[U-Boot] MPC8377: USB breaks board

[U-Boot] MPC8377: USB breaks board

[Andre Schwarz]

Jocke,

sorry to bother you again, but it might relate to our last discussion.

Actually my MPC8377 based board is up and running with basic functionality.
Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
defined serial line is pretty dead again.

This means again there's something wrong at the very beggining.

Diff'ing the System.map gives almost identical layout up to cpu_init_f 
... which adds USB specific code making the rest move, of course.


The only difference is the location of __got2_entries :

WORKING:
000004a5 A __fixup_entries
0000070b A __got2_entries
fc000044 T version_string
[...]

NOT WORKING:
000004a5 A __fixup_entries
00000712 A __got2_entries
fc000044 T version_string
[...]


What makes __got2_entries move ?

Why is __fixup_entries completely unaligned ? Looks pretty unusual to me.


Any idea if this can lead to malfunction ?



Regards,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

[U-Boot] MPC8377: USB breaks board

[Joakim Tjernlund]

Andre Schwarz <andre.schwarz@matrix-vision.de> wrote on 2010/11/05 15:51:13:
>
> Jocke,
>
> sorry to bother you again, but it might relate to our last discussion.
>
> Actually my MPC8377 based board is up and running with basic functionality.
> Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
> defined serial line is pretty dead again.
>
> This means again there's something wrong at the very beggining.
>
> Diff'ing the System.map gives almost identical layout up to cpu_init_f
> ... which adds USB specific code making the rest move, of course.
>
>
> The only difference is the location of __got2_entries :
>
> WORKING:
> 000004a5 A __fixup_entries
> 0000070b A __got2_entries
> fc000044 T version_string
> [...]
>
> NOT WORKING:
> 000004a5 A __fixup_entries
> 00000712 A __got2_entries
> fc000044 T version_string
> [...]
>
>
> What makes __got2_entries move ?

*_entries are not addresses, they are a count(how many entries)
See the linker file.

[U-Boot] MPC8377: USB breaks board

[Joakim Tjernlund]

Andre Schwarz <andre.schwarz@matrix-vision.de> wrote on 2010/11/05 15:51:13:
>
> Jocke,
>
> sorry to bother you again, but it might relate to our last discussion.
>
> Actually my MPC8377 based board is up and running with basic functionality.
> Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
> defined serial line is pretty dead again.

Perhaps USB becomes the console in this case and that is why you don't see anything?

[U-Boot] MPC8377: USB breaks board

[Andre Schwarz]

On 11/05/2010 04:00 PM, Joakim Tjernlund wrote:
> Andre Schwarz<andre.schwarz@matrix-vision.de>  wrote on 2010/11/05 15:51:13:
>    
>> Jocke,
>>
>> sorry to bother you again, but it might relate to our last discussion.
>>
>> Actually my MPC8377 based board is up and running with basic functionality.
>> Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
>> defined serial line is pretty dead again.
>>
>> This means again there's something wrong at the very beggining.
>>
>> Diff'ing the System.map gives almost identical layout up to cpu_init_f
>> ... which adds USB specific code making the rest move, of course.
>>
>>
>> The only difference is the location of __got2_entries :
>>
>> WORKING:
>> 000004a5 A __fixup_entries
>> 0000070b A __got2_entries
>> fc000044 T version_string
>> [...]
>>
>> NOT WORKING:
>> 000004a5 A __fixup_entries
>> 00000712 A __got2_entries
>> fc000044 T version_string
>> [...]
>>
>>
>> What makes __got2_entries move ?
>>      
> *_entries are not addresses, they are a count(how many entries)
> See the linker file.
>
>    
ok - I'm no expert ... that's why I have to ask.

Cheers,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

[U-Boot] MPC8377: USB breaks board

[Andre Schwarz]

On 11/05/2010 04:07 PM, Joakim Tjernlund wrote:
> Andre Schwarz<andre.schwarz@matrix-vision.de>  wrote on 2010/11/05 15:51:13:
>    
>> Jocke,
>>
>> sorry to bother you again, but it might relate to our last discussion.
>>
>> Actually my MPC8377 based board is up and running with basic functionality.
>> Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
>> defined serial line is pretty dead again.
>>      
> Perhaps USB becomes the console in this case and that is why you don't see anything?
>
>    

Hmm - how can this happen ?

I have ...

#define CONFIG_CONS_INDEX       1
#define CONFIG_SYS_NS16550
#define CONFIG_SYS_NS16550_SERIAL
#define CONFIG_SYS_NS16550_REG_SIZE     1
#define CONFIG_SYS_NS16550_CLK          get_bus_freq(0)

#define CONFIG_SYS_BAUDRATE_TABLE  \
         {300, 600, 1200, 2400, 4800, 9600, 19200, 38400, 57600, 115200}

#define CONFIG_SYS_NS16550_COM1 (CONFIG_SYS_IMMR+0x4500)
#define CONFIG_SYS_NS16550_COM2 (CONFIG_SYS_IMMR+0x4600)

#define CONFIG_CONSOLE          ttyS0
#define CONFIG_BAUDRATE         115200

... like on any other MPC83xx based board.


