[RFC PATCH] ARM: Support allocating crashkernel above 4G for LPAE

[Baoquan He]

On 08/06/24 at 09:34am, Russell King (Oracle) wrote:
> On Tue, Aug 06, 2024 at 04:11:30PM +0800, Baoquan He wrote:
> > I am fine with it. BUT have you addressed Russell's concern, e.g how to
> > test it actually?
> 
> Thanks for bringing that up.
> 
> Let me reinforce my position on this. I will _not_ be accepting a patch
> that allows the crash kernel to be placed into high memory on 32-bit
> ARM unless it has been thoroughly tested to prove that it can actually
> work.
> 
> Right now, I don't believe it can work as placing the kernel in highmem
> likely means it will be located *outside* of the lower 4GiB of physical
> memory which is all that will be accessible when the MMU is turned off.
> This is a pre-condition to boot a kernel - the kernel image _must_ be
> located within a region of memory which is exposed to the CPU when the
> MMU is turned off.
> 
> Unless it can be proven that placing the kernel in highmem means that
> the kernel will be located in the lower 4GiB of physical memory space
> with the MMU off, then further work on this patch is a waste of time.

Yeah, totally agree.