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:   Sat, 15 Jul 2023 05:07:03 +0100
From:   Matthew Wilcox <willy@infradead.org>
To:     Sidhartha Kumar <sidhartha.kumar@oracle.com>
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        akpm@linux-foundation.org, linmiaohe@huawei.com,
        naoya.horiguchi@nec.com
Subject: Re: [PATCH mm-unstable 8/8] mm/memory-failure: convert
 unpoison_memory() to folios
Message-ID: <ZLIbZygG7LqSI9xe@casper.infradead.org>
References: <20230112204608.80136-1-sidhartha.kumar@oracle.com>
 <20230112204608.80136-9-sidhartha.kumar@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230112204608.80136-9-sidhartha.kumar@oracle.com>
Precedence: bulk

On Thu, Jan 12, 2023 at 02:46:08PM -0600, Sidhartha Kumar wrote:
> @@ -2348,7 +2347,6 @@ int unpoison_memory(unsigned long pfn)
>  		return -ENXIO;
>  
>  	p = pfn_to_page(pfn);
> -	page = compound_head(p);
>  	folio = page_folio(p);
>  
>  	mutex_lock(&mf_mutex);
> @@ -2360,31 +2358,31 @@ int unpoison_memory(unsigned long pfn)
>  		goto unlock_mutex;
>  	}
>  
> -	if (!PageHWPoison(p)) {
> +	if (!folio_test_hwpoison(folio)) {

This is wrong.  You need to test the individual page for PageHWPoison()
instead of testing the folio.  I understand that this will work for
hugetlbfs pages, but for THP, it's the individual pages that get poisoned,
and folio_test_hwpoison() on a THP actually only tests the head page.

Ideas for preventing this mistake in the future gratefully received.