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;
Content-Type: text/plain;
        charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 12.4 \(3445.104.11\))
Subject: Re: [5.6.0-rc7] Kernel crash while running ndctl tests
From:   Sachin Sant <sachinp@linux.vnet.ibm.com>
In-Reply-To: <87a746cdva.fsf@linux.ibm.com>
Date:   Tue, 24 Mar 2020 15:06:31 +0530
Cc:     LKML <linux-kernel@vger.kernel.org>, linuxppc-dev@lists.ozlabs.org,
        Dan Williams <dan.j.williams@intel.com>,
        Baoquan He <bhe@redhat.com>, linux-nvdimm@lists.01.org
Content-Transfer-Encoding: 8BIT
References: <CF20E440-4DCB-4EFD-88B6-0AB98DC7FBD4@linux.vnet.ibm.com>
 <87a746cdva.fsf@linux.ibm.com>
To:     "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
Message-Id: <33E32320-C371-4A41-A3E1-4B9D2DDAFBFC@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On 24-Mar-2020, at 2:45 PM, Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> wrote:
> 
> Sachin Sant <sachinp@linux.vnet.ibm.com> writes:
> 
>> While running ndctl[1] tests against 5.6.0-rc7 following crash is encountered.
>> 
>> Bisect leads me to  commit d41e2f3bd546 
>> mm/hotplug: fix hot remove failure in SPARSEMEM|!VMEMMAP case
>> 
>> Reverting this commit helps and the tests complete without any crash.
> 
> 
> Can you try this change?
> 
> diff --git a/mm/sparse.c b/mm/sparse.c
> index aadb7298dcef..3012d1f3771a 100644
> --- a/mm/sparse.c
> +++ b/mm/sparse.c
> @@ -781,6 +781,8 @@ static void section_deactivate(unsigned long pfn, unsigned long nr_pages,
> 			ms->usage = NULL;
> 		}
> 		memmap = sparse_decode_mem_map(ms->section_mem_map, section_nr);
> +		/* Mark the section invalid */
> +		ms->section_mem_map &= ~SECTION_HAS_MEM_MAP;
> 	}
> 
> 	if (section_is_early && memmap)
> 

This patch works for me. The test ran successfully without any crash/failure.

Thanks
-Sachin

> a pfn_valid check involves pnf_section_valid() check if section is
> having MEM_MAP. In this case we did end up  setting the ms->uage = NULL.
> So when we do that tupdate the section to not have MEM_MAP.
> 
> -aneesh