public inbox for ecos-discuss@sourceware.org
 help / color / mirror / Atom feed
* [ECOS] discontinguous memory setup
@ 2001-09-20  3:11 Stefan Eletzhofer
  2001-09-20  7:37 ` Jonathan Larmour
  0 siblings, 1 reply; 2+ messages in thread
From: Stefan Eletzhofer @ 2001-09-20  3:11 UTC (permalink / raw)
  To: ecos

Hello,
we use RedBoot do boot linux on our SA1110/1111 dev board, which is roughly
assabet+neponset. For now, we use RedBoot-Assabet config as a starting point.
RedBoot currently runs and boots linux correctly, uses flash etc (CF slot not
detected, tough).

Now i have two Questions:
1)
I hacked hal_platform_setup.h to provide correct memory timings & SDRAM
configuration, our board uses 2 banks of ram, each 16 MB:
0xc000000 - 16Mb
0xd000000 - 16Mb

I setup SA11X0_DRAM_CONFIGURATION correctly, but somehow RedBoot seems to
change the settings from 0x72557255 to 0x72547255 (Bank 2 Enabled -> disabled).

I  have not changed plf_mmap.h to reflect my memory settings, as i wonder how
RedBoot detects available memory (or is it a compile-time option?).

Does RedBoot support such discontinguous memory setups?

2)
If we decide to use RedBoot, how do i add Support for our Board? I understand
that hacking assabet code is not the Right Thing To Do, is it?

Thank you,
Stefan Eletzhofer

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [ECOS] discontinguous memory setup
  2001-09-20  3:11 [ECOS] discontinguous memory setup Stefan Eletzhofer
@ 2001-09-20  7:37 ` Jonathan Larmour
  0 siblings, 0 replies; 2+ messages in thread
From: Jonathan Larmour @ 2001-09-20  7:37 UTC (permalink / raw)
  To: stefan.eletzhofer; +Cc: ecos

Stefan Eletzhofer wrote:
> 
> Hello,
> we use RedBoot do boot linux on our SA1110/1111 dev board, which is roughly
> assabet+neponset. For now, we use RedBoot-Assabet config as a starting point.
> RedBoot currently runs and boots linux correctly, uses flash etc (CF slot not
> detected, tough).
> 
> Now i have two Questions:
> 1)
> I hacked hal_platform_setup.h to provide correct memory timings & SDRAM
> configuration, our board uses 2 banks of ram, each 16 MB:
> 0xc000000 - 16Mb
> 0xd000000 - 16Mb
> 
> I setup SA11X0_DRAM_CONFIGURATION correctly, but somehow RedBoot seems to
> change the settings from 0x72557255 to 0x72547255 (Bank 2 Enabled -> disabled).

Look a few lines further up at:

        // Enable DRAM controller
        ldr     r1,=SA11X0_DRAM_CONFIGURATION
        ldr     r2,=0x72547255
        str     r2,[r1]

That value should really be a macro, but that's the one you want to change.
 
> I  have not changed plf_mmap.h to reflect my memory settings, as i wonder how
> RedBoot detects available memory (or is it a compile-time option?).
> 
> Does RedBoot support such discontinguous memory setups?

For the assabet target, the memory map is all hard-coded I believe.

RedBoot has some trouble with discontiguous memory - it makes assumptions
when checking validity of RAM addresses. You may want to disable
CYGSEM_REDBOOT_VALIDATE_USER_RAM_LOADS

> 2)
> If we decide to use RedBoot, how do i add Support for our Board? I understand
> that hacking assabet code is not the Right Thing To Do, is it?

No reason why not unless you want to contribute it back. Starting from an
existing port is exactly the right thing to do. Or you can just create a
new target and package in ecos.db and copy your current "assabet" directory
to something else.

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2001-09-20  7:37 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-09-20  3:11 [ECOS] discontinguous memory setup Stefan Eletzhofer
2001-09-20  7:37 ` Jonathan Larmour

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).