Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179390
Jonas Juffinger, Lukas Lamster, Andreas Kogler, Maria Eichlseder, Moritz Lipp, D. Gruss
In this paper, we present CSI:Rowhammer, a principled hardware-software co-design Rowhammer mitigation with cryptographic security and integrity guarantees, that does not focus on any specific properties of Rowhammer. We design a new memory error detection mechanism based on a low-latency cryptographic MAC and an exception mechanism initiating a software-level correction routine. The exception handler uses a novel instruction-set extension for the error correction and resumes execution afterward. In contrast to regular ECC-DRAM that remains exploitable if more than 2 bits are flipped, CSI:Rowhammer maintains the security level of the cryptographic MAC. We evaluate CSI:Rowhammer in a gem5 proof-of-concept implementation. Under normal conditions, we see latency overheads below 0.75% and no memory overhead compared to off-the-shelf ECC-DRAM. While the average latency to correct a single bitflip is below 20 ns (compared to a range from a few nanoseconds to several milliseconds for state-of-the-art ECC memory), CSI:Rowhammer can detect any number of bitflips with overwhelming probability and correct at least 8 bitflips in practical time constraints.
{"title":"CSI:Rowhammer – Cryptographic Security and Integrity against Rowhammer","authors":"Jonas Juffinger, Lukas Lamster, Andreas Kogler, Maria Eichlseder, Moritz Lipp, D. Gruss","doi":"10.1109/SP46215.2023.10179390","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179390","url":null,"abstract":"In this paper, we present CSI:Rowhammer, a principled hardware-software co-design Rowhammer mitigation with cryptographic security and integrity guarantees, that does not focus on any specific properties of Rowhammer. We design a new memory error detection mechanism based on a low-latency cryptographic MAC and an exception mechanism initiating a software-level correction routine. The exception handler uses a novel instruction-set extension for the error correction and resumes execution afterward. In contrast to regular ECC-DRAM that remains exploitable if more than 2 bits are flipped, CSI:Rowhammer maintains the security level of the cryptographic MAC. We evaluate CSI:Rowhammer in a gem5 proof-of-concept implementation. Under normal conditions, we see latency overheads below 0.75% and no memory overhead compared to off-the-shelf ECC-DRAM. While the average latency to correct a single bitflip is below 20 ns (compared to a range from a few nanoseconds to several milliseconds for state-of-the-art ECC memory), CSI:Rowhammer can detect any number of bitflips with overwhelming probability and correct at least 8 bitflips in practical time constraints.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130487080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179483
Lucien K. L. Ng, Sherman S. M. Chow
We studied 53 privacy-preserving neural-network papers in 2016-2022 based on cryptography (without trusted processors or differential privacy), 16 of which only use homomorphic encryption, 19 use secure computation for inference, and 18 use non-colluding servers (among which 12 support training), solving a wide variety of research problems. We dissect their cryptographic techniques and "love-hate relationships" with machine learning alongside a genealogy highlighting noteworthy developments. We also re-evaluate the state of the art under WAN. We hope this can serve as a go-to guide connecting different experts in related fields.
{"title":"SoK: Cryptographic Neural-Network Computation","authors":"Lucien K. L. Ng, Sherman S. M. Chow","doi":"10.1109/SP46215.2023.10179483","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179483","url":null,"abstract":"We studied 53 privacy-preserving neural-network papers in 2016-2022 based on cryptography (without trusted processors or differential privacy), 16 of which only use homomorphic encryption, 19 use secure computation for inference, and 18 use non-colluding servers (among which 12 support training), solving a wide variety of research problems. We dissect their cryptographic techniques and \"love-hate relationships\" with machine learning alongside a genealogy highlighting noteworthy developments. We also re-evaluate the state of the art under WAN. We hope this can serve as a go-to guide connecting different experts in related fields.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"222 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114310646","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179328
Xinyi Wang, Cen Zhang, Yeting Li, Zhiwu Xu, Shuailin Huang, Yi Liu, Yican Yao, Yang Xiao, Yanyan Zou, Y. Liu, Wei Huo
Regular expression Denial-of-Service (ReDoS) is one kind of algorithmic complexity attack. For a vulnerable regex, attackers can craft certain strings to trigger the super-linear worst-case matching time, which causes denial-of-service to regex engines. Various ReDoS detection approaches have been proposed recently. Among them, hybrid approaches which absorb the advantages of both static and dynamic approaches have shown their performance superiority. However, two key challenges still hinder the effectiveness of the detection: 1) Existing modelings summarize localized vulnerability patterns based on partial features of the vulnerable regex; 2) Existing attack string generation strategies are ineffective since they neglected the fact that non-vulnerable parts of the regex may unexpectedly invalidate the attack string (we name this kind of invalidation as disturbance.)Rengar is our hybrid ReDoS detector with new vulnerability modeling and disturbance free attack string generator. It has the following key features: 1) Benefited by summarizing patterns from full features of the vulnerable regex, its modeling is a more precise interpretation of the root cause of ReDoS vulnerability. The modeling is more descriptive and precise than the union of existing modelings while keeping conciseness; 2) For each vulnerable regex, its generator automatically checks all potential disturbances and composes generation constraints to avoid possible disturbances.Compared with nine state-of-the-art tools, Rengar detects not only all vulnerable regexes they found but also 3 – 197 times more vulnerable regexes. Besides, it saves 57.41% – 99.83% average detection time compared with tools containing a dynamic validation process. Using Rengar, we have identified 69 zero-day vulnerabilities (21 CVEs) affecting popular projects which have more than dozens of millions weekly download count.
{"title":"Effective ReDoS Detection by Principled Vulnerability Modeling and Exploit Generation","authors":"Xinyi Wang, Cen Zhang, Yeting Li, Zhiwu Xu, Shuailin Huang, Yi Liu, Yican Yao, Yang Xiao, Yanyan Zou, Y. Liu, Wei Huo","doi":"10.1109/SP46215.2023.10179328","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179328","url":null,"abstract":"Regular expression Denial-of-Service (ReDoS) is one kind of algorithmic complexity attack. For a vulnerable regex, attackers can craft certain strings to trigger the super-linear worst-case matching time, which causes denial-of-service to regex engines. Various ReDoS detection approaches have been proposed recently. Among them, hybrid approaches which absorb the advantages of both static and dynamic approaches have shown their performance superiority. However, two key challenges still hinder the effectiveness of the detection: 1) Existing modelings summarize localized vulnerability patterns based on partial features of the vulnerable regex; 2) Existing attack string generation strategies are ineffective since they neglected the fact that non-vulnerable parts of the regex may unexpectedly invalidate the attack string (we name this kind of invalidation as disturbance.)Rengar is our hybrid ReDoS detector with new vulnerability modeling and disturbance free attack string generator. It has the following key features: 1) Benefited by summarizing patterns from full features of the vulnerable regex, its modeling is a more precise interpretation of the root cause of ReDoS vulnerability. The modeling is more descriptive and precise than the union of existing modelings while keeping conciseness; 2) For each vulnerable regex, its generator automatically checks all potential disturbances and composes generation constraints to avoid possible disturbances.Compared with nine state-of-the-art tools, Rengar detects not only all vulnerable regexes they found but also 3 – 197 times more vulnerable regexes. Besides, it saves 57.41% – 99.83% average detection time compared with tools containing a dynamic validation process. Using Rengar, we have identified 69 zero-day vulnerabilities (21 CVEs) affecting popular projects which have more than dozens of millions weekly download count.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114695903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179401
