[PATCH v6 5/6] mm/vmalloc: map contiguous pages in batches for vmap() if possible

[Dev Jain]

On 09/07/26 1:08 pm, Wen Jiang wrote:
> From: "Barry Song (Xiaomi)" <baohua at kernel.org>
> 
> In many cases, the pages passed to vmap() may include high-order
> pages. For example, the systemheap often allocates pages in descending
> order: order 8, then 4, then 0. Currently, vmap() iterates over every
> page individually—even pages inside a high-order block are handled
> one by one.
> 
> This patch detects physically contiguous pages (regardless of whether
> they are compound or non-compound) by scanning with
> num_pages_contiguous(), and maps them as a single contiguous block
> whenever possible. The mapping order is determined by taking the
> minimum of the contiguous page count and the pfn alignment, allowing
> graceful degradation when pfn alignment is less than the contiguous
> range.
> 
> Pages with the same page_shift are coalesced and mapped via
> vmap_pages_range_noflush_walk() to avoid page table rewalk.
> 
> As users typically allocate memory in descending orders (e.g.
> 8 → 4 → 0), once an order-0 page is encountered, we stop scanning
> for contiguous pages since subsequent pages are likely order-0 as well.
> 
> Signed-off-by: Barry Song (Xiaomi) <baohua at kernel.org>
> Co-developed-by: Dev Jain <dev.jain at arm.com>
> Signed-off-by: Dev Jain <dev.jain at arm.com>
> Signed-off-by: Wen Jiang <jiangwen6 at xiaomi.com>
> Tested-by: Xueyuan Chen <xueyuan.chen21 at gmail.com>
> Tested-by: Leo Yan <leo.yan at arm.com>
> ---
>  mm/vmalloc.c | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 85 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index d2a4d649af549..db0492151ad08 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -3543,6 +3543,89 @@ void vunmap(const void *addr)
>  }
>  EXPORT_SYMBOL(vunmap);
>  
> +static inline unsigned int vm_shift(pgprot_t prot, unsigned long size)
> +{
> +	if (arch_vmap_pmd_supported(prot) && size >= PMD_SIZE)
> +		return PMD_SHIFT;
> +
> +	return arch_vmap_pte_supported_shift(size);
> +}
> +
> +static inline int get_vmap_batch_order(struct page **pages,
> +		pgprot_t prot, unsigned int max_steps, unsigned int idx)
> +{
> +	unsigned int nr_contig;
> +	int order;
> +
> +	if (!IS_ENABLED(CONFIG_HAVE_ARCH_HUGE_VMAP))
> +		return 0;
> +
> +	nr_contig = num_pages_contiguous(&pages[idx], max_steps);
> +	if (nr_contig < 2)
> +		return 0;
> +
> +	order = ilog2(nr_contig);
> +
> +	/* Limit order by pfn alignment */
> +	order = min_t(int, order, __ffs(page_to_pfn(pages[idx])));

As said previously, you need to account for the case where pfn is zero.


> +
> +	if (vm_shift(prot, PAGE_SIZE << order) == PAGE_SHIFT)
> +		return 0;

Correct. There is no point in clamping the order by vm_shift(), because then
we won't be able to batch multiple cont blocks. So in case (PAGE_SIZE << order) is
a quantity less than CONT_PTE_SHIFT, simply bail out from batching.


> +
> +	return order;
> +}
> +
> +static int vmap_pages_range_batched(unsigned long addr, unsigned long end,
> +		pgprot_t prot, struct page **pages)
> +{
> +	unsigned int count = (end - addr) >> PAGE_SHIFT;
> +	unsigned int prev_shift = 0, idx = 0;
> +	unsigned long map_addr = addr, batch_end = addr;
> +	int err;
> +
> +	err = kmsan_vmap_pages_range_noflush(addr, end, prot, pages,
> +						PAGE_SHIFT, GFP_KERNEL);

Nit: Indentation. PAGE_SHIFT should start right below addr.

Apart from these two points, nothing jumped at me. So after these
are taken care of:

Reviewed-by: Dev Jain <dev.jain at arm.com>



> +	if (err)
> +		goto out;
> +
> +	for (unsigned int i = 0; i < count; ) {
> +		unsigned int shift = PAGE_SHIFT +
> +			get_vmap_batch_order(pages, prot, count - i, i);
> +
> +		if (!i)
> +			prev_shift = shift;
> +
> +		if (shift != prev_shift) {
> +			err = vmap_pages_range_noflush_walk(map_addr, batch_end,
> +					prot, pages + idx, prev_shift);
> +			if (err)
> +				goto out;
> +			prev_shift = shift;
> +			map_addr = batch_end;
> +			idx = i;
> +		}
> +
> +		/*
> +		 * Once small pages are encountered, the remaining pages
> +		 * are likely small as well.
> +		 */
> +		if (shift == PAGE_SHIFT)
> +			break;
> +
> +		batch_end += 1UL << shift;
> +		i += 1U << (shift - PAGE_SHIFT);
> +	}
> +
> +	/* Remaining */
> +	if (map_addr < end)
> +		err = vmap_pages_range_noflush_walk(map_addr, end,
> +				prot, pages + idx, prev_shift);
> +
> +out:
> +	flush_cache_vmap(addr, end);
> +	return err;
> +}
> +
>  /**
>   * vmap - map an array of pages into virtually contiguous space
>   * @pages: array of page pointers
> @@ -3586,8 +3669,8 @@ void *vmap(struct page **pages, unsigned int count,
>  		return NULL;
>  
>  	addr = (unsigned long)area->addr;
> -	if (vmap_pages_range(addr, addr + size, pgprot_nx(prot),
> -				pages, PAGE_SHIFT) < 0) {
> +	if (vmap_pages_range_batched(addr, addr + size, pgprot_nx(prot),
> +				pages) < 0) {
>  		vunmap(area->addr);
>  		return NULL;
>  	}