Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <20180314005350.6xdda2uqzuy4n3o6@sasha-lappy>
References: <20180131210300.22963-1-pasha.tatashin@oracle.com>
 <20180131210300.22963-2-pasha.tatashin@oracle.com> <20180313234333.j3i43yxeawx5d67x@sasha-lappy>
 <CAGM2reaPK=ZcLBOnmBiC2-u86DZC6ukOhL1xxZofB2OTW3ozoA@mail.gmail.com> <20180314005350.6xdda2uqzuy4n3o6@sasha-lappy>
From:   Pavel Tatashin <pasha.tatashin@oracle.com>
Date:   Tue, 13 Mar 2018 21:02:53 -0400
Message-ID: <CAGM2reYo2EbH0W70rJGSGWRBAO=upcNDanBoCQgve+eQ_94C8A@mail.gmail.com>
Subject: Re: [PATCH v2 1/2] mm: uninitialized struct page poisoning sanity checking
To:     Sasha Levin <Alexander.Levin@microsoft.com>
Cc:     "steven.sistare@oracle.com" <steven.sistare@oracle.com>,
        "daniel.m.jordan@oracle.com" <daniel.m.jordan@oracle.com>,
        "akpm@linux-foundation.org" <akpm@linux-foundation.org>,
        "mgorman@techsingularity.net" <mgorman@techsingularity.net>,
        "mhocko@suse.com" <mhocko@suse.com>,
        "linux-mm@kvack.org" <linux-mm@kvack.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
        "vbabka@suse.cz" <vbabka@suse.cz>,
        "bharata@linux.vnet.ibm.com" <bharata@linux.vnet.ibm.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Mar 13, 2018 at 8:53 PM, Sasha Levin
<Alexander.Levin@microsoft.com> wrote:
> On Tue, Mar 13, 2018 at 08:38:57PM -0400, Pavel Tatashin wrote:
>>Hi Sasha,
>>
>>It seems the patch is doing the right thing, and it catches bugs. Here
>>we access uninitialized struct page. The question is why this happens?
>
> Not completely; note that we die on an invalid reference rather than
> assertion failure.

I think that invalid reference happens within assertion failure, as
far as I can tell, it is dump_page() where we get the invalid
reference, but to get to dump_page() from get_nid_for_pfn() we must
have triggered the assertion.

>
>>register_mem_sect_under_node(struct memory_block *mem_blk, int nid)
>>   page_nid = get_nid_for_pfn(pfn);
>>
>>node id is stored in page flags, and since struct page is poisoned,
>>and the pattern is recognized, the panic is triggered.
>>
>>Do you have config file? Also, instructions how to reproduce it?
>
> Attached the config. It just happens on boot.

Thanks, I will try in qemu.

Pasha