Hi Geoff,
On 17/06/16 00:48, Geoff Levand wrote:
> Add three new files, kexec.h, machine_kexec.c and relocate_kernel.S to the
> arm64 architecture that add support for the kexec re-boot mechanism
> (CONFIG_KEXEC) on arm64 platforms.
> diff --git a/arch/arm64/kernel/machine_kexec.c b/arch/arm64/kernel/machine_kexec.c
> new file mode 100644
> index 0000000..a6dcb9c
> --- /dev/null
> +++ b/arch/arm64/kernel/machine_kexec.c
[ ... ]
> +
> +/**
> + * kexec_list_flush - Helper to flush the kimage list and source pages to PoC.
> + */
> +static void kexec_list_flush(struct kimage *kimage)
> +{
> + kimage_entry_t *entry;
> +
> + for (entry = &kimage->head; ; entry++) {
> + unsigned int flag;
> + void *addr;
> +
> + __flush_dcache_area(entry, sizeof(kimage_entry_t));
> +
> + flag = *entry & IND_FLAGS;
> + if (flag == IND_DONE)
> + break;
> +
> + addr = phys_to_virt(*entry & PAGE_MASK);
> +
> + switch (flag) {
> + case IND_INDIRECTION:
> + /* Set entry point just before the new list page. */
> + entry = (kimage_entry_t *)addr - 1;
Ah, after this restructuring we end up cleaning this page as we fall-through,
and one kimage_entry_t at a time as we walk the list.
break;
?
> + case IND_SOURCE:
> + __flush_dcache_area(addr, PAGE_SIZE);
> + break;
> + case IND_DESTINATION:
> + break;
> + default:
> + BUG();
> + }
> + }
> +}
> diff --git a/arch/arm64/kernel/relocate_kernel.S b/arch/arm64/kernel/relocate_kernel.S
> new file mode 100644
> index 0000000..51b73cd
> --- /dev/null
> +++ b/arch/arm64/kernel/relocate_kernel.S
[ ... ]
> +.ltorg
> +
> +.align 3 /* To keep the 64-bit values below naturally aligned. */
> +
> +.Lcopy_end:
> +.org KEXEC_CONTROL_PAGE_SIZE
On 16/06/16 23:41, Geoff Levand wrote:
> This is to check if arm64_relocate_new_kernel gets too
> big. The assembler should give an error if the location
> counter is set backwards.
Cunning. (probably worth a comment).
This looks like it is need because we hard-coded KEXEC_CONTROL_PAGE_SIZE in
kexec.h, so the code must be smaller than that size.
It looks like this value is only ever passed to get_order(), could we just pass
the actual size? It shouldn't matter if it grows >4K, as the core code will
allocate enough memory, and the memcpy() in machine_kexec() uses the size too.
Reviewed-by: James Morse <james.morse at arm.com>
Thanks,
James
