Fwd: OpenBMC workshop agenda (was: "Fwd: Foxconn/Rackspace workshop agenda")

[Nathan House]

Forwarding this here.  Looking for a little guidance on io map definitions extracted from an x86 platform.

Thanks!

 Nate House


Begin forwarded message:

Subject: Re: OpenBMC workshop agenda (was: "Fwd: Foxconn/Rackspace workshop agenda")
From: Nathan House <nathan.house at rackspace.com<mailto:nathan.house at rackspace.com>>
Date: December 16, 2015 at 4:00:51 PM CST
Cc: Adi Gangidi <adi.gangidi at rackspace.com<mailto:adi.gangidi at rackspace.com>>, Kenneth Wilke <kenneth.wilke at RACKSPACE.COM<mailto:kenneth.wilke at RACKSPACE.COM>>, Major Hayden <major.hayden at rackspace.com<mailto:major.hayden at rackspace.com>>, Antony Messerli <amesserl at rackspace.com<mailto:amesserl at rackspace.com>>, Norman James/Austin/IBM <njames at us.ibm.com<mailto:njames at us.ibm.com>>, <jk at ozlabs.org<mailto:jk at ozlabs.org>>, <causten at us.ibm.com<mailto:causten at us.ibm.com>>, <pwilliams at us.ibm.com<mailto:pwilliams at us.ibm.com>>
To: Aaron Sullivan <aaron.sullivan at rackspace.com<mailto:aaron.sullivan at rackspace.com>>

Hey Guys,

In my attempts to keep some momentum going following the workshop I was digging on one of our x86 platforms ast2400 vendor provided images.  Looking for what needs to change to make a working image for them etc.  I think with the dev address defs it would get half way if not more.  Loaded their bmc fw bin into binwalk and discovered some interesting binary data that might have what is needed.  Below is a gist of the contents which appear might be refs to addresses for different io and sensors.

https://gist.github.com/naterh/c00caf9c940578eec138

What needs to be added to the skeleton to make this happen.  I’m thinking it would need a new module in bin/ and maybe a bit of work in objects/ and its sub dirs.  Outside of that project and looking at the existing working platforms, I was going to try to base off of palmetto and see where that gets me.  Anything else I’m missing?

Feel free to include or drop off anyone from this thread as some of the addresses might be incorrect.  Was something I’d forgot to get before taking off.

Thanks!

Nate House

On Dec 10, 2015, at 2:52 PM, Aaron Sullivan <aaron.sullivan at rackspace.com<mailto:aaron.sullivan at rackspace.com>> wrote:

See below.  Still might change, but just a bit.

Sent from my mobile
Monday Dec 14th

Time Topic Details Presenter

9:00am 15 min Kickoff Intros, present objective, agenda Norm

9:15 45 min Dbus Dbus tutorial. Summarize bindings (python, sdbus, gdbus) Patrick or Norm?
Even though we aren't using gdbus, they will come across it
when searching and don't want any confusion

10:00 15 min Dbus and IPMI Highlight differences from IPMI, such as inventory vs sensors Norm

10:15 15 min BREAK

10:30 1 hr Dbus example Walk through a current dbus implementation Chris

11:30 1.5 hr LUNCH

1:00pm 30 min Githup overview Review of the gihub repositories and what each contains Brad

1:30 1 hr Tree code struct Review of the openbmc tree code structure in yocto env Brad

2:30 15 min BREAK

2:45 1 hr Dbus objects Overview of all created dbus objects Whoever created the dbus object can talk for it

3:45 1hr Roundtable Questions to Foxconn/Rackspace, such as mfg tools,
how they provision their systems,
how their admins support the systems, access to root level
things on dbus such as flashing, should that be something clients have access to?
Any features they want us to work / focus on
NTP clock, MAC addresses, Rackspace's Ironic

4:45 15 min WRAP-UP Questions

5:00 END



Tuesday Dec 15th

Time Topic Details Presenter

9:00am 15 min Intro Agenda for the day

9:15 1 hr Hwmon Hwmon tutorial and overview of OCC driver Li Yi

10:15 15 min BREAK

10:30 45 min Kernel - Cover new kernel advantages, limitations, device tree overview - Jeremy

11:15 45 min Event logs Event logs and fru isolation strategy Chris

12:00 1 hr LUNCH

1:00pm 1 hr Rest How to access methods through REST Brad

2:00 15 min Rest advantages Show how seamlessly the REST implementation is Brad

2:15 15 min BREAK

2:30 1.5 hr Hands-on Write a dbus implementation (TBD) - System info (free text area) Chris

4:00 30 min Future Where we're going next and where we're heading ???
- Discuss x86 and maybe refactoring Facebook Openbmc
for Yosemite to dbus architecture and new kernel

4:30 30 min WRAP-UP Questions, workshop recap, conclusion

5:00 END


2 topics to add:
- Cover new kernel advantages, limitations, device tree overview - Jeremy
- Discuss x86 and maybe refactoring Facebook Openbmc for Yosemite to dbus architecture and new kernel




<graycol.gif>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openbmc/attachments/20151216/5e4f51cb/attachment.html>