Re: [PATCH v3 2/4] riscv: Add support for kexec handover (KHO)

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [High] riscv: kho: BUG_ON() in linear_mapping_pa_to_va() during
  efi_init()
- [Medium] riscv: kho: kho_reserve_bootmem() fails when DRAM starts at
  or above 4 GiB
--

commit 9359adbe095f746cadb4028ea618b75e0af09036
Author: Chen Pei <cp0613@linux.alibaba.com>

riscv: Add support for kexec handover (KHO)

This commit enables ARCH_SUPPORTS_KEXEC_HANDOVER on 64-bit RISC-V systems
with MMU support.

> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index bde1f1885038f..bffef6b903e13 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -1125,6 +1125,9 @@ config ARCH_SELECTS_KEXEC_FILE
>  config ARCH_SUPPORTS_KEXEC_PURGATORY
>  	def_bool ARCH_SUPPORTS_KEXEC_FILE
>  
> +config ARCH_SUPPORTS_KEXEC_HANDOVER
> +	def_bool 64BIT && MMU
> +
>  config ARCH_SUPPORTS_CRASH_DUMP
>  	def_bool y

[Severity: High]
Does enabling KEXEC_HANDOVER here hit BUG_ON(!kernel_map.va_pa_offset) in
linear_mapping_pa_to_va() during early boot on UEFI systems with
CONFIG_DEBUG_VIRTUAL=y?

In setup_arch(), parse_dtb() populates kho_in.fdt_phys, and then
efi_init() runs before paging_init():

arch/riscv/kernel/setup.c:setup_arch() {
    ...
	parse_dtb();
    ...
	efi_init();
	paging_init();
    ...
}

efi_init() -> reserve_regions() -> is_kho_boot() -> kho_get_fdt() calls
phys_to_virt(kho_in.fdt_phys), which translates to __va() ->
linear_mapping_pa_to_va().  On 64-bit RISC-V, kernel_map.va_pa_offset is
still 0 at that point because it is not initialized until paging_init() ->
setup_bootmem().

[Severity: Medium]
Can kho_reserve_bootmem() fail on 64-bit RISC-V systems where DRAM starts
at or above 4 GiB?

RISC-V does not define ARCH_LOW_ADDRESS_LIMIT, so it defaults to
0xffffffffUL in include/linux/memblock.h.  When all DRAM starts at or
above 4 GiB, bootmem_size_update() sets bootmem_size_lowmem to 0, and
kho_reserve_bootmem() unconditionally calls memblock_phys_alloc_range()
below ARCH_LOW_ADDRESS_LIMIT:

kernel/liveupdate/kexec_handover.c:kho_reserve_bootmem() {
    ...
	size = bootmem_size_lowmem;
	addr = memblock_phys_alloc_range(size, BOOTMEM_ALIGNMENT_BYTES, 0,
					 ARCH_LOW_ADDRESS_LIMIT);
	if (!addr) {
		pr_err("Failed to reserve lowmem bootmem\n");
		goto err_free_bootmem_desc;
	}
    ...
}

Because no memory exists below 4 GiB, the allocation returns 0 and aborts
KHO initialization.  Conversely, when all DRAM lies below 4 GiB,
bootmem_size_global is 0 and kho_reserve_bootmem() records a 0-byte entry
in kho_bootmem[], which causes kho_walk_bootmem() to build an inverted
struct resource with end = start - 1.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261106071459.2003-1-cp0613@linux.alibaba.com?part=2