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;
Date:   Tue, 24 May 2022 13:23:30 -0700
From:   Andrew Morton <akpm@linux-foundation.org>
To:     Zi Yan <ziy@nvidia.com>
Cc:     Zi Yan <zi.yan@sent.com>, Qian Cai <quic_qiancai@quicinc.com>,
        David Hildenbrand <david@redhat.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org,
        virtualization@lists.linux-foundation.org,
        Vlastimil Babka <vbabka@suse.cz>,
        Mel Gorman <mgorman@techsingularity.net>,
        Eric Ren <renzhengeek@gmail.com>,
        Mike Rapoport <rppt@kernel.org>,
        Oscar Salvador <osalvador@suse.de>,
        Christophe Leroy <christophe.leroy@csgroup.eu>
Subject: Re: [PATCH] mm: fix a potential infinite loop in
 start_isolate_page_range().
Message-Id: <20220524132330.eaf1366967d2fa927fdaf995@linux-foundation.org>
In-Reply-To: <20220524194756.1698351-1-zi.yan@sent.com>
References: <20220524194756.1698351-1-zi.yan@sent.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Tue, 24 May 2022 15:47:56 -0400 Zi Yan <zi.yan@sent.com> wrote:

> From: Zi Yan <ziy@nvidia.com>
> 
> In isolate_single_pageblock() called by start_isolate_page_range(),
> there are some pageblock isolation issues causing a potential
> infinite loop when isolating a page range. This is reported by Qian Cai.
> 
> 1. the pageblock was isolated by just changing pageblock migratetype
>    without checking unmovable pages. Calling set_migratetype_isolate() to
>    isolate pageblock properly.
> 2. an off-by-one error caused migrating pages unnecessarily, since the page
>    is not crossing pageblock boundary.
> 3. migrating a compound page across pageblock boundary then splitting the
>    free page later has a small race window that the free page might be
>    allocated again, so that the code will try again, causing an potential
>    infinite loop. Temporarily set the to-be-migrated page's pageblock to
>    MIGRATE_ISOLATE to prevent that and bail out early if no free page is
>    found after page migration.
> 
> An additional fix to split_free_page() aims to avoid crashing in
> __free_one_page(). When the free page is split at the specified
> split_pfn_offset, free_page_order should check both the first bit of
> free_page_pfn and the last bit of split_pfn_offset and use the smaller one.
> For example, if free_page_pfn=0x10000, split_pfn_offset=0xc000,
> free_page_order should first be 0x8000 then 0x4000, instead of 0x4000 then
> 0x8000, which the original algorithm did.
> 
> ...
>
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -1114,13 +1114,16 @@ void split_free_page(struct page *free_page,
>  	unsigned long flags;
>  	int free_page_order;
>  
> +	if (split_pfn_offset == 0)
> +		return;
> +
>  	spin_lock_irqsave(&zone->lock, flags);
>  	del_page_from_free_list(free_page, zone, order);
>  	for (pfn = free_page_pfn;
>  	     pfn < free_page_pfn + (1UL << order);) {
>  		int mt = get_pfnblock_migratetype(pfn_to_page(pfn), pfn);
>  
> -		free_page_order = ffs(split_pfn_offset) - 1;
> +		free_page_order = min(pfn ? __ffs(pfn) : order, __fls(split_pfn_offset));

Why is it testing the zeroness of `pfn' here?  Can pfn==0 even happen? 
If so, it's a legitimate value so why does it get special-cased?