Re: [PATCH v1] mm/memblock: Correct totalram_pages accounting with KMSAN

[Mike Rapoport <rppt@kernel.org>]

On Wed, Sep 17, 2025 at 02:32:50PM +0200, Alexander Potapenko wrote:
> When KMSAN is enabled, `kmsan_memblock_free_pages()` can hold back pages
> for metadata instead of returning them to the early allocator. The callers,
> however, would unconditionally increment `totalram_pages`, assuming the
> pages were always freed. This resulted in an incorrect calculation of the
> total available RAM, causing the kernel to believe it had more memory than
> it actually did.
> 
> This patch refactors `memblock_free_pages()` to return the number of pages
> it successfully frees. If KMSAN stashes the pages, the function now
> returns 0; otherwise, it returns the number of pages in the block.
> 
> The callers in `memblock.c` have been updated to use this return value,
> ensuring that `totalram_pages` is incremented only by the number of pages
> actually returned to the allocator. This corrects the total RAM accounting
> when KMSAN is active.
> 
> Cc: Aleksandr Nogikh <nogikh@google.com>
> Fixes: 3c2065098260 ("init: kmsan: call KMSAN initialization routines")
> Signed-off-by: Alexander Potapenko <glider@google.com>
> ---
>  mm/internal.h |  4 ++--
>  mm/memblock.c | 18 +++++++++---------
>  mm/mm_init.c  |  9 +++++----
>  3 files changed, 16 insertions(+), 15 deletions(-)
> 
> diff --git a/mm/internal.h b/mm/internal.h
> index 45b725c3dc030..ae1ee6e02eff9 100644
> --- a/mm/internal.h
> +++ b/mm/internal.h
> @@ -742,8 +742,8 @@ static inline void clear_zone_contiguous(struct zone *zone)
>  extern int __isolate_free_page(struct page *page, unsigned int order);
>  extern void __putback_isolated_page(struct page *page, unsigned int order,
>  				    int mt);
> -extern void memblock_free_pages(struct page *page, unsigned long pfn,
> -					unsigned int order);
> +extern unsigned long memblock_free_pages(struct page *page, unsigned long pfn,
> +					 unsigned int order);

No need for extern, the inconsistency is fine here.

>  extern void __free_pages_core(struct page *page, unsigned int order,
>  		enum meminit_context context);
>  
> diff --git a/mm/memblock.c b/mm/memblock.c
> index 117d963e677c9..de7ff644d8f4f 100644
> --- a/mm/memblock.c
> +++ b/mm/memblock.c
> @@ -1834,10 +1834,9 @@ void __init memblock_free_late(phys_addr_t base, phys_addr_t size)
>  	cursor = PFN_UP(base);
>  	end = PFN_DOWN(base + size);
>  
> -	for (; cursor < end; cursor++) {
> -		memblock_free_pages(pfn_to_page(cursor), cursor, 0);
> -		totalram_pages_inc();
> -	}
> +	for (; cursor < end; cursor++)
> +		totalram_pages_add(
> +			memblock_free_pages(pfn_to_page(cursor), cursor, 0));
>  }
>  
>  /*
> @@ -2259,9 +2258,11 @@ static void __init free_unused_memmap(void)
>  #endif
>  }
>  
> -static void __init __free_pages_memory(unsigned long start, unsigned long end)
> +static unsigned long __init __free_pages_memory(unsigned long start,
> +						unsigned long end)
>  {
>  	int order;
> +	unsigned long freed = 0;
>  
>  	while (start < end) {
>  		/*
> @@ -2279,10 +2280,11 @@ static void __init __free_pages_memory(unsigned long start, unsigned long end)
>  		while (start + (1UL << order) > end)
>  			order--;
>  
> -		memblock_free_pages(pfn_to_page(start), start, order);
> +		freed += memblock_free_pages(pfn_to_page(start), start, order);
>  
>  		start += (1UL << order);
>  	}
> +	return freed;
>  }
>  
>  static unsigned long __init __free_memory_core(phys_addr_t start,
> @@ -2297,9 +2299,7 @@ static unsigned long __init __free_memory_core(phys_addr_t start,
>  	if (start_pfn >= end_pfn)
>  		return 0;
>  
> -	__free_pages_memory(start_pfn, end_pfn);
> -
> -	return end_pfn - start_pfn;
> +	return __free_pages_memory(start_pfn, end_pfn);
>  }
>  
>  static void __init memmap_init_reserved_pages(void)
> diff --git a/mm/mm_init.c b/mm/mm_init.c
> index 5c21b3af216b2..9883612768511 100644
> --- a/mm/mm_init.c
> +++ b/mm/mm_init.c
> @@ -2548,24 +2548,25 @@ void *__init alloc_large_system_hash(const char *tablename,
>  	return table;
>  }
>  
> -void __init memblock_free_pages(struct page *page, unsigned long pfn,
> -							unsigned int order)
> +unsigned long __init memblock_free_pages(struct page *page, unsigned long pfn,
> +					 unsigned int order)

Please either align this with 'struct' or drop spaces and keep only tabs.

>  {
>  	if (IS_ENABLED(CONFIG_DEFERRED_STRUCT_PAGE_INIT)) {
>  		int nid = early_pfn_to_nid(pfn);
>  
>  		if (!early_page_initialised(pfn, nid))
> -			return;
> +			return 0;
>  	}
>  
>  	if (!kmsan_memblock_free_pages(page, order)) {
>  		/* KMSAN will take care of these pages. */
> -		return;
> +		return 0;
>  	}
>  
>  	/* pages were reserved and not allocated */
>  	clear_page_tag_ref(page);
>  	__free_pages_core(page, order, MEMINIT_EARLY);
> +	return 1UL << order;
>  }
>  
>  DEFINE_STATIC_KEY_MAYBE(CONFIG_INIT_ON_ALLOC_DEFAULT_ON, init_on_alloc);

-- 
Sincerely yours,
Mike.