This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: ARM 920t queries with eCos


> >"When system is booting MMU is off and bootROM is mapped at address 0.
> >Therefore bootMonitor (the boot loader) gets executed. After systems
boots,
> >MMU is turned on and a remapping happens. During remapping, RAM is mapped
to
> >address 0 and bootROM maps to somewhere else
>
> Yes.
>
> >(I suppose this should be EBI reason. Actully it is already there, at
boot time this is aliased to
> >bottom). The address I see in target.ld reflact address after remapping
has
> >taken place (MMU on)."
>
> Can you explain me, How the RAM gets to 0x0000.0000 during boot time? Is
it by some Hardware means in the EBI (like in Atmel ARM chips)?
Yes. I have 4-pole DIP swith on Integrator board, which changes memory
mapping.

>
> There are several advantages in having RAM mapped to VMA 0x0.
> Since you mentioned about the bootMonitor; Each third party boot monitor
(like Angel) might have its own method of dealing with the HW vectors (0x00
~ 0x20). It would be very difficult for eCOS or any other generic piece of
SW for that matter to work with each of these monitors. It might be a bad
idea to completely eliminate the onboard monitor (but platforms of different
archs do), so we keep the monitor until Redboot starts up & after which
complete control over the HW vectors is assumed by redboot by ensuring 100%
code path within Redboot/eCOS. If we have RAM at 0x0, then Reboot can fix
all the HW vectors with its own code, there by eliminating the monitor
completely. The onboard third party monitor gets dummied out at this point
<after Redboot starts>.
>
> If the RAM is at 0x0 even before Redboot starts, then VMA 0x0000.0000 will
or can be mapped to physical 0x0 (RAM).

This is really good information. Thanks. Does it  mean I can do following?
Excuse me if it is silly one.
I having bootMonitor as bootloader on Integrator board. It is very poor
compared to RedBoot as it does not have any network and debugging support.
So I thought of installing RedBoot on board. Rather than overwriting the
bootMonitor what I think to download RedBoot in flash through bootMonitor.
On next reset I map address 0 to flash (using 4-pole DIP swith), so that
redboot will be executed rather than bootMonitor. Is it ok?

Regards
Mohanlal

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


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]