Re: [PATCH RESEND v3] kho: init alloc tags when restoring pages from reserved memory

[Pratyush Yadav <pratyush@kernel.org>]

On Thu, Jan 22 2026, ranxiaokai627@163.com wrote:

> From: Ran Xiaokai <ran.xiaokai@zte.com.cn>
>
> Memblock pages (including reserved memory) should have their allocation
> tags initialized to CODETAG_EMPTY via clear_page_tag_ref() before being
> released to the page allocator. When kho restores pages through
> kho_restore_page(), missing this call causes mismatched
> allocation/deallocation tracking and below warning message:
>
> alloc_tag was not set
> WARNING: include/linux/alloc_tag.h:164 at ___free_pages+0xb8/0x260, CPU#1: swapper/0/1
> RIP: 0010:___free_pages+0xb8/0x260
>  kho_restore_vmalloc+0x187/0x2e0
>  kho_test_init+0x3c4/0xa30
>  do_one_initcall+0x62/0x2b0
>  kernel_init_freeable+0x25b/0x480
>  kernel_init+0x1a/0x1c0
>  ret_from_fork+0x2d1/0x360
>
> Add missing clear_page_tag_ref() annotation in kho_restore_page() to
> fix this.
>
> Fixes: fc33e4b44b27 ("kexec: enable KHO support for memory preservation")
> Cc: <stable@vger.kernel.org>
> Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
> ---
>
> v2 -> v3: 
>  - also call clear_page_tag_ref() for non-compound order-0 tail pages
>
>  kernel/liveupdate/kexec_handover.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
>
> diff --git a/kernel/liveupdate/kexec_handover.c b/kernel/liveupdate/kexec_handover.c
> index d4482b6e3cae..96767b106cac 100644
> --- a/kernel/liveupdate/kexec_handover.c
> +++ b/kernel/liveupdate/kexec_handover.c
> @@ -255,6 +255,14 @@ static struct page *kho_restore_page(phys_addr_t phys, bool is_folio)
>  	if (is_folio && info.order)
>  		prep_compound_page(page, info.order);
>  
> +	/* Always mark headpage's codetag as empty to avoid accounting mismatch */
> +	clear_page_tag_ref(page);
> +	if (!is_folio) {
> +		/* Also do that for the non-compound tail pages */
> +		for (unsigned int i = 1; i < nr_pages; i++)
> +			clear_page_tag_ref(page + i);
> +	}
> +

I think it would be a little bit better if we just did this in the loop
above instead of looping again. But I think it is fine for now. I can
fix it up in the next release.

Reviewed-by: Pratyush Yadav <pratyush@kernel.org>

>  	adjust_managed_page_count(page, nr_pages);
>  	return page;
>  }

-- 
Regards,
Pratyush Yadav