TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Android Encryption Demystified

40 点作者 sashk大约 8 年前

6 条评论

strcat大约 8 年前
There are some inaccuracies and misleading information here:<p>- Encryption keys are derived from various inputs including the user credentials, not stored in the TEE. The TEE is involved in key derivation and is really supposed to use a hardware-bound key not directly accessible to software including itself but it&#x27;s an implementation detail that varies by device.<p>- Pixel phones ship with file-based encryption. It&#x27;s not an option. A phone either uses FDE or FBE. If it uses FBE, then it supports Direct Boot (partial functionally before the user credentials for encryption are enabled via device-encrypted storage class) and per-profile encryption keys. It enables a bunch of possible improvements like authenticated encryption down the road, but isn&#x27;t much of a security improvement itself. Nexus 5X and 6P only offered a partial implementation as preview for developers tucked away in the hidden developer options, not a user-facing option.<p>- Credential-based encryption is enabled by default when setting a lockscreen method.<p>- Android has a Keystore, that&#x27;s not exclusive to iOS.<p>- Android doesn&#x27;t use ECB even though it&#x27;s implied that only iOS uses unique keys per block. Android does too.<p>There&#x27;s some more, but I don&#x27;t have time to go through and nitpick. It&#x27;s clearly written based on interpreting other people&#x27;s blog posts, etc. rather than direct knowledge of how it works or even reading the documentation. It doesn&#x27;t even sound like they have experience using an Android device based on some statements they make.<p>The post totally misses out on things like key derivation and which data is kept at rest. The article misses the real remaining iOS encryption advantages (FBE data classes that Android hasn&#x27;t added yet and more of the key derivation work is hardware-bound) and just tries to make it sound good by stating things that are not unique to iOS.
ickwabe大约 8 年前
I think this session from I&#x2F;O 2017 provides a better overview in some regards. Admittedly it&#x27;s brief on details but the overview is helpful.<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=C9_ytg6MUP0" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=C9_ytg6MUP0</a>
sofaofthedamned大约 8 年前
I can&#x27;t quite understand why Elcomsoft are still going.<p>1. They&#x27;re being stiffed by Google and Apple upping their security game, hence the amount of complaints on this blog in the comments<p>2. Their users all seem dodgy as fuck, it&#x27;s not even funny.<p>3. Passwords and 2fa are making their brute force methods completely outdated<p>Why haven&#x27;t they shut down yet?
评论 #14405042 未加载
评论 #14406075 未加载
评论 #14407012 未加载
gambiting大约 8 年前
My problem is that I had full encryption enabled on Marshmallow, upgraded to nougat and my system settings say that encryption is enabled even though I strongly suspect it isn&#x27;t - my phone used to ask me for pin before booting up, but now it doesn&#x27;t, so I suspect the encryption was removed while upgrading but the system settings are showing the old value for whatever reason. And I can&#x27;t really wipe my entire phone just to re-enable the encryption, it&#x27;s a lot of work to set up all apps again. Xperia Z5.<p>My point is that android encryption has problems if you can just lose the encryption between os version upgrades.
评论 #14405107 未加载
评论 #14405183 未加载
评论 #14406114 未加载
mnm1大约 8 年前
How does it work with a pattern lock? Is the pattern used to calculate the encryption key?
评论 #14408674 未加载
评论 #14404809 未加载
astrobase_go大约 8 年前
this is one the only articles i&#x27;ve ever read where &quot;honking&quot; is used in a serious capacity as an adjective.