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;
Message-ID: <1f26411c-34b9-14e8-66eb-4854d5b1fbbb@linux.ibm.com>
Date:   Wed, 2 Aug 2023 15:57:49 +0530
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.13.0
Subject: Re: [next-20230731] Kdump fails to capture vmcore (powerpc)
To:     Sachin Sant <sachinp@linux.ibm.com>,
        linuxppc-dev <linuxppc-dev@lists.ozlabs.org>, linux-mm@kvack.org
Cc:     "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>,
        Mahesh J Salgaonkar <mahesh@linux.ibm.com>,
        Sourabh Jain <sourabhjain@linux.ibm.com>,
        open list <linux-kernel@vger.kernel.org>
References: <F996AB3C-400E-447A-B142-9CB3600BCDD8@linux.ibm.com>
Content-Language: en-US
From:   Hari Bathini <hbathini@linux.ibm.com>
In-Reply-To: <F996AB3C-400E-447A-B142-9CB3600BCDD8@linux.ibm.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 31/07/23 7:39 pm, Sachin Sant wrote:
> Kernel Crash dump(kdump) and firmware assisted dump on powerpc
> fails to capture vmcore on recent linux-next builds.
> 
> Starting Kdump Vmcore Save Service...
> systemd[1]: Starting Kdump Vmcore Save Service...
> kdump[599]: Kdump is using the default log level(3).
> kdump[635]: saving to /sysroot/var/crash/127.0.0.1-2023-07-31-09:01:28/
> kdump[639]: Remounting the dump target in rw mode.
> kdump[643]: saving vmcore-dmesg.txt to /sysroot/var/crash/127.0.0.1-2023-07-31-09:01:28/
> kdump[649]: saving vmcore-dmesg.txt complete
> kdump[651]: saving vmcore
> kdump.sh[652]: readpage_elf: Attempt to read non-existent page at 0x4000000000000000.
> kdump.sh[652]: readmem: type_addr: 0, addr:0, size:8
> kdump.sh[652]: get_vmemmap_list_info: Can't get vmemmap region addresses
> kdump.sh[652]: get_machdep_info_ppc64: Can't get vmemmap list info.
> kdump.sh[652]: makedumpfile Failed.
> kdump[654]: saving vmcore failed, exitcode:1
> kdump[656]: saving the /run/initramfs/kexec-dmesg.log to /sysroot/var/crash/127.0.0.1-2023-07-31-09:01:28//
> kdump[663]: saving vmcore failed
> systemd[1]: kdump-capture.service: Main process exited, code=exited, status=1/FAILURE
> systemd[1]: kdump-capture.service: Failed with result 'exit-code'.
> systemd[1]: Failed to start Kdump Vmcore Save Service.
> [FAILED] Failed to start Kdump Vmcore Save Service.
> 
> Git bisect points to following patch:
> 
> 8dc9a0ad0c3e7f43e4e091e4e24634e21ce17a54 is the first bad commit
> commit 8dc9a0ad0c3e7f43e4e091e4e24634e21ce17a54
>      powerpc/book3s64/vmemmap: switch radix to use a different vmemmap handling function
> 

Earlier, makedumpfile made use of vmemmap_list to populate all the
vmemmap regions but that changed for radix. It uses the init_mm pgtable
for vmemmap regions. A fix is needed in makedumpfile to catch up with
the above change.

Thanks
Hari