Arnd / Steve,
Thanks for sending out your notes from the recent ELC.
Mark
On Sun, Apr 17, 2016 at 12:01 PM, Steve Muckle <steve.muckle@linaro.org> wrote:
Hello all,
I also attended ELC a couple weeks ago. Here are my highlights.
In general ELC didn't offer a huge amount of presented
content or attendees relevant for my scheduler/power management
interests, but there were still some interesting talks and
opportunities to catch up with folks.
* Dirk Hohndel interviews Linus Torvalds
A couple fun comments - "something weasel" is reserved for "good
kernels," small sensors will probably never run linux, Linus loves
doing userspace because it's so easy and like a vacation - but "you
can't spend your life doing that."
* kernelci.org BOF by Kevin Hilman
Kevin gave a great overview of kernelci.org - how it started, the
current status and next steps. There was plenty of engagement from
the audience along with interest from folks in contributing.
* Google ProjectARA Power Management Challenges by Patrick Titiano
Patrick's talk clearly explains how BayLibre are dealing with
ARA's unique archicture from a power management standpoint.
* SCHED_DEADLINE: A Status Update by Juri Lelli
I especially enjoyed this talk as it was very detailed and also the
most relevant to my work. The theory behind SCHED_DEADLINE was
explained, including why earliest deadline first can be insufficient
and how constant bandwidth server addresses that and why bandwidth
reclaiming is useful. The future work for SCHED_DEADLINE was
outlined.
* Static Code Checking in the Linux Kernel by Arnd Bergmann
This was an interesting look at Arnd's battle to eliminate warnings
and errors. There's an assortment of useful bits of knowledge here
such as things gcc doesn't like and what certain tools are best
suited for.
* Securing Embedded Linux by Michael E. Anderson
A good introductory talk on security, with a little bit of
embedded/IOT spin.
* Understanding a Real-Time System by Steven Rostedt
My first time attending a talk by Steven Rostedt - he reminded me of
my favorite college professors due to his energy and enthusiasm for
the material. Also, the talk was quite useful, describing the
difference between real time and real fast along with all the
sources of variation in execution speed in the system (bad for
real-time/determinism).
In the hallway track I was able to sync up with Mike Turquette from
BayLibre to discuss immediate plans for schedutil (scheduler-guided
CPU frequency). Also took the ARM guys to dinner at Stone Brewing, my
favorite brewery and a rather good restaurant as well :) .
cheers,
Steve
> _______________________________________________
On Sat, Apr 16, 2016 at 08:42:05PM +0200, Arnd Bergmann wrote:
> Hi everyone,
>
> I attended the Embedded Linux Conference earlier this month and have now
> returned home from the extended trip, so it seems like a good time to
> share my experience.
>
> The event was a combination of ELC with the OpenIoTSummit, so it was
> huge with 800 participants and nine parallel tracks. Often enough there
> were multiple interesting topics at the same time, and as usual the
> "hallway track" -- meeting people in person to discuss further topics
> is the most important one.
>
> Slides for most talks have now become available at
> http://events.linuxfoundation.org/events/embedded-linux-conference
> and videos were promised to get posted soonish. Some personal
> highlights include:
>
> * Porting Linux to a new archicture by Joël Porquet
> http://events.linuxfoundation.org/sites/events/files/slides/elc-2016.pdf
> I've talked about this topic before myself at previous conferences,
> and his talk was particularly well structured and easy to follow,
> covering all the important parts. Probably not what most of you need
> for your work though.
>
> * Lessons from Ion by Laura Abbott
> http://events.linuxfoundation.org/sites/events/files/slides/elc16.pdf
> Memory management for GPUs remains an interesting topic, and Laura
> gave an excellent retrospective of what happened with Google's
> Ion approach, what went wrong, and what problems remain to be
> solved. Unfortunately only about 10 people were in the room, as it
> was the very last session and it conflicted with Grant Likely's
> popular talk about "Hardware Design for Linux Engineers".
>
> * Understanding a real-time system by Steven Rostedt
> http://events.linuxfoundation.org/sites/events/files/slides/elc-understanding-rt-system-2016.pdf
> Stephen's talks are always worth attending, even when there is little
> actual new information in them. See the video if you have anything to
> do with realtime systems
>
> * OpenBMC by Tian Fang
> http://events.linuxfoundation.org/sites/events/files/slides/OpenBMC_2016ELC.pdf
> Facebook are doing one of two projects to bring a proper open source
> and secure OS into the baseboard management controller. This is
> incredibly important work, and I'd recommend reading the LWN article
> about the talk at https://lwn.net/Articles/682944/.
> One outcome was that this article has triggered the other OpenBMC team
> (at IBM) to speed up working on their kernel patches to get support for
> the popular AST2xxx SoC family mainlined, we should see patches soon.
> Note that this is the same BMC that is used to manage most ARM servers,
> and it normally runs a closed-down ARM9 or ARM11 Linux distro itself.
>
> I initially planned to attend without giving a talk myself because I could
> not think of a good topic, but Mark Brown had the idea that I could talk
> about my side project of fixing build warnings, and that worked out great
> in the end. There is even an LWN article about it at
> http://lwn.net/Articles/683476/ (for non-subscribers: https://lwn.net/SubscriberLink/683476/180f1aedcd1d0b5b/ ) and my slides are
> at http://events.linuxfoundation.org/sites/events/files/slides/Static%20code%20checking.pdf
> The room was almost full, and the topic was well received (though I still
> think my two presentations last year were much better).
>
> I was talking at the same time as Mark Rutland and Karen Sandler, and I
> probably want to see both of their presentations once they become available
> for streaming.
>
> Arnd
> Conf mailing list
> Conf@lists.linaro.org
> https://lists.linaro.org/mailman/listinfo/conf
--
Mark Orvek
EVP, Engineering
M: +1.408.313.6988 IRC: morvek Skype: morveklinaro.org │ Leading Collaboration in the ARM Ecosystem
_______________________________________________ linaro-kernel mailing list linaro-kernel@lists.linaro.org https://lists.linaro.org/mailman/listinfo/linaro-kernel