Re: [PATCH] mm/page_alloc: don't warn about large allocations with __GFP_NOFAIL

[Michal Hocko <mhocko@suse.com>]

On Wed 05-11-25 15:41:06, libaokun@huaweicloud.com wrote:
> From: Baokun Li <libaokun1@huawei.com>
> 
> Filesystems use __GFP_NOFAIL to allocate block-sized folios for metadata
> reads at critical points, since they cannot afford to go read-only,
> shut down, or enter an inconsistent state due to memory pressure.
> 
> Currently, attempting to allocate page units greater than order-1 with
> the __GFP_NOFAIL flag triggers a WARN_ON() in __alloc_pages_slowpath().
> However, filesystems supporting large block sizes (blocksize > PAGE_SIZE)
> can easily require allocations larger than order-1.
> 
> As Matthew Wilcox noted, if we have a filesystem with 64KiB sectors, there
> will be many clean folios in the page cache that are 64KiB or larger.
> 
> With gfp flags and order already included in the OOM report, both
> Vlastimil Babka and Michal Hocko suggested that we can take the risk of
> removing this warning first and then observe whether a large number of
> related OOM reports appear.
> 
> If that happens, we can consider adding special handling in other places.
> 
> Suggested-by: Matthew Wilcox <willy@infradead.org>
> Link: https://lore.kernel.org/all/aQPX1-XWQjKaMTZB@casper.infradead.org
> Suggested-by: Vlastimil Babka <vbabka@suse.cz>
> Link: https://lore.kernel.org/all/188a95ba-6384-4319-bb74-c0d9ec6c4079@suse.cz
> Suggested-by: Michal Hocko <mhocko@suse.com>
> Link: https://lore.kernel.org/all/aQotQBjnDDeL_wHx@tiehlicka
> Signed-off-by: Baokun Li <libaokun1@huawei.com>

Thanks for referencing above links which are providing a useful insight
for future reference. I would just add a link to Matthew explanation why
kvmalloc is not an option
Link: https://lore.kernel.org/all/aQTHMI3t5mNXp0M1@casper.infradead.org/T/#u
Acked-by: Michal Hocko <mhocko@suse.com>

> ---
> 
> RFC: https://lore.kernel.org/all/20251031061350.2052509-1-libaokun@huaweicloud.com
> 
>  mm/page_alloc.c | 5 -----
>  1 file changed, 5 deletions(-)
> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index fb91c566327c..e4efda1158b2 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -4683,11 +4683,6 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>  	int reserve_flags;
>  
>  	if (unlikely(nofail)) {
> -		/*
> -		 * We most definitely don't want callers attempting to
> -		 * allocate greater than order-1 page units with __GFP_NOFAIL.
> -		 */
> -		WARN_ON_ONCE(order > 1);
>  		/*
>  		 * Also we don't support __GFP_NOFAIL without __GFP_DIRECT_RECLAIM,
>  		 * otherwise, we may result in lockup.
> -- 
> 2.46.1

-- 
Michal Hocko
SUSE Labs