In this post, I describe a little about me, GSoC and VKMS. I want to share a little about my experience with GSoC and DRM community, an overview of the VKMS project, report project progress, and provide a roadmap for future posts.
When I was in the graduate school, I became aware of Operating Systems (OS) field for the first time, since then, my most wanted professional desire is: work with OS from theoretical to the practical level. Nonetheless, entering in this field is not an easy task when you are not in a favorable environment, but you know… I could not give up on something that I’m passionate about. To solve this problem, I tried to improve my knowledge in OS by starting a master’s degree. On the one hand, my master provides some understanding of the OS state-of-the-art; on the other side, it not offered me practical skills in a real-world OS. I tried to start to contribute to Linux Kernel to fulfill my urge to work in a production-oriented OS, but it was hard to start it alone. However, the conditions begin to change for me at the end of 2017 due to a conference named “Linux Developer Conference Brazil,” (linuxdev-br)[1] in this event I had the fantastic opportunity to talk with many Brazilians kernel developer. I also heard about the experience of people that started to work with Linux Kernel due to GSoC. Finally, I had the opportunity to meet Gustavo Padovan in the conference, I got his contact and interacted with him with questions related to DRM tasks.
Since linuxdev-br I started to put as much effort as I could to become a Kernel developer. During this processes, I often mailed Gustavo to get some help and some ideas on stuff that I could do on DRM subsystem. He recommended me to look at a TODO list on the site, and there I heard for the first time about Virtual KMS (VKMS). In summary, from this idea in the TODO list started my journey with DRM in GSoC.
The Kernel Mode-Setting (KMS) is a mechanism that enables a process to command the kernel to set a mode (screen resolution, color depth, and rate) which is in a range of values supported by graphics cards and the display screen [2]. Create a Virtual KMS (VKMS) bring some benefits, for developers:
In this regard, VKMS have to fake outputs and provide just a basic set of KMS elements such as a single CRTC, encoder, connector, and plane (all up and running). Finally, this module has a potential to evaluate to receive many more features in the future.
Due to the importance of VKMS project, some people are directly engaged in the project. First, there are two interns from a different programs working together which is Haneen (Outreachy) and I (GSoC). Officially Sean Paul mentor Haneen, and Gustavo Padovan mentor me. Additionally, Daniel Vetter provided a lot of support and work as a mentor for both of us. Additionally, many different people from the community helped us with valuable information.
For the first task, Daniel suggested that we create an elementary DRM module which could be used by Haneen and me during the project. The requirements provided by Daniel for this module was:
These tasks were really challenging for me due to my lack of experience and knowledge. As a result, I put a lot effort to try to understand the details in the DRM subsystem; for me, the task was hard but really enjoyable since mentors and the community helped me with my questions. Additionally, work with Haneen is fantastic since she always helps me to understand some of my mistakes and also provided me some great feedbacks. After three weeks of interactions, we create the basic structure of VKMS which can be checked in the “drm-misc” on branch “topic/vkms” [3].
Now the work was split into two parts: add a fake page flip simulation ( real and virtual hardware), and CRC. I am working on the first task and Haneen on the second one. With these two features, VKMS will become valuable and ready for expansion regarding features. I believe these two tasks will take much more time than the first task, and it will be much more exciting.
We wish to make VKMS pass on various of the IGT tests. As a result, we had to learn how to use the IGT and how it works. For example, we had to change the target module in the __open_driver function to use VKMS. From my studies in the IGT source code, I noticed a lot of things that I could work for helping to improve the project. I recently decided to try to upstream some of my contributions, some of them are:
I have other small things on my local branches to try to upstream, but I decided to wait for a little to get feedback in the previous patches.
Since the beginning of the GSoC, I put a lot of effort to dive into the DRM. I faced some challenges related to basic stuff, such as understand the basic concepts and jargon, to technical problems associated with starting a new module from scratch. Inspired by all the obstacle that Haneen and I faced, I believe that could be worth for newcomer a series of posts based on VKMS. From my current experience, I want to write the following posts: