Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <ed212c6b-a10a-0267-fb20-9a1691b9ba45@bytedance.com>
Date:   Wed, 11 May 2022 13:12:55 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:91.0)
 Gecko/20100101 Thunderbird/91.9.0
Subject: Re: [PATCH 1/2] mm: fix missing handler for __GFP_NOWARN
Content-Language: en-US
To:     akpm@linux-foundation.org
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        akinobu.mita@gmail.com, vbabka@suse.cz, gregkh@linuxfoundation.org,
        jirislaby@kernel.org, rostedt@goodmis.org
References: <20220510113809.80626-1-zhengqi.arch@bytedance.com>
From:   Qi Zheng <zhengqi.arch@bytedance.com>
In-Reply-To: <20220510113809.80626-1-zhengqi.arch@bytedance.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 2022/5/10 7:38 PM, Qi Zheng wrote:
> We expect no warnings to be issued when we specify __GFP_NOWARN, but
> currently in paths like alloc_pages() and kmalloc(), there are still
> some warnings printed, fix it.

Hi Andrew,

Maybe we only need to deal with memory allocation failures, such as
should_fail() (This leads to deadlock in PATCH[2/2]). These 
WARN_ON_ONCE()s report usage problems and should not be printed. If they
are printed, we should fix these usage problems.

