By definition, ‘unplanned’ session for people to come and discuss anything around AOSP - including AOSP for members, AOSP for non-Android Linux devs, OPTEE/Security, AOSP TV, to name a few. An initial list of topics for discussion could include:
- An Intro: Uses of AOSP codebase at Linaro and members
- AOSP for Linux developers (an intro to the AOSP codebase for regular Linux guys?
- AOSP TV
- Perhaps Completely open graphics stack
This session will detail and respond to questions regarding the latest developments in the MMC/SD stack and how we are moving forward with block multiqueueing (blk-mq) and the Budget Fair Queueing (BFQ) patch set
LMG team, in collaboration with Toolchains team, has worked on multiple areas related to clang, gcc and native AOSP builds. Bero and Renato will take us through what has been done, and what we need to continue to do. This presentation will focus on LMG efforts around the following:
- Clang CI efforts
- Getting AOSP to build with gcc 6/7
- Building kernels with clang
- Native AOSP builds and development
The SKS/KeyGen2 project is about establishing an security architecture, provisioning and management scheme for cryptographic keys targeting a wide variety of applications including on-line banking, payments, e-government access, and enterprise login. A TEE (possibly aided by a local security processor) is a core component of the envisioned architecture. In order to enable easy enrollment, a browser-based provisioning protocol is another core component. Since a cryptographic key (unlike a file), usually represents a relationship to a remote party which also typically imply a policy for "their" keys, the system supports key ACLs which through an OS/TEE layer governs which applications a key may be used with. A consequence of this arrangement is that cryptographic keys become first-class OS objects like files. The protocol and basic key store is already running as an application which is used for testing and evaluation. What's missing is the OS/TEE/Browser integration, something which requires a set of rather different
This session will talk about efforts around increasing Android testing - frequency, automation and coverage. The discussion will include details about:
- Current testing Linaro does
- Changes we’re working on for testing AOSP/master
- More kernel focused testing & integrating community tests into android
- “Sharding” CTS test runs in the cloud for faster results
The European Space Agency’s (ESA) current mission Gaia and upcoming one Euclid will generate massive amounts of astronomical data to be made freely available on-line through powerful data management systems. This big data era brings a change of paradigm and associated big challenges for building these missions’ archives. The current use case of “bring the data to the user” is being replaced by the new concept of “bring the user to the data”. This new generation of science archives is transforming into science exploitation and collaborative platforms, where the users can send her/his code to the data, run it on computing and storage services provided directly by the archive (where the data reside), and easily share all this with other scientists they’re collaborating with.
We have recently achieved a critical milestone of having core enablement for SBSA (Server Base System Architecture) and SBBR (Server Base Boot Requirements) in upstream kernels "out of the box". These initial standards address minimal functionality and will soon undergo additional development as we seek to cover more complex use case requirements, such as NFV, and those introduced by the emergence of workload acceleration (CCIX, FPGAs, etc.). In this session, we will gather feedback and input on where to go next.