5 ms·
One thing to note is that (last I heard) both Chrome and Firefox sandbox EME modules fairly tightly. Flash is a browser plugin, which means that it usually inje
by geofft 12y ago
One thing to note is that (last I heard) both Chrome and Firefox sandbox EME modules fairly tightly. Flash is a browser plugin, which means that it usually injects code into the browser itself, and runs with full privileges on your computer, just as much as your browser does. This is what makes Flash such fertile ground for exploits of all kinds, and also makes it bad for your privacy because it has direct access to your webcam, microphone, clipboard, supercookies, etc. — anything the browser can do, Flash can do without asking. If it asks, it's out of the kindness of its heart, not because the browser has any say.
Chrome and Firefox's sandboxes, meanwhile, are both open-source. You can inspect what powers the EME module might possibly have, and know that it can't gain any more. A vulnerability in the code is unlikely to be able to do anything other than pirate your download of Game of Thrones — and that's assuming it even has general-purpose network access. Ideally, a vulnerability would be able to do nothing other than modify the video you see, but the remote site could achieve that by encoding a modified video in the first place.
As far as the general moral arguments about DRM go, it's true that the new boss is the same as the old boss. But the bulk of the EFF's argument against Flash in this blog post is about security, not about open content, and it's important to acknowledge that EME is a significant step forward. The new boss is sitting in a tightly locked cage.
- detrino 12y agoFlash for Chrome uses the PPAPI and is sandboxed.
- geofft 12y agoThat is a fair point, but I would also consider that a (limited) victory on par with the (limited) victory of moving from Flash to EME, and I wish it got more press. The people who are implementing sandboxes are doing massively effective work to improve the security and privacy of everyone using the web. Chrome has been quite active here and I don't think they can be thanked enough. That said, PPAPI is Chrome-only; EME is a standard, and Mozilla's implementation even outlines a way for other browsers to integrate the EME module Mozilla uses, plus their sandbox, into their own browsers.
- aikah 12y agothere is no victory here. EME is as bad as flash if not worse.People who refuse to acknowledge that simple fact are as guilty as vendors pushing these technologies. >> EME is a standard, That's the heart of the problem. It guarantees the fact that web technologies are not open technologies anymore. >> Mozilla's implementation Mozilla dropped the ball years ago when it comes to proprietary techs. I see little difference between Firefox and Chrome on that matter.
- dark_photon 12y agoI think the point is that Chrome uses PPAPI for both Flash and EME. Firefox also uses some kind of sandbox for plugins (like Flash) but yes, their planned sandbox for EME is stricter.
- ris 12y agoPeople talk a lot about "sandboxing" things as though it's magical fairy dust.
- jumpwah 12y agoWhen Cameron wanted encryption to be illegal, would that mean EME would also have been illegal?
- geofft 12y agoDidn't he only want encryption without a government backdoor to be illegal? I don't really think anyone has a problem with giving the latest episodes of Homeland and Downton Abbey to our political elite.
- ianlevesque 12y agoIt definitely is in a tightly locked cage. Firefox even generates the device fingerprint itself rather than allow any machine access. An EME module cannot access the Internet, nor store data to the hard drive. There's a detailed writeup here: https://hacks.mozilla.org/2014/05/reconciling-mozillas-mission-and-w3c-eme/ https://hacks.mozilla.org/2014/05/reconciling-mozillas-missi...
- belorn 12y agoAn interesting technical claim to sandbox a process in such a way that it has free access to display drivers and hardware modules (tpm's) but not the host system, network or hard drive. I would like to see a security audit on that claim.
- ianlevesque 12y agoI'm no expert on the Firefox source code but it appears the relevant code is here https://hg.mozilla.org/mozilla-central/file/193c4c5c7ec2/media/gmp-clearkey/0.1 https://hg.mozilla.org/mozilla-central/file/193c4c5c7ec2/med...
- SixSigma 12y agoThis is a bad thing. No saving means : No time shifting, no device shifting, no legal fair use.
- DougBTX 12y agoExaggerating claims beyond a certain point makes them unconvincing. For example, the BBC's iPlayer is designed for time shifting and device shifting, and it has always used restrictive DