Re: Linux on MPC5200 - LITE5200EVAL

Re: Linux on MPC5200 - LITE5200EVAL

[Wolfgang Denk]

In message <402E0D8C.4080801@246tNt.com> you wrote:
>
> The weird problem with u-boot ( not booting If I set bootargs manually )
> and the kernel crash are both solved. It was a compiler issue. I used a
> toolchain I used to build binaries for a "normal" mac based on gcc 3.4 (
> cvs ) and it was a really bad idea.

I see.

>  - The PCI slot is 3.3 v and apparently I don't have a single PCI card
> that seems to work on it ... When I plug one ( that I thought was 3.3v
> ), nothing boot, sometimes even the led don't power on ...

Be careful to put it correctly. The slot bracket has  to  be  on  the
side where the UART / USB / CAN connctors are. Do not insert the card
with the slot bracket pointing to the IDE connector.

Also be VERY carefully  which  card  you  insert.  make  sure  it  is
_really_ a 3.3V compatible card. [I fried a board by trying a Digitus
"DC  SER2 Rev. 2.0 Serial PCI I/O Card" - this card has the key slots
for a 3.3/5V combi card, but the V_I/O pins are connected to  the  5V
rail. The RAM chips didn't like to see 5V ...]

>  - When I change options in the kernel to support new stuff I need, like
> activating USB host & IDE controller, the kernel don't boot at all ( or
> at least, I don't see on the serial console it does ). It's the same
> problem with activating drivers for the PCI bus ... What is weird is
> that it's not an error on boot, it's just that after u-boot messages,
> there is nothing at all.

Can you attach a debugger? Or even try  a  post-mortem  dump  of  the
logbif area? Both PCI and USB are working fine for me.

> I'm not sure my hardware is fully OK, but I did mtest on u-boot  for
> about 10 hours and it didn't complained. It's a fresh new board from
> motorola...

Maybe you can ask for a replacement. Your problems don't look  normal
to me.

Best regards,

Wolfgang Denk

--
See us @ Embedded World, Nuremberg, Feb 17 - 19,  Hall 12.0 Booth 440
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Being schizophrenic is better than living alone.

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

Re: Linux on MPC5200 - LITE5200EVAL

[Sylvain Munaut]

Hi

>> - The PCI slot is 3.3 v and apparently I don't have a single PCI card
>>that seems to work on it ... When I plug one ( that I thought was 3.3v
>>), nothing boot, sometimes even the led don't power on ...
>>
>>
>
>Be careful to put it correctly. The slot bracket has  to  be  on  the
>side where the UART / USB / CAN connctors are. Do not insert the card
>with the slot bracket pointing to the IDE connector.
>
>Also be VERY carefully  which  card  you  insert.  make  sure  it  is
>_really_ a 3.3V compatible card. [I fried a board by trying a Digitus
>"DC  SER2 Rev. 2.0 Serial PCI I/O Card" - this card has the key slots
>for a 3.3/5V combi card, but the V_I/O pins are connected to  the  5V
>rail. The RAM chips didn't like to see 5V ...]
>
>
Yeah I can only imagine ...
I don't think my only test fried the card, it show the same behavior
before and after so I guess ( hope ? ) I'm lucky.

The slot does indeed work, I just came across a Intel EEPro100 card,
like the one I used on other dev kit. I'm sure it's 3.3v/5v compatible
and it's recognized without problems.

>> - When I change options in the kernel to support new stuff I need, like
>>activating USB host & IDE controller, the kernel don't boot at all ( or
>>at least, I don't see on the serial console it does ). It's the same
>>problem with activating drivers for the PCI bus ... What is weird is
>>that it's not an error on boot, it's just that after u-boot messages,
>>there is nothing at all.
>>
>>
>
>Can you attach a debugger? Or even try  a  post-mortem  dump  of  the
>logbif area? Both PCI and USB are working fine for me.
>
>
I don't have a debugger ... But I think I'll invest in one soon ! ;)
How can I dump the logbuf area ? I saw a discussion on this ml about
this buffer, is it fixed in memory ? How to find it's address ?

>>I'm not sure my hardware is fully OK, but I did mtest on u-boot  for
>>about 10 hours and it didn't complained. It's a fresh new board from
>>motorola...
>>
>>
>
>Maybe you can ask for a replacement. Your problems don't look  normal
>to me.
>
>
>
Could you send me a kernel image you compiled with the config at
http://www.246tNt.com/mpc5200_config ?
Maybe my compiler is wrong again. The ELDK binaries didn't work ( I
don't have the good glibc version on my machine ). I'm downloading what
I need to try to compile it right now ...
I've also put http://www.246tNt.com/uImage wich is not bootable on my
board ...

Basically the config is default + USB + IDE ( HD & CDRom ) + SCSI disk
support ( for memory sticks on USB ) + Misc filesystems and partition
support.