Thanks,
Qi


> 
> Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
> ---
>   include/linux/fault-inject.h |  2 ++
>   lib/fault-inject.c           |  3 +++
>   mm/failslab.c                |  3 +++
>   mm/internal.h                | 11 +++++++++++
>   mm/page_alloc.c              | 22 ++++++++++++----------
>   5 files changed, 31 insertions(+), 10 deletions(-)
> 
> diff --git a/include/linux/fault-inject.h b/include/linux/fault-inject.h
> index 2d04f6448cde..9f6e25467844 100644
> --- a/include/linux/fault-inject.h
> +++ b/include/linux/fault-inject.h
> @@ -20,6 +20,7 @@ struct fault_attr {
>   	atomic_t space;
>   	unsigned long verbose;
>   	bool task_filter;
> +	bool no_warn;
>   	unsigned long stacktrace_depth;
>   	unsigned long require_start;
>   	unsigned long require_end;
> @@ -39,6 +40,7 @@ struct fault_attr {
>   		.ratelimit_state = RATELIMIT_STATE_INIT_DISABLED,	\
>   		.verbose = 2,						\
>   		.dname = NULL,						\
> +		.no_warn = false,					\
>   	}
>   
>   #define DECLARE_FAULT_ATTR(name) struct fault_attr name = FAULT_ATTR_INITIALIZER
> diff --git a/lib/fault-inject.c b/lib/fault-inject.c
> index ce12621b4275..423784d9c058 100644
> --- a/lib/fault-inject.c
> +++ b/lib/fault-inject.c
> @@ -41,6 +41,9 @@ EXPORT_SYMBOL_GPL(setup_fault_attr);
>   
>   static void fail_dump(struct fault_attr *attr)
>   {
> +	if (attr->no_warn)
> +		return;
> +
>   	if (attr->verbose > 0 && __ratelimit(&attr->ratelimit_state)) {
>   		printk(KERN_NOTICE "FAULT_INJECTION: forcing a failure.\n"
>   		       "name %pd, interval %lu, probability %lu, "
> diff --git a/mm/failslab.c b/mm/failslab.c
> index f92fed91ac23..58df9789f1d2 100644
> --- a/mm/failslab.c
> +++ b/mm/failslab.c
> @@ -30,6 +30,9 @@ bool __should_failslab(struct kmem_cache *s, gfp_t gfpflags)
>   	if (failslab.cache_filter && !(s->flags & SLAB_FAILSLAB))
>   		return false;
>   
> +	if (gfpflags & __GFP_NOWARN)
> +		failslab.attr.no_warn = true;
> +
>   	return should_fail(&failslab.attr, s->object_size);
>   }
>   
> diff --git a/mm/internal.h b/mm/internal.h
> index cf16280ce132..7a268fac6559 100644
> --- a/mm/internal.h
> +++ b/mm/internal.h
> @@ -35,6 +35,17 @@ struct folio_batch;
>   /* Do not use these with a slab allocator */
>   #define GFP_SLAB_BUG_MASK (__GFP_DMA32|__GFP_HIGHMEM|~__GFP_BITS_MASK)
>   
> +#define WARN_ON_ONCE_GFP(cond, gfp)	({				\
> +	static bool __section(".data.once") __warned;			\
> +	int __ret_warn_once = !!(cond);					\
> +									\
> +	if (unlikely(!(gfp & __GFP_NOWARN) && __ret_warn_once && !__warned)) { \
> +		__warned = true;					\
> +		WARN_ON(1);						\
> +	}								\
> +	unlikely(__ret_warn_once);					\
> +})
> +
>   void page_writeback_init(void);
>   
>   static inline void *folio_raw_mapping(struct folio *folio)
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 0e42038382c1..2bf4ce4d0e2f 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -3722,7 +3722,7 @@ struct page *rmqueue(struct zone *preferred_zone,
>   	 * We most definitely don't want callers attempting to
>   	 * allocate greater than order-1 page units with __GFP_NOFAIL.
>   	 */
> -	WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
> +	WARN_ON_ONCE_GFP((gfp_flags & __GFP_NOFAIL) && (order > 1), gfp_flags);
>   
>   	do {
>   		page = NULL;
> @@ -3799,6 +3799,9 @@ static bool __should_fail_alloc_page(gfp_t gfp_mask, unsigned int order)
>   			(gfp_mask & __GFP_DIRECT_RECLAIM))
>   		return false;
>   
> +	if (gfp_mask & __GFP_NOWARN)
> +		fail_page_alloc.attr.no_warn = true;
> +
>   	return should_fail(&fail_page_alloc.attr, 1 << order);
>   }
>   
> @@ -4346,7 +4349,8 @@ __alloc_pages_may_oom(gfp_t gfp_mask, unsigned int order,
>   	 */
>   
>   	/* Exhausted what can be done so it's blame time */
> -	if (out_of_memory(&oc) || WARN_ON_ONCE(gfp_mask & __GFP_NOFAIL)) {
> +	if (out_of_memory(&oc) ||
> +	    WARN_ON_ONCE_GFP(gfp_mask & __GFP_NOFAIL, gfp_mask)) {
>   		*did_some_progress = 1;
>   
>   		/*
> @@ -4902,8 +4906,8 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>   	 * We also sanity check to catch abuse of atomic reserves being used by
>   	 * callers that are not in atomic context.
>   	 */
> -	if (WARN_ON_ONCE((gfp_mask & (__GFP_ATOMIC|__GFP_DIRECT_RECLAIM)) ==
> -				(__GFP_ATOMIC|__GFP_DIRECT_RECLAIM)))
> +	if (WARN_ON_ONCE_GFP((gfp_mask & (__GFP_ATOMIC|__GFP_DIRECT_RECLAIM)) ==
> +				(__GFP_ATOMIC|__GFP_DIRECT_RECLAIM), gfp_mask))
>   		gfp_mask &= ~__GFP_ATOMIC;
>   
>   retry_cpuset:
> @@ -5117,7 +5121,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>   		 * All existing users of the __GFP_NOFAIL are blockable, so warn
>   		 * of any new users that actually require GFP_NOWAIT
>   		 */
> -		if (WARN_ON_ONCE(!can_direct_reclaim))
> +		if (WARN_ON_ONCE_GFP(!can_direct_reclaim, gfp_mask))
>   			goto fail;
>   
>   		/*
> @@ -5125,7 +5129,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>   		 * because we cannot reclaim anything and only can loop waiting
>   		 * for somebody to do a work for us
>   		 */
> -		WARN_ON_ONCE(current->flags & PF_MEMALLOC);
> +		WARN_ON_ONCE_GFP(current->flags & PF_MEMALLOC, gfp_mask);
>   
>   		/*
>   		 * non failing costly orders are a hard requirement which we
> @@ -5133,7 +5137,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>   		 * so that we can identify them and convert them to something
>   		 * else.
>   		 */
> -		WARN_ON_ONCE(order > PAGE_ALLOC_COSTLY_ORDER);
> +		WARN_ON_ONCE_GFP(order > PAGE_ALLOC_COSTLY_ORDER, gfp_mask);
>   
>   		/*
>   		 * Help non-failing allocations by giving them access to memory
> @@ -5379,10 +5383,8 @@ struct page *__alloc_pages(gfp_t gfp, unsigned int order, int preferred_nid,
>   	 * There are several places where we assume that the order value is sane
>   	 * so bail out early if the request is out of bound.
>   	 */
> -	if (unlikely(order >= MAX_ORDER)) {
> -		WARN_ON_ONCE(!(gfp & __GFP_NOWARN));
> +	if (WARN_ON_ONCE_GFP(order >= MAX_ORDER, gfp))
>   		return NULL;
> -	}
>   
>   	gfp &= gfp_allowed_mask;
>   	/*

-- 
Thanks,
Qi