Haoyang Li, Qingqing Ye, Haibo Hu, Jin Li, Leixia Wang, Chengfang Fang, Jie Shi
Federated Learning (FL), the de-facto distributed machine learning paradigm that locally trains datasets at individual devices, is vulnerable to backdoor model poisoning attacks. By compromising or impersonating those devices, an attacker can upload crafted malicious model updates to manipulate the global model with backdoor behavior upon attacker-specified triggers. However, existing backdoor attacks require more information on the victim FL system beyond a practical black-box setting. Furthermore, they are often specialized to optimize for a single objective, which becomes ineffective as modern FL systems tend to adopt in-depth defense that detects backdoor models from different perspectives. Motivated by these concerns, in this paper, we propose 3DFed, an adaptive, extensible, and multi-layered framework to launch covert FL backdoor attacks in a black-box setting. 3DFed sports three evasion modules that camouflage backdoor models: backdoor training with constrained loss, noise mask, and decoy model. By implanting indicators into a backdoor model, 3DFed can obtain the attack feedback in the previous epoch from the global model and dynamically adjust the hyper-parameters of these backdoor evasion modules. Through extensive experimental results, we show that when all its components work together, 3DFed can evade the detection of all state-of-the-art FL backdoor defenses, including Deepsight, Foolsgold, FLAME, FL-Detector, and RFLBAT. New evasion modules can also be incorporated in 3DFed in the future as it is an extensible framework.
{"title":"3DFed: Adaptive and Extensible Framework for Covert Backdoor Attack in Federated Learning","authors":"Haoyang Li, Qingqing Ye, Haibo Hu, Jin Li, Leixia Wang, Chengfang Fang, Jie Shi","doi":"10.1109/SP46215.2023.10179401","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179401","url":null,"abstract":"Federated Learning (FL), the de-facto distributed machine learning paradigm that locally trains datasets at individual devices, is vulnerable to backdoor model poisoning attacks. By compromising or impersonating those devices, an attacker can upload crafted malicious model updates to manipulate the global model with backdoor behavior upon attacker-specified triggers. However, existing backdoor attacks require more information on the victim FL system beyond a practical black-box setting. Furthermore, they are often specialized to optimize for a single objective, which becomes ineffective as modern FL systems tend to adopt in-depth defense that detects backdoor models from different perspectives. Motivated by these concerns, in this paper, we propose 3DFed, an adaptive, extensible, and multi-layered framework to launch covert FL backdoor attacks in a black-box setting. 3DFed sports three evasion modules that camouflage backdoor models: backdoor training with constrained loss, noise mask, and decoy model. By implanting indicators into a backdoor model, 3DFed can obtain the attack feedback in the previous epoch from the global model and dynamically adjust the hyper-parameters of these backdoor evasion modules. Through extensive experimental results, we show that when all its components work together, 3DFed can evade the detection of all state-of-the-art FL backdoor defenses, including Deepsight, Foolsgold, FLAME, FL-Detector, and RFLBAT. New evasion modules can also be incorporated in 3DFed in the future as it is an extensible framework.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124118422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179461
Marzieh Bitaab, Haehyun Cho, Adam Oest, Zhuo Lyu, Wei Wang, Jorij Abraham, Ruoyu Wang, Tiffany Bao, Yan Shoshitaishvili, Adam Doupé
Despite recent advancements in malicious website detection and phishing mitigation, the security ecosystem has paid little attention to Fraudulent e-Commerce Websites (FCWs), such as fraudulent shopping websites, fake charities, and cryptocurrency scam websites. Even worse, there are no active large-scale mitigation systems or publicly available datasets for FCWs.In this paper, we first propose an efficient and automated approach to gather FCWs through crowdsourcing. We identify eight different types of non-phishing FCWs and derive key defining characteristics. Then, we find that anti-phishing mitigation systems, such as Google Safe Browsing, have a detection rate of just 0.46% on our dataset. We create a classifier, BEYOND PHISH, to identify FCWs using manually defined features based on our analysis. Validating BEYOND PHISH on never-before-seen (untrained and untested data) through a user study indicates that our system has a high detection rate and a low false positive rate of 98.34% and 1.34%, respectively. Lastly, we collaborated with a major Internet security company, Palo Alto Networks, as well as a major financial services provider, to evaluate our classifier on manually labeled real-world data. The model achieves a false positive rate of 2.46% and a 94.88% detection rate, showing potential for real-world defense against FCWs.
{"title":"Beyond Phish: Toward Detecting Fraudulent e-Commerce Websites at Scale","authors":"Marzieh Bitaab, Haehyun Cho, Adam Oest, Zhuo Lyu, Wei Wang, Jorij Abraham, Ruoyu Wang, Tiffany Bao, Yan Shoshitaishvili, Adam Doupé","doi":"10.1109/SP46215.2023.10179461","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179461","url":null,"abstract":"Despite recent advancements in malicious website detection and phishing mitigation, the security ecosystem has paid little attention to Fraudulent e-Commerce Websites (FCWs), such as fraudulent shopping websites, fake charities, and cryptocurrency scam websites. Even worse, there are no active large-scale mitigation systems or publicly available datasets for FCWs.In this paper, we first propose an efficient and automated approach to gather FCWs through crowdsourcing. We identify eight different types of non-phishing FCWs and derive key defining characteristics. Then, we find that anti-phishing mitigation systems, such as Google Safe Browsing, have a detection rate of just 0.46% on our dataset. We create a classifier, BEYOND PHISH, to identify FCWs using manually defined features based on our analysis. Validating BEYOND PHISH on never-before-seen (untrained and untested data) through a user study indicates that our system has a high detection rate and a low false positive rate of 98.34% and 1.34%, respectively. Lastly, we collaborated with a major Internet security company, Palo Alto Networks, as well as a major financial services provider, to evaluate our classifier on manually labeled real-world data. The model achieves a false positive rate of 2.46% and a 94.88% detection rate, showing potential for real-world defense against FCWs.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122224880","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Continuous Integration (CI) is a widely-adopted software development practice for automated code integration. A typical CI workflow involves multiple independent stakeholders, including code hosting platforms (CHPs), CI platforms (CPs), and third party services. While CI can significantly improve development efficiency, unfortunately, it also exposes new attack surfaces. As the code executed by a CI task may come from a less-trusted user, improperly configured CI with weak isolation mechanisms might enable attackers to inject malicious code into victim software by triggering a CI task. Also, one insecure stakeholder can potentially affect the whole process. In this paper, we systematically study potential security threats in CI workflows with multiple stakeholders and major CP components considered. We design and develop an analysis tool, CInspector, to investigate potential vulnerabilities in seven popular CPs, when integrated with three mainstream CHPs. We find that all CPs have the risk of token leakage caused by improper resource sharing and isolation, and many of them utilize over-privileged tokens with improper validity periods. We further reveal four novel attack vectors that allow attackers to escalate their privileges and stealthy inject malicious code by executing a piece of code in a CI task. To understand the potential impact, we conduct a large-scale measurement on the three mainstream CHPs, scrutinizing over 1.69 million repositories. Our quantitative analysis demonstrates that some very popular repositories and large organizations are affected by these attacks. We have duly reported the identified vulnerabilities to CPs and received positive responses.
{"title":"Continuous Intrusion: Characterizing the Security of Continuous Integration Services","authors":"Yacong Gu, Lingyun Ying, Huajun Chai, Chu Qiao, Haixin Duan, Xing Gao","doi":"10.1109/SP46215.2023.10179471","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179471","url":null,"abstract":"Continuous Integration (CI) is a widely-adopted software development practice for automated code integration. A typical CI workflow involves multiple independent stakeholders, including code hosting platforms (CHPs), CI platforms (CPs), and third party services. While CI can significantly improve development efficiency, unfortunately, it also exposes new attack surfaces. As the code executed by a CI task may come from a less-trusted user, improperly configured CI with weak isolation mechanisms might enable attackers to inject malicious code into victim software by triggering a CI task. Also, one insecure stakeholder can potentially affect the whole process. In this paper, we systematically study potential security threats in CI workflows with multiple stakeholders and major CP components considered. We design and develop an analysis tool, CInspector, to investigate potential vulnerabilities in seven popular CPs, when integrated with three mainstream CHPs. We find that all CPs have the risk of token leakage caused by improper resource sharing and isolation, and many of them utilize over-privileged tokens with improper validity periods. We further reveal four novel attack vectors that allow attackers to escalate their privileges and stealthy inject malicious code by executing a piece of code in a CI task. To understand the potential impact, we conduct a large-scale measurement on the three mainstream CHPs, scrutinizing over 1.69 million repositories. Our quantitative analysis demonstrates that some very popular repositories and large organizations are affected by these attacks. We have duly reported the identified vulnerabilities to CPs and received positive responses.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134352641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
FMCW radars are integral to automotive driving for robust and weather-resistant sensing of surrounding objects. However, these radars are vulnerable to spoofing attacks that can cause sensor malfunction and potentially lead to accidents. Previous attempts at spoofing FMCW radars using an attacker device have not been very effective due to the need for synchronization between the attacker and the victim. We present a novel spoofing mechanism called mmSpoof that does not require synchronization and is resilient to various security features and countermeasures of the victim radar. Our spoofing mechanism uses a "reflect array" based attacker device that reflects the radar signal with appropriate modulation to spoof the victim’s radar. We provide insights and mechanisms to flexibly spoof any distance and velocity on the victim’s radar using a unique frequency shift at the mmSpoof’s reflect array. We design a novel algorithm to estimate this frequency shift without assuming prior information about the victim’s radar. We show the effectiveness of our spoofing using a compact and mobile setup with commercial-off-the-shelf components in realistic automotive driving scenarios with commercial radars.
{"title":"mmSpoof: Resilient Spoofing of Automotive Millimeter-wave Radars using Reflect Array","authors":"Rohith Reddy Vennam, I. Jain, Kshitiz Bansal, Joshua Orozco, Puja Shukla, Aanjhan Ranganathan, Dinesh Bharadia","doi":"10.1109/SP46215.2023.10179371","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179371","url":null,"abstract":"FMCW radars are integral to automotive driving for robust and weather-resistant sensing of surrounding objects. However, these radars are vulnerable to spoofing attacks that can cause sensor malfunction and potentially lead to accidents. Previous attempts at spoofing FMCW radars using an attacker device have not been very effective due to the need for synchronization between the attacker and the victim. We present a novel spoofing mechanism called mmSpoof that does not require synchronization and is resilient to various security features and countermeasures of the victim radar. Our spoofing mechanism uses a \"reflect array\" based attacker device that reflects the radar signal with appropriate modulation to spoof the victim’s radar. We provide insights and mechanisms to flexibly spoof any distance and velocity on the victim’s radar using a unique frequency shift at the mmSpoof’s reflect array. We design a novel algorithm to estimate this frequency shift without assuming prior information about the victim’s radar. We show the effectiveness of our spoofing using a compact and mobile setup with commercial-off-the-shelf components in realistic automotive driving scenarios with commercial radars.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134369776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179399
Lukas Gerlach, Daniel Weber, Ruiyi Zhang, Michael Schwarz
Microarchitectural attacks threaten the security of computer systems even in the absence of software vulnerabilities. Such attacks are well explored on x86 and ARM CPUs, with a wide range of proposed but not-yet deployed hardware countermeasures. With the standardization of the RISC-V instruction set architecture and the announcement of support for the architecture by major processor vendors, RISC-V CPUs are on the verge of becoming ubiquitous. However, the microarchitectural attack surface of the first commercially-available RISC-V hardware CPUs still needs to be explored.This paper analyzes the two commercially-available off-the-shelf 64-bit RISC-V (hardware) CPUs used in most RISC-V systems running a full-fledged commodity Linux system. We evaluate the microarchitectural attack surface and introduce 3 new microarchitectural attack techniques: Cache+Time, a novel cache-line-granular cache attack without shared memory, Flush+Fault exploiting the Harvard cache architecture for Flush+Reload, and CycleDrift exploiting unprivileged access to instruction-retirement information. We also show that many known attacks apply to these RISC-V CPUs, mainly due to non-existing hardware countermeasures and instruction-set subtleties that do not consider the microarchitectural attack surface. We demonstrate our attacks in 6 case studies, including the first RISC-V-specific microarchitectural KASLR break and a CycleDrift-based method for detecting kernel activity. Based on our analysis, we stress the need to consider the microarchitectural attack surface during every step of a CPU design, including custom ISA extensions.
{"title":"A Security RISC: Microarchitectural Attacks on Hardware RISC-V CPUs","authors":"Lukas Gerlach, Daniel Weber, Ruiyi Zhang, Michael Schwarz","doi":"10.1109/SP46215.2023.10179399","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179399","url":null,"abstract":"Microarchitectural attacks threaten the security of computer systems even in the absence of software vulnerabilities. Such attacks are well explored on x86 and ARM CPUs, with a wide range of proposed but not-yet deployed hardware countermeasures. With the standardization of the RISC-V instruction set architecture and the announcement of support for the architecture by major processor vendors, RISC-V CPUs are on the verge of becoming ubiquitous. However, the microarchitectural attack surface of the first commercially-available RISC-V hardware CPUs still needs to be explored.This paper analyzes the two commercially-available off-the-shelf 64-bit RISC-V (hardware) CPUs used in most RISC-V systems running a full-fledged commodity Linux system. We evaluate the microarchitectural attack surface and introduce 3 new microarchitectural attack techniques: Cache+Time, a novel cache-line-granular cache attack without shared memory, Flush+Fault exploiting the Harvard cache architecture for Flush+Reload, and CycleDrift exploiting unprivileged access to instruction-retirement information. We also show that many known attacks apply to these RISC-V CPUs, mainly due to non-existing hardware countermeasures and instruction-set subtleties that do not consider the microarchitectural attack surface. We demonstrate our attacks in 6 case studies, including the first RISC-V-specific microarchitectural KASLR break and a CycleDrift-based method for detecting kernel activity. Based on our analysis, we stress the need to consider the microarchitectural attack surface during every step of a CPU design, including custom ISA extensions.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131766413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179359
Mathew Hogan, Yan Michalevsky, Saba Eskandarian
We introduce new compression side-channel attacks against database storage engines that simultaneously support compression of database pages and encryption at rest. Given only limited, indirect access to an encrypted and compressed database table, our attacks extract arbitrary plaintext with high accuracy. We demonstrate accurate and performant attacks on the InnoDB storage engine variants found in MariaDB and MySQL as well as the WiredTiger storage engine for MongoDB.Our attacks overcome obstacles unique to the database setting that render previous techniques developed to attack TLS ineffective. Unlike the web setting, where the exact length of a compressed and encrypted message can be observed, we make use of only approximate ciphertext size information gleaned from file sizes on disk. We amplify this noisy signal and combine it with new attack heuristics tailored to the database setting to extract secret plaintext. Our attacks can detect whether a random string appears in a table with > 90% accuracy and extract 10-character random strings from encrypted tables with > 95% success.
{"title":"DBREACH: Stealing from Databases Using Compression Side Channels","authors":"Mathew Hogan, Yan Michalevsky, Saba Eskandarian","doi":"10.1109/SP46215.2023.10179359","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179359","url":null,"abstract":"We introduce new compression side-channel attacks against database storage engines that simultaneously support compression of database pages and encryption at rest. Given only limited, indirect access to an encrypted and compressed database table, our attacks extract arbitrary plaintext with high accuracy. We demonstrate accurate and performant attacks on the InnoDB storage engine variants found in MariaDB and MySQL as well as the WiredTiger storage engine for MongoDB.Our attacks overcome obstacles unique to the database setting that render previous techniques developed to attack TLS ineffective. Unlike the web setting, where the exact length of a compressed and encrypted message can be observed, we make use of only approximate ciphertext size information gleaned from file sizes on disk. We amplify this noisy signal and combine it with new attack heuristics tailored to the database setting to extract secret plaintext. Our attacks can detect whether a random string appears in a table with > 90% accuracy and extract 10-character random strings from encrypted tables with > 95% success.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123166402","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-01DOI: 10.1109/SP46215.2023.10179326
Yingchen Wang, Riccardo Paccagnella, Alan Wandke, Zhao Gang, Grant Garrett-Grossman, Christopher W. Fletcher, David Kohlbrenner, H. Shacham
The recent Hertzbleed disclosure demonstrates how remote-timing analysis can reveal secret information previously only accessible to local-power analysis. At worst, this constitutes a fundamental break in the constant-time programming principles and the many deployed programs that rely on them. But all hope is not lost. Hertzbleed relies on a coarse-grained, noisy channel that is difficult to exploit. Indeed, the Hertzbleed paper required a bespoke cryptanalysis to attack a specific cryptosystem (SIKE). Thus, it remains unclear if Hertzbleed represents a threat to the broader security ecosystem.In this paper, we demonstrate that Hertzbleed’s effects are wide ranging, not only affecting cryptosystems beyond SIKE, but also programs beyond cryptography, and even computations occurring outside the CPU cores. First, we demonstrate how latent gadgets in other cryptosystem implementations— specifically "constant-time" ECDSA and Classic McEliece— can be combined with existing cryptanalysis to bootstrap Hertzbleed attacks on those cryptosystems. Second, we demonstrate how power consumption on the integrated GPU influences frequency on the CPU—and how this can be used to perform the first cross-origin pixel stealing attacks leveraging "constant-time" SVG filters on Google Chrome.
{"title":"DVFS Frequently Leaks Secrets: Hertzbleed Attacks Beyond SIKE, Cryptography, and CPU-Only Data","authors":"Yingchen Wang, Riccardo Paccagnella, Alan Wandke, Zhao Gang, Grant Garrett-Grossman, Christopher W. Fletcher, David Kohlbrenner, H. Shacham","doi":"10.1109/SP46215.2023.10179326","DOIUrl":"https://doi.org/10.1109/SP46215.2023.10179326","url":null,"abstract":"The recent Hertzbleed disclosure demonstrates how remote-timing analysis can reveal secret information previously only accessible to local-power analysis. At worst, this constitutes a fundamental break in the constant-time programming principles and the many deployed programs that rely on them. But all hope is not lost. Hertzbleed relies on a coarse-grained, noisy channel that is difficult to exploit. Indeed, the Hertzbleed paper required a bespoke cryptanalysis to attack a specific cryptosystem (SIKE). Thus, it remains unclear if Hertzbleed represents a threat to the broader security ecosystem.In this paper, we demonstrate that Hertzbleed’s effects are wide ranging, not only affecting cryptosystems beyond SIKE, but also programs beyond cryptography, and even computations occurring outside the CPU cores. First, we demonstrate how latent gadgets in other cryptosystem implementations— specifically \"constant-time\" ECDSA and Classic McEliece— can be combined with existing cryptanalysis to bootstrap Hertzbleed attacks on those cryptosystems. Second, we demonstrate how power consumption on the integrated GPU influences frequency on the CPU—and how this can be used to perform the first cross-origin pixel stealing attacks leveraging \"constant-time\" SVG filters on Google Chrome.","PeriodicalId":439989,"journal":{"name":"2023 IEEE Symposium on Security and Privacy (SP)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134509502","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：