Re: [ECOS] Can not get redboot to load / run linux binaries on IXP420

[Gary Thomas <gary@mlbassoc.com>]

On Tue, 2005-07-26 at 12:48 -0500, Don Osburn wrote:
> Mark,
> 
> Thanks again for the help so far.  I am really stumped.  I can not get
> redboot to load / run linux on our platform.  I have tried a number of
> things with no luck.  Can you or someone help out (please)?
> 
> Here's my config;  I am using Redboot v2.01.  Our platform is an IXP420 with
> 32MB RAM.
> 
> I originally used the redboot_ROM.bin image from the npe_enable_bin/ixdp425
> directory to boot our platform.  It booted properly (even though we only
> have 32MB of RAM).  I then attempted a
> 
> >load -r -v -b 0x01600000 zImage
> which worked fine.  Then I attempted a
> >load -r -v -b 0x00800000 ramdisk.gz
> which would hang up every time.  It appeared like the problem was that the
> load was writing to RAM being used by redboot, and that was causing it to
> hang.
> 
> Since I knew we had a different memory config, I rebuilt redboot, modifying
> the ixdp425.h file's SDRAM configs.  Specifically, I changed the following
> #defines;
> 
> #define SDRAM_SIZE                         0x2000000  // 32MB
> #define IXP425_SDRAM_CONFIG_INIT  (SDRAM_CONFIG_CAS_3 |
> SDRAM_CONFIG_4x16Mx16)
> 
> I used the SDRAM_CONFIG_4x16Mx16 because we determined from chapter 7 of the
> developer's manual that we needed to set the SDRAM config register to 0x11.
> The rebuild worked, and I was able to update the redboot image successfully.
> 
> Unfortunately, when I attempted to load the linux images, I got the exact
> same result as before.  zImage would load fine, but ramdisk.gz would hang
> loading it at 0x00800000.
> 
> At this point, I decided to try something different.  The ramdisk.gz image
> will load if I put it somewhere else.  Therefore, I did the following:
> 
> >load -r -v -b 0x01600000 zImage
> >fis create -b 0x01600000 -l 0xc0000 -e 0x00800000 -r 0x01600000 zimage
> >load -r -v -b 0x1000000 ramdisk.gz
> >fis create -b 0x1000000 -l 0x2e0000 -e 0x00800000 -r 0x1000000 ramdisk
> >fis load zimage
> >fis load ramdisk
> >exec -b 0x01600000
> 
> This time the kernel started to boot, but then it failed stating it could
> not find a ramdisk at location 0.  Given that, I reloaded the images and
> this time did a
> >exec -b 0x01600000 -l 0x2e0000 -r 0x100000 -x 0x2e0000
> and received the exact same results.  The kernel started, then died when it
> could not find a ramdisk a location 0.
> 
> There is one last thing we attempted.  We found a note in a mini-howto
> document that said for Redboot version 2.01, you are supposed to use
> 0x11600000 and 0x10800000 instead of 0x01600000 and 0x00800000 respectively.
> I attempted the load using these values and got similar results.  (It would
> hang during the load.)
> 
> At this point I am totally stumped.  How am I supposed to load the zImage
> and ramdisk.gz images into my IXP420 board?  How do I get linux up and
> running on this board using Redboot?
> 
> Help would be greatly appreciated.  This is grinding our project to a halt.

Are you sure that your memory is actually working as planned?
It sounds to me like you're getting some reflections when you
put things at 0x800000 and this is overwriting memory used by
RedBoot (which will be in low RAM and is documented when you
boot)

I'd try running a memory test like the one at:
  .../ecos/packages/hal/arm/edb7xxx/current/tests/dram_test.c

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------


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