[U-Boot-Users] what is wrong when calloc return bogus?

Wolfgang Denk wd at denx.de
Fri Aug 15 22:39:17 CEST 2003


In message <FD2AC9A020DDD51194710008C7089B200BEE22B9 at dlee17.itg.ti.com> you wrote:
> This has come up a couple times recently. I'll agree that from a consistency
> standpoint it would be nice if this were same for all architectures, and it
> will probably end up this way, but is it really "broken", what is not
> currently functioning, other than a proper mapping from the documentation to
> reality.

The design idea of U-Boot is that you  boot  from  flash,  initialize
RAM,  measure  the  size  of available RAM, and then use RAm in a top
down way. If we exclude more exotic features like pRAM,  log  buffer,
or  frame buffer memory that shall be reused by the Linux kernel this
means copy U-Boot code, data and bss to the end of RAM, below it  you
place  the malloc arena, and below that the stack (growing downward).
This is easy to implement, and provides maximum flexibility  even  in
cases where the memory size of a board is unknown at compile time.

Also, it allows the aforementioned  "exotic"  features  like  passing
information  from  U-Boot  (like POST results) in a syslog compatible
way to Linux, or initializing a framebuffer in U-Boot (with a  splash
screen  for  example)  and  keep  this  image  steady until the Linux
application start, or pRAM, where data can be stored which  will  not
be changed by a warmboot.



Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd at denx.de
Technology is dominated by those who manage what they do  not  under-
stand.




More information about the U-Boot mailing list