<div dir="ltr"><div>Design template Review is available here<br></div><div><br></div><div><a href="https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21772" target="_blank">https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/21772</a></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, May 16, 2019 at 6:31 PM Andrew Geissler <<a href="mailto:geissonator@gmail.com" target="_blank">geissonator@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Thu, May 16, 2019 at 1:36 AM Deepak Kodihalli<br>
<<a href="mailto:dkodihal@linux.vnet.ibm.com" target="_blank">dkodihal@linux.vnet.ibm.com</a>> wrote:<br>
><br>
> On 15/05/19 6:09 PM, Jayanth Othayoth wrote:<br>
> > ## Problem Description<br>
> > Issue #457: Add support to debug unresponsive host.<br>
> ><br>
> > Scope: High level design direction to solve this problem,<br>
> ><br>
> > ## Background and References<br>
> > There are situation at customer places where OPAL/Linux goes<br>
> > unresponsive causing a system hang. And there is no way to figure out<br>
> > what went wrong with Linux kernel or OPAL. Looking for a way to trigger<br>
> > a dump capture on Linux host so that we can capture the OS dump for post<br>
> > analysis.<br>
> ><br>
> > ## Proposed Design for POWER processor based systems:<br>
> > Get all Host CPUs in reset vector and Linux then has a mechanism to<br>
> > patch it into panic-kdump path to trigger dump capture. This will enable<br>
> > us to analyze and fix customer issue where we see Linux hang and<br>
> > unresponsive system.<br>
> ><br>
> > ### Redfish Schema used:<br>
> > * Reference: DSP2046 2018.3,<br>
> > * ComputerSystem 1.6.0 schema provides an action called<br>
> > #ComputerSystem.Reset”, This action is used to reset the system.<br>
> > ResetType parameter is used for indicating type of reset need to be<br>
> > performed. In this use case we can use “Nmi” type<br>
> > * Nmi: Generate a Diagnostic Interrupt (usually an NMI on x86<br>
> > systems) to cease normal operations, perform diagnostic actions and<br>
> > typically halt the system.<br>
> > * ### d-bus :<br>
> ><br>
> > Option 1: Extending the existing d-bus interface state.Host name<br>
> > space (<br>
> > /openbmc/phosphor-dbus-interfaces/xyz/openbmc_project/State/Host.interface.yaml<br>
> > ) to support new RequestedHostTransition property called “Nmi”. d-bus<br>
> > backend can internally invoke processor specific target to do Sreset(<br>
> > equivalent to x86 NMI) and associated actions.<br>
><br>
> I don't prefer this option, because this would mean adding host specific<br>
> code in phoshor-state-manager, which I think until now is host agnostic.<br>
<br>
Yeah, this was my main concern with tying it into phosphor-state-manager.<br>
The fact Redfish put it in with their other state related commands (which<br>
are implemented by phosphor-state-manager) is the only reason I'm a little<br>
wishy-washy here. We could just create a generic systemd target "host-nmi"<br>
or something and phosphor-state-manager could just call that to abstract<br>
any of the specifics, but it sill doesn't really feel like it fits to me.<br>
<br>
I think I prefer option 2, and then we can just map bmcweb to that API when<br>
the Redfish command comes in. Sounds like for ppc64 systems we can just<br>
use pdbg to issue the NMI.<br>
<br>
> So for that reason, Option 2 sounds better. There are some good<br>
> questions from Neeraj as well, so I would suggest adding this as a<br>
> design template on Gerrit to gather better feedback.<br>
><br>
> Thanks,<br>
> Deepak<br>
><br>
> > Option 2: Introducing new d-bus interface in the control.state namespace<br>
> > (<br>
> > /openbmc/phosphor-dbus-interfaces/xyz/openbmc_project/Control/Host/NMI.interface.yaml)<br>
> > namespace and implement the new d-bus back-end for respective processor<br>
> > specific targets.<br>
><br>
</blockquote></div>