[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z92Wf8SDt8lzFs7F@arm.com>
Date: Fri, 21 Mar 2025 16:40:31 +0000
From: Catalin Marinas <catalin.marinas@....com>
To: Zhenhua Huang <quic_zhenhuah@...cinc.com>
Cc: anshuman.khandual@....com, david@...hat.com, will@...nel.org,
ryan.roberts@....com, mark.rutland@....com, ardb@...nel.org,
yangyicong@...ilicon.com, joey.gouly@....com,
quic_cgoldswo@...cinc.com, quic_sudaraja@...cinc.com,
akpm@...ux-foundation.org, linux-mm@...ck.org,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
quic_tingweiz@...cinc.com
Subject: Re: [PATCH] arm64: mm: Correct the update of max_pfn
On Fri, Mar 21, 2025 at 03:00:19PM +0800, Zhenhua Huang wrote:
> Hotplugged memory can be smaller than the original memory. For example,
> on my target:
>
> root@...ericarmv8:~# cat /sys/kernel/debug/memblock/memory
> 0: 0x0000000064005000..0x0000000064023fff 0 NOMAP
> 1: 0x0000000064400000..0x00000000647fffff 0 NOMAP
> 2: 0x0000000068000000..0x000000006fffffff 0 DRV_MNG
> 3: 0x0000000088800000..0x0000000094ffefff 0 NONE
> 4: 0x0000000094fff000..0x0000000094ffffff 0 NOMAP
> max_pfn will affect read_page_owner. Therefore, it should first compare and
> then select the larger value for max_pfn.
>
> Fixes: 8fac67ca236b ("arm64: mm: update max_pfn after memory hotplug")
> Signed-off-by: Zhenhua Huang <quic_zhenhuah@...cinc.com>
> ---
> arch/arm64/mm/mmu.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 1dfe1a8efdbe..310ff75891ef 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -1361,7 +1361,8 @@ int arch_add_memory(int nid, u64 start, u64 size,
> __remove_pgd_mapping(swapper_pg_dir,
> __phys_to_virt(start), size);
> else {
> - max_pfn = PFN_UP(start + size);
> + /* Address of hotplugged memory can be smaller */
> + max_pfn = max(max_pfn, PFN_UP(start + size));
> max_low_pfn = max_pfn;
> }
I think this makes sense and also matches what x86 does. I'll queue it
at -rc1 as a fix.
Thanks.
--
Catalin
Powered by blists - more mailing lists