[ECOS] Atmel EB40A eval board SRAM confusion

[ECOS] Atmel EB40A eval board SRAM confusion

[Tim Drury]

I just got an EB40A (not EB40) board.  According to the docs the onboard 
SRAM is not there by default from the factory - only footprints in which to
add your own.  However, mine has two Alliance AS7C34096 (512Kx8) srams
on board.  Lucky?  Not sure sure.  They are supposedly mapped to
0x02000000, but gdb gives "RDI_read: data abort, no memory at specified
address?" error.  

I did a sanity check with gdb to be sure it was working and was able to 
read/write to the PIO registers to turn the LEDs on and off.  Obviously
the eCOS EB40 target.ld _should_ work since it is trying to load the
image at 0x02000000 but I get an RDI_write error when I initially load
an image.  I'm using Angel right now, I haven't attempted RedBoot yet.

I've emailed Atmel asking about the SRAM, but I was wondering if anyone
else on the list has the EB40A and has working SRAM on board.  Is there
anything else I can try to get access to the SRAM?  Has anyone with the
EB40 had same or similar problems?

-tim


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Atmel EB40A eval board SRAM confusion

[Scott Dattalo]

On Sat, 8 Jun 2002, Tim Drury wrote:

> 
> I just got an EB40A (not EB40) board.  According to the docs the onboard 
> SRAM is not there by default from the factory - only footprints in which to
> add your own.  However, mine has two Alliance AS7C34096 (512Kx8) srams
> on board.  Lucky?  Not sure sure.  They are supposedly mapped to
> 0x02000000, but gdb gives "RDI_read: data abort, no memory at specified
> address?" error.  
> 
> I did a sanity check with gdb to be sure it was working and was able to 
> read/write to the PIO registers to turn the LEDs on and off.  Obviously
> the eCOS EB40 target.ld _should_ work since it is trying to load the
> image at 0x02000000 but I get an RDI_write error when I initially load
> an image.  I'm using Angel right now, I haven't attempted RedBoot yet.
> 
> I've emailed Atmel asking about the SRAM, but I was wondering if anyone
> else on the list has the EB40A and has working SRAM on board.  Is there
> anything else I can try to get access to the SRAM?  Has anyone with the
> EB40 had same or similar problems?

Tim,

I just got my EB40 (not EB40A) yesterday and I did not see this error. On
the EB40 there is a strap (LK1) to select the size of RAM. You may wish to
check verify that (or the EB40A equivalent jumper) is set correctly. 
Also, can you probe around other parts of the RAM with gdb (or Angel) to
find where the RAM really resides?  I know, it's the blind leading the
blind, but I'm getting an EB40A shortly too...

Scott


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Atmel EB40A eval board SRAM confusion

[Kjell Svensson]

Hi Tim,

I don't have an EB40A here at the moment, but when I used those about a 
half year ago, they certainly did have their external SRAMs mapped 
starting from 0x02000000.
Anyway, I'd first suggest that You should check in the EBI-regs of the 
AT91, to see what base address the SRAMs are mapped at.

And, yes - You should definitely feel lucky if You have been able to get 
Your hands on the EB40A model, 1meg is definitely more useful than the 
ordinary 256k ;-)

Cheers, /Kjell
-- 
Kjell Svensson                 Embedded Technology Manager
Techtribe Solutions AB         Tel:  +46 (0)31 706 06 00
Flöjelbergsgatan 12            GSM:  +46 (0)70 270 76 66
SE-431 37  MÖLNDAL             Mail: kjell@techtribe.se
Sweden




