On Fri, Jul 17, 2009 at 01:37:14PM +0200, Wolfgang Denk wrote:
Dear Detlev Zundel,
In message m2r5wfzh0l.fsf@ohwell.denx.de you wrote:
I also think that making our lifes easier is a very good thing. Our "it's only a bootloader" codebase is pretty complex already so I welcome making it more robust. I was kind of surprised the other day that null-pointer jumps(*) did not result in immediate "bad programmer, no cookie this time" messages but in random crashes.
Well, what do you suggest? remember for example that booting a Linux kernel means jumping to it's entry point address - which happens to be 0 - or is this NULL ? How do you detemine the difference?
The MMU configuration that the OS wants on entry does not necessarily have to be the same one that the majority of U-Boot code runs with.
-Scott