[PATCH 08/16] arm64/kexec: Add core kexec support

Pratyush Anand panand at redhat.com
Tue Oct 20 01:56:20 PDT 2015


Hi Geoff,

Thanks for the patches.

On 19/10/2015:11:38:53 PM, Geoff Levand wrote:
> +static void soft_restart(unsigned long addr)
> +{
> +	setup_mm_for_reboot();
> +	cpu_soft_restart(virt_to_phys(cpu_reset), addr,
> +		is_hyp_mode_available());

So now we do not flush cache for any memory region. Shouldn't we still flush
at least kernel and purgatory segments. 

kexec-tools loads a new kernel and purgatory executable. Some of those bits
might still be only in D-cache and we disable D-cache before control is passed
to the purgatory binary. Purgatory and some initial part of kernel code is
executed with D-cache disabled. So, We might land into a situation where correct
code is not executed while D-cache is disabled, no?

~Pratyush



More information about the kexec mailing list