Have I missed something else ?


Regards,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

[U-Boot] MPC8377: USB breaks board

[Joakim Tjernlund]

Andre Schwarz <andre.schwarz@matrix-vision.de> wrote on 2010/11/05 17:13:34:
> On 11/05/2010 04:07 PM, Joakim Tjernlund wrote:
> > Andre Schwarz<andre.schwarz@matrix-vision.de>  wrote on 2010/11/05 15:51:13:
> >
> >> Jocke,
> >>
> >> sorry to bother you again, but it might relate to our last discussion.
> >>
> >> Actually my MPC8377 based board is up and running with basic functionality.
> >> Now I'm trying to add USB support - but as soon as CONFIG_USB_EHCI_FSL is
> >> defined serial line is pretty dead again.
> >>
> > Perhaps USB becomes the console in this case and that is why you don't see anything?
> >
> >
>
> Hmm - how can this happen ?

Beats me, it was just an idea. I am not using USB at all
so I really don't know what I am talking about :)

[U-Boot] MPC8377: USB breaks board

[Andre Schwarz]

Kim, Timur,

some comments on this from your side would really help !


The issue can be summarized like this :

- MPC8377 based board is basically running fine with both U-Boot and Linux.
- Activating USB support within U-Boot leads to *very* early crash :
      No console output - not even PCI clocks enabled.
- Within Linux USB is working without problems, i.e. there's no hardware 
issue.

I have an SMSC USB3300 PHY connected via ULPI and a self powered SMSC 
2514 1:4 Hub.
This is exactly the same as on mvblm7 board (=MPC8343).

This behaviour reminds me of e.g. one or more missing sync instructions 
during early init as discussed with Jocke in the past weeks.

It is very unlikely that I'm the only one facing this problem.


Can you at least tell me whether there are 837x boards known to work 
with latest code utilizing USB and Nor-Boot ?


Regards,
Andr?


MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

[U-Boot] MPC8377: USB breaks board

[Timur Tabi]

On Thu, Nov 11, 2010 at 6:24 AM, Andre Schwarz
<andre.schwarz@matrix-vision.de> wrote:

> - Activating USB support within U-Boot leads to *very* early crash :
> ? ? ?No console output - not even PCI clocks enabled.

I don't know anything about USB or the 8377, but it could be that
you're U-Boot image is too large.  Try disabling some other large
feature (like PCI and Ethernet support) and booting that image.

-- 
Timur Tabi
Linux kernel developer at Freescale

[U-Boot] MPC8377: USB breaks board

[Schwarz, Andre]

Timur,
?

> > - Activating USB support within U-Boot leads to *very* early crash :
> > ? ? ?No console output - not even PCI clocks enabled.
>
> I don't know anything about USB or the 8377, but it could be that
> you're U-Boot image is too large.? Try disabling some other large
> feature (like PCI and Ethernet support) and booting that image.?
I'm at 300KiB at the moment with plenty of flash.
Is there anything specific you're thinking about ?
?
?
Will try to dig into it by disabling USB init functions step-by-step then ...
?
?
Cheers,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

[U-Boot] MPC8377: USB breaks board

[Timur Tabi]

On Thu, Nov 11, 2010 at 2:11 PM, Schwarz,Andre
<andre.schwarz@matrix-vision.de> wrote:

> I'm at 300KiB at the moment with plenty of flash.
>
> Is there anything specific you're thinking about ?

No, it's just wild speculation.

-- 
Timur Tabi
Linux kernel developer at Freescale

[U-Boot] MPC8377: USB breaks board

[Andre Schwarz]

On 11/11/2010 01:24 PM, Andre Schwarz wrote:
> Kim, Timur,
>
> some comments on this from your side would really help !
>
>
> The issue can be summarized like this :
>
> - MPC8377 based board is basically running fine with both U-Boot and Linux.
> - Activating USB support within U-Boot leads to *very* early crash :
>        No console output - not even PCI clocks enabled.
> - Within Linux USB is working without problems, i.e. there's no hardware
> issue.
>
> I have an SMSC USB3300 PHY connected via ULPI and a self powered SMSC
> 2514 1:4 Hub.
> This is exactly the same as on mvblm7 board (=MPC8343).
>
> This behaviour reminds me of e.g. one or more missing sync instructions
> during early init as discussed with Jocke in the past weeks.
>
> It is very unlikely that I'm the only one facing this problem.
>    

got it - line 330 in arch/powerpc/cpu/mpc83xx/cpu_init.c does some EHCI 
configs masked by

#ifndef CONFIG_MPC834x

Removing this makes the board work again.

It still hangs during "usb start" but this is most propably due to 
missing initialization.

Will have a look and send a proper patch.

After all I do like mondays ;-)

Cheers,
Andr?


MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner