Abstract
Mobile emulation, which creates full-fledged software mobile devices on a physical PC/server, is pivotal to the mobile ecosystem, especially for PC-based mobile gaming, app debugging, and malware detection. Unfortunately, existing mobile emulators perform poorly on graphics-intensive apps in terms of either efficiency or compatibility or both. To address this, we introduce graphics projection, a novel graphics virtualization mechanism that adds a small-size projection space inside the guest memory of a virtual mobile device. The projection space processes graphics operations involving control contexts and resource handles without host interactions. Novel flow control and data teleporting mechanisms are devised to match the decoupled graphics processing rates of the virtual device and the host GPU to maximize performance. The resulting new Android emulator, dubbed Trinity, exhibits an average of 93.3% native hardware performance and 97.2% app support, in some cases outperforming other emulators by more than an order of magnitude. It has been adopted by Huawei DevEco Studio, a major Android IDE with millions of developers.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings of the 16th USENIX Symposium on Operating Systems Design and Implementation, OSDI 2022 |
| Publisher | USENIX Association |
| Pages | 285-301 |
| Number of pages | 17 |
| ISBN (Electronic) | 9781939133281 |
| State | Published - 2022 |
| Event | 16th USENIX Symposium on Operating Systems Design and Implementation, OSDI 2022 - Carlsbad, United States Duration: Jul 11 2022 → Jul 13 2022 |
Publication series
| Name | Proceedings of the 16th USENIX Symposium on Operating Systems Design and Implementation, OSDI 2022 |
|---|
Conference
| Conference | 16th USENIX Symposium on Operating Systems Design and Implementation, OSDI 2022 |
|---|---|
| Country/Territory | United States |
| City | Carlsbad |
| Period | 7/11/22 → 7/13/22 |
Bibliographical note
Funding Information:We would like to express our deepest appreciation to our shepherd, Jason Nieh, who was very responsive during our interactions with him and provided us with valuable suggestions, which have significantly improved our paper. We also thank the anonymous reviewers for their constructive suggestions. We thank Wei Liu and Xinlei Yang for their help in data collection and analysis. This work is supported in part by the National Key R&D Program of China under grant 2021YFB2900100, as well as the National Natural Science Foundation of China (NSFC) under grant 61902211.
Publisher Copyright:
© 2022 by The USENIX Association. All rights reserved.