Philip Blundell writes:
> >boot parameter. This results in hard-to-debug bad behavior during
> >bootup sequence -- around memory mapping setting.
>
> Yeah. It would be nice if someone could contribute some memory auto-sizing
> code actually. It shouldn't be that hard. One could probably assume a
> minimum of 4MB and go up from there.
I disagree here. One some architectures, eg EBSA285, the SDRAM size must be
known before the SDRAM is in use, and that means that the boot loader must
know. With the DC21285, it is not possible to probe the memory size and
organisation without disabling the SDRAM banks.
This information might as well be passed to the kernel by the loader code
anyway therefore.
Ok, you could find out what's currently set up, but I think that some of
the EBSA285 PBLs use fixed memory sizes, so increasing the SDRAMs with some
autoprobing code in the kernel will not be beneficial. (you'll still only
have 16MB available).
_____
|_____| ------------------------------------------------- ---+---+-
| | Russell King [EMAIL PROTECTED] --- ---
| | | | http://www.arm.linux.org.uk/~rmk/armlinux.html / / |
| +-+-+ --- -+-
/ | THE developer of ARM Linux |+| /|\
/ | | | --- |
+-+-+ ------------------------------------------------- /\\\ |
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