On http://www.246tNt.com/mpc5200_config  is a sample .config file that
makes the kernel to not boot at all.


Sylvain Munaut


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

Re: Linux on MPC5200 - LITE5200EVAL

[Dale Farnsworth]

On Sat, Feb 14, 2004 at 02:53:03PM +0000, Wolfgang Denk wrote:
>
> In message <402E0D8C.4080801@246tNt.com> you wrote:
> >
> > The weird problem with u-boot ( not booting If I set bootargs manually )
> > and the kernel crash are both solved. It was a compiler issue. I used a
> > toolchain I used to build binaries for a "normal" mac based on gcc 3.4 (
> > cvs ) and it was a really bad idea.
>
> I see.
>
> >  - The PCI slot is 3.3 v and apparently I don't have a single PCI card
> > that seems to work on it ... When I plug one ( that I thought was 3.3v
> > ), nothing boot, sometimes even the led don't power on ...
>
> Be careful to put it correctly. The slot bracket has  to  be  on  the
> side where the UART / USB / CAN connctors are. Do not insert the card
> with the slot bracket pointing to the IDE connector.
>
> Also be VERY carefully  which  card  you  insert.  make  sure  it  is
> _really_ a 3.3V compatible card. [I fried a board by trying a Digitus
> "DC  SER2 Rev. 2.0 Serial PCI I/O Card" - this card has the key slots
> for a 3.3/5V combi card, but the V_I/O pins are connected to  the  5V
> rail. The RAM chips didn't like to see 5V ...]
>
> >  - When I change options in the kernel to support new stuff I need, like
> > activating USB host & IDE controller, the kernel don't boot at all ( or
> > at least, I don't see on the serial console it does ). It's the same
> > problem with activating drivers for the PCI bus ... What is weird is
> > that it's not an error on boot, it's just that after u-boot messages,
> > there is nothing at all.
>
> Can you attach a debugger? Or even try  a  post-mortem  dump  of  the
> logbif area? Both PCI and USB are working fine for me.
>
> > I'm not sure my hardware is fully OK, but I did mtest on u-boot  for
> > about 10 hours and it didn't complained. It's a fresh new board from
> > motorola...
>
> Maybe you can ask for a replacement. Your problems don't look  normal
> to me.

I've seen the same symptoms Sylvain described on the five Lite5200
boards that I have tried.  Eepro100 network cards work fine, but the 3
brands of IDE cards I tried are not even visible in PCI config space.
U-boot doesn't see them and Linux doesn't see them, booted either from
dBUG or U-boot.

Actually, one Lite5200/IDE combo comes close.  Parts of the IDE cards
PCI config info do read correctly sometimes.  Different parts each time.
I think this is likely a hardware issue with Lite5200 boards.

-Dale Farnsworth

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

Re: Linux on MPC5200 - LITE5200EVAL & New bestcomm code ?

[Sylvain Munaut]

[-- Attachment #1: Type: text/plain, Size: 864 bytes --]

Hi every one.


A little update on my problems. Apparently the problem was the u-boot
command I used to boot :

I used :
     tftp 100000 mpc5200/uImage; tftp 300000 mpc5200/pRamdisk;bootm
100000 300000

And now I use :
     tftp 200000 mpc5200/uImage; tftp 400000 mpc5200/pRamdisk;bootm
200000 400000

and it works just fine ... Apparently when the kernel was too big, the
area near 100000 is probably used by u-boot in some way, maybe during
decompression and it doesnt work. I've check with multiple different
memory tests that the concerned memory region is fine. So it is a
software issue. Maybe u-boot should give some warning about that ( if
it's the cause, not yet sure ). But that's off topic here.



Now, I'm just looking forward for new BestComm code, to get AC97 and IDE
DMA running ( quite slow and CPU intensive right now ... ).



Sylvain Munaut



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 190 bytes --]

Re: Linux on MPC5200 - LITE5200EVAL & New bestcomm code ?

[Wolfgang Denk]

In message <40458A8C.8020009@246tNt.com> you wrote:
>
> and it works just fine ... Apparently when the kernel was too big, the
> area near 100000 is probably used by u-boot in some way, maybe during

Well, it is not used by U-Boot itself, but when you have a big  Linux
kernel it starts overwriting itself when you uncompress it to address
0.

> software issue. Maybe u-boot should give some warning about that ( if
> it's the cause, not yet sure ). But that's off topic here.

U-Boot does not know the uncompressed size in advance. And  anyway  -
it's  a  pilot  error  :-)  In modification of anmother famous quote:
"U-Boot was not designed  to  stop  you  from  doing  stupid  things,
because that would also stop you from doing clever things."

> Now, I'm just looking forward for new BestComm code, to get AC97 and IDE
> DMA running ( quite slow and CPU intensive right now ... ).

:-(

Best regards,

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Q:  How many DEC repairman does it take to fix a flat ?
A:  Five; four to hold the car up and one to swap tires.

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