with the folllowing
config file :

[INIT]
; Setup MMU info
WM32    0x000000f4  0x00000000  ;invalidate kernel  page table base
WM32    0x000000f8  0x00000000  ;invalidate process page table base
WM32    0x000000f0  0xc00000f4  ;invalidate page table base

[TARGET]
MMU         XLAT 0xC0000000     ;enable virtual address mode
PTBASE      0x000000f0          ;address where kernel/user stores
pointer to page table

the hhl 2.0 from monta vista with the following head_4xx.S

	/* Add helper information for the Abatron bdiGDB debugger.
	*/
	lis	r5, abatron_pteptrs@ha
	ori	r5, r5, abatron_pteptrs@l
	stw	r5, 0xf0(r0)	/* This much match your Abatron config */
	lis	r6, swapper_pg_dir@h
	ori	r6, r6, swapper_pg_dir@l
	tophys(r5, r5)
	stw	r6, 0(r5)

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

Re: Abatron Bdi2000 and 405GP

[Dan Malek]

ibm405gp wrote:

> ....i should be able to read any virtual adress, even if
> this adress is not present in the TLB. Am i right ? Any suggestion ?

Yes, yes.  Make sure you have the proper Abatron firmware.  We kind of
screwed up the Abatron folks when we moved some bits in the Linux PTE,
so be patient with them.  It wasn't their fault.

> 	lis	r5, abatron_pteptrs@ha


Uh, wait a minute....I'll bet you have the wrong kernel sources.  There is
a different set of sources for the 4xx (and other processors) from MontaVista
that fixes this cut'n'paste error I made here.  Ensure you are using the
proper source CD that matches the processor LSPs you have.


	-- Dan


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

405GP interrupts tip and looking for a little console help

[Mark Pilon]

Hi,

I'm bringing up the _devel kernel on a custom ppc405 board and
wanted to offer a little tip which might save some of you some
time.

if your boot rom doesn't have much to do w/ interrupts (and mine
doesn't) nothing will set the UIC polarity register so that the
UIC responds correctly to on-chip interrupts.  I mention the boot
rom as it seems linux depends upon this setup happening upstream
of it.

In my case, failing to set all the UIC_PR bits for on-chip peripherals
to 1 meant I kept getting unexpected interrupts.  the first
occasion of this was initing the ethernet controller ...

in any case, write the correct value for your processor to
DCRN_UICPR -- w/ a mask of 1s for each on-chip peripheral.

AND I'd appreciate a little help on getting my console to
talk w/ me -- I boot, mount a remote target filesystem w/ nfs,
and get console messages right up to:

NET4: Unix domain sockets 1.0/SMP for Linux
NET4.0.
Looking up port of RPC 100003/2 on
192.168.200.4
Looking up port of RPC 100005/1 on
192.168.200.4
VFS: Mounted root (nfs
filesystem).
Freeing unused kernel memory: 56k
init
echo
*
^^^^^^----- I typed this last line and it echo'd,
I passed in a command init=/sbin/sash, but don't get its prompt,

I debugged it to the point of seeing the exec of /bin/sash,
and would appreciate any views as to what I've forgotten here.

thanks,

Mark

--
Mark Pilon

Minolta-QMS
P.O. Box 37
Fallon, MT.  59326-0037

1-406-853-0433

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

Re: 405GP interrupts tip and looking for a little console help

[Phillip Lougher]

On Wed, 7 Nov 2001, Mark Pilon wrote:

>
> Hi,
>
> AND I'd appreciate a little help on getting my console to
> talk w/ me -- I boot, mount a remote target filesystem w/ nfs,
> and get console messages right up to:
>
> NET4: Unix domain sockets 1.0/SMP for Linux
> NET4.0.
> Looking up port of RPC 100003/2 on
> 192.168.200.4
> Looking up port of RPC 100005/1 on
> 192.168.200.4
> VFS: Mounted root (nfs
> filesystem).
> Freeing unused kernel memory: 56k
> init
> echo
> *
> ^^^^^^----- I typed this last line and it echo'd,
> I passed in a command init=/sbin/sash, but don't get its prompt,
>
> I debugged it to the point of seeing the exec of /bin/sash,
> and would appreciate any views as to what I've forgotten here.
>

I had this problem a while ago on a Redwood-4 board using an early
revision of the processor (STB3400 a variant of the 405GC).  Like
you I tried sash and debugged it up to the point of seeing the exec, then
nothing else would happen.

In my case, the exec was going ahead, but because of a processor
hardware bug, the process was being killed before any output was
produced.  Hence the apparent console output problem.

I finally identified that exec'ing and console output was working by
writing a very small init process (< 2K text/data) that did nothing
but print a string by calling the write system call directly in assembler.
This was small enough that it executed without causing the hardware bug.

Upgrading to a later revision of the processor cured the problem.

This may be nothing to do with your problem, but it may be worth a try.

Phillip

-----
Phillip Lougher

Zarlink Semiconductor,
Portskewitt, Wales, UK.


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