Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Andrea Arcangeli <aarcange@redhat.com>
To:     Mel Gorman <mgorman@suse.de>,
        Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        Qian Cai <cai@lca.pw>
Cc:     Vlastimil Babka <vbabka@suse.cz>, Michal Hocko <mhocko@kernel.org>,
        David Hildenbrand <david@redhat.com>,
        linux-kernel@vger.kernel.org, Mike Rapoport <rppt@linux.ibm.com>,
        Baoquan He <bhe@redhat.com>
Subject: [PATCH 0/1] VM_BUG_ON_PAGE(!zone_spans_pfn) in set_pfnblock_flags_mask
Date:   Sat, 21 Nov 2020 14:45:05 -0500
Message-Id: <20201121194506.13464-1-aarcange@redhat.com>
In-Reply-To: <8C537EB7-85EE-4DCF-943E-3CC0ED0DF56D@lca.pw>
References: <8C537EB7-85EE-4DCF-943E-3CC0ED0DF56D@lca.pw>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hello,

After hitting this twice on two different systems, I'm now running
with the tentative fix applied, but it's not a meaningful test since
it's non reproducible.

However it is possible to inject this bug if you do "grep E820
/proc/iomem" and then find a phys addr there with a struct page
(i.e. pfn_valid) in a zone, with this change:

	min_pfn = pageblock_start_pfn(cc->free_pfn - (distance >> 1));
+	if (cc->zone is the zone where the e820 physaddr has a pfn_valid)
+	   min_pfn = physaddr_of_E820_non_RAM_page_with_valid_pfn >> PAGE_SHIFT;

I didn't try to inject the bug to validate the fix and it'd be great
if someone can try that to validate this or any other fix.

Andrea Arcangeli (1):
  mm: compaction: avoid fast_isolate_around() to set pageblock_skip on
    reserved pages

 mm/compaction.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)