Tim Drury wrote:
> I just got an EB40A (not EB40) board.  According to the docs the onboard 
> SRAM is not there by default from the factory - only footprints in which to
> add your own.  However, mine has two Alliance AS7C34096 (512Kx8) srams
> on board.  Lucky?  Not sure sure.  They are supposedly mapped to
> 0x02000000, but gdb gives "RDI_read: data abort, no memory at specified
> address?" error.  
> 
> I did a sanity check with gdb to be sure it was working and was able to 
> read/write to the PIO registers to turn the LEDs on and off.  Obviously
> the eCOS EB40 target.ld _should_ work since it is trying to load the
> image at 0x02000000 but I get an RDI_write error when I initially load
> an image.  I'm using Angel right now, I haven't attempted RedBoot yet.
> 
> I've emailed Atmel asking about the SRAM, but I was wondering if anyone
> else on the list has the EB40A and has working SRAM on board.  Is there
> anything else I can try to get access to the SRAM?  Has anyone with the
> EB40 had same or similar problems?
> 
> -tim
> 
> 





-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Atmel EB40A eval board SRAM confusion

[Tim Drury]

I feel a little stupid.  Not 10 minutes after I sent that email that I thought
about looking at the EBI registers (it has been over 2 years since I've done
embedded work and I feel I've forgotten everything).  And, sure enough,
the CS1 config register was not setup for the SRAM.  For those that find
themselves in a similar situation, I put 0x02002001 in that register and got
the SRAM up.  I'm not sure about the wait states (all off) but it is 10ns SRAM
and appears to be reading/writing.

I haven't gotten "hello world" running yet as I ran out of time last night,
but the code image did appear to get into the SRAM.  I set a breakpoint
on the printf() and when I ran the code there appeared to be no error
messages in the gdb console, but it claimed "RDI_setbreak: no more
break/watch points available".  I look into this one more today.

Thanks for the help,

-tim


On Saturday 08 June 2002 05:09 am, Kjell Svensson wrote:
> Hi Tim,
>
> I don't have an EB40A here at the moment, but when I used those about a
> half year ago, they certainly did have their external SRAMs mapped
> starting from 0x02000000.
> Anyway, I'd first suggest that You should check in the EBI-regs of the
> AT91, to see what base address the SRAMs are mapped at.
>
> And, yes - You should definitely feel lucky if You have been able to get
> Your hands on the EB40A model, 1meg is definitely more useful than the
> ordinary 256k ;-)
>
> Cheers, /Kjell


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Atmel EB40A eval board SRAM confusion

[Gary Thomas]

On Sat, 2002-06-08 at 06:25, Tim Drury wrote:
> 
> I feel a little stupid.  Not 10 minutes after I sent that email that I thought
> about looking at the EBI registers (it has been over 2 years since I've done
> embedded work and I feel I've forgotten everything).  And, sure enough,
> the CS1 config register was not setup for the SRAM.  For those that find
> themselves in a similar situation, I put 0x02002001 in that register and got
> the SRAM up.  I'm not sure about the wait states (all off) but it is 10ns SRAM
> and appears to be reading/writing.
> 
> I haven't gotten "hello world" running yet as I ran out of time last night,
> but the code image did appear to get into the SRAM.  I set a breakpoint
> on the printf() and when I ran the code there appeared to be no error
> messages in the gdb console, but it claimed "RDI_setbreak: no more
> break/watch points available".  I look into this one more today.
> 
> Thanks for the help,
> 

What are you using for GDB?  Note that if it's Angel, eCos currently
has no support for I/O.  In other words, eCos will be expecting to talk
"normal GDB protocol" to print messages, which is not what Angel uses.
Your choices are to get RedBoot running on the board (which supports the
normal GDB protocol), or add Angel GDB support to eCos [a non-trivial
task].

> -tim
> 
> 
> On Saturday 08 June 2002 05:09 am, Kjell Svensson wrote:
> > Hi Tim,
> >
> > I don't have an EB40A here at the moment, but when I used those about a
> > half year ago, they certainly did have their external SRAMs mapped
> > starting from 0x02000000.
> > Anyway, I'd first suggest that You should check in the EBI-regs of the
> > AT91, to see what base address the SRAMs are mapped at.
> >
> > And, yes - You should definitely feel lucky if You have been able to get
> > Your hands on the EB40A model, 1meg is definitely more useful than the
> > ordinary 256k ;-)
> >
> > Cheers, /Kjell
> 
> 
> -- 
> Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss



-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss