Accessing kernel panic information without BMC console?

[Neeraj Ladkani]

Any data on how heavy is RamOops ( w.r.t memory ) ? 

mem=128M ramoops.mem_address=0x8000000 ramoops.ecc=1

Neeraj

-----Original Message-----
From: openbmc <openbmc-bounces+neladk=microsoft.com at lists.ozlabs.org> On Behalf Of Kun Yi
Sent: Monday, September 30, 2019 11:08 AM
To: Yong Li <yong.b.li at linux.intel.com>
Cc: Andrew Jeffery <andrew at aj.id.au>; OpenBMC Maillist <openbmc at lists.ozlabs.org>; Vijay Khemka <vijaykhemka at fb.com>
Subject: Re: Accessing kernel panic information without BMC console?

On Sun, Sep 29, 2019 at 12:47 AM Yong Li <yong.b.li at linux.intel.com> wrote:
>
> I am working on this ramoops feature too, and it works on ast2500, just enable some kernel config and modify the dts.

Thanks for sharing, Yong Li. I have yet to look into this option.
Would you be able to share how much RAM you are using and how you solve the persistent RAM problem?

>
> I will send out a patch about this change later.
>
> Thanks,
> Yong
> -----Original Message-----
> From: openbmc 
> <openbmc-bounces+yong.b.li=linux.intel.com at lists.ozlabs.org> On Behalf 
> Of Vijay Khemka
> Sent: Saturday, September 28, 2019 6:28 AM
> To: Kun Yi <kunyi at google.com>; Joel Stanley <joel at jms.id.au>; Andrew 
> Jeffery <andrew at aj.id.au>; OpenBMC Maillist <openbmc at lists.ozlabs.org>
> Subject: Re: Accessing kernel panic information without BMC console?
>
>
>
> On 9/27/19, 3:05 PM, "openbmc on behalf of Kun Yi" <openbmc-bounces+vijaykhemka=fb.com at lists.ozlabs.org on behalf of kunyi at google.com> wrote:
>
>     Hello there,
>
>     Wonder whether anyone has had experience persisting kernel panic
>     information or sending them through network? For a lot of our devices
>     the console is either unconnected or served by obmc-console-client
>     only, which wouldn't be able to capture kernel oops.
>
>     We are starting to look into these tools:
>
>     https://www.kernel.org/doc/html/v5.3/admin-guide/ramoops.html
>     Ramoops says it requires persistent RAM.. Which may make it infeasible
>     since we don't persist memory (or, we try not to).
>
>     https://urldefense.proofpoint.com/v2/url?u=https-3A__linux.die.net_man_8_netdump&d=DwIBaQ&c=5VD0RTtNlTh3ycd41b3MUw&r=v9MU0Ki9pWnTXCWwjHPVgpnCR80vXkkcrIaqU7USl5g&m=0H6Yf0igviHCOSHAbOPQxxM-_B0Lh1EyZm4dpdMqe2g&s=H3n57FRZwV0z_wOHqmaiRa-kQ6h3doWg712SV4ez-GU&e=
>     netdump seems promising.
>
>     https://www.kernel.org/doc/html/v5.3/admin-guide/kdump/kdump.html
>     It seems we need more pieces to get kdump working. Have anyone tried
>     kexec/kdump on their platforms?
>
> I have used kdump in past was very happy with it. Yes, it needs lot more pieces to make it work.
>
>
>
>     Regards,
>     Kun
>
>
>


--
Regards,
Kun