TY - GEN
T1 - Stealing webpages rendered on your browser by exploiting GPU vulnerabilities
AU - Lee, Sangho
AU - Kim, Youngsok
AU - Kim, Jangwoo
AU - Kim, Jong
N1 - Publisher Copyright:
© 2014 IEEE.
Copyright:
Copyright 2014 Elsevier B.V., All rights reserved.
PY - 2014/11/13
Y1 - 2014/11/13
N2 - Graphics processing units (GPUs) are important components of modern computing devices for not only graphics rendering, but also efficient parallel computations. However, their security problems are ignored despite their importance and popularity. In this paper, we first perform an in-depth security analysis on GPUs to detect security vulnerabilities. We observe that contemporary, widely-used GPUs, both NVIDIA's and AMD's, do not initialize newly allocated GPU memory pages which may contain sensitive user data. By exploiting such vulnerabilities, we propose attack methods for revealing a victim program's data kept in GPU memory both during its execution and right after its termination. We further show the high applicability of the proposed attacks by applying them to the Chromium and Firefox web browsers which use GPUs for accelerating webpage rendering. We detect that both browsers leave rendered webpage textures in GPU memory, so that we can infer which web pages a victim user has visited by analyzing the remaining textures. The accuracy of our advanced inference attack that uses both pixel sequence matching and RGB histogram matching is up to 95.4%.
AB - Graphics processing units (GPUs) are important components of modern computing devices for not only graphics rendering, but also efficient parallel computations. However, their security problems are ignored despite their importance and popularity. In this paper, we first perform an in-depth security analysis on GPUs to detect security vulnerabilities. We observe that contemporary, widely-used GPUs, both NVIDIA's and AMD's, do not initialize newly allocated GPU memory pages which may contain sensitive user data. By exploiting such vulnerabilities, we propose attack methods for revealing a victim program's data kept in GPU memory both during its execution and right after its termination. We further show the high applicability of the proposed attacks by applying them to the Chromium and Firefox web browsers which use GPUs for accelerating webpage rendering. We detect that both browsers leave rendered webpage textures in GPU memory, so that we can infer which web pages a victim user has visited by analyzing the remaining textures. The accuracy of our advanced inference attack that uses both pixel sequence matching and RGB histogram matching is up to 95.4%.
UR - http://www.scopus.com/inward/record.url?scp=84914100506&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84914100506&partnerID=8YFLogxK
U2 - 10.1109/SP.2014.9
DO - 10.1109/SP.2014.9
M3 - Conference contribution
AN - SCOPUS:84914100506
T3 - Proceedings - IEEE Symposium on Security and Privacy
SP - 19
EP - 33
BT - Proceedings - IEEE Symposium on Security and Privacy
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 35th IEEE Symposium on Security and Privacy, SP 2014
Y2 - 18 May 2014 through 21 May 2014
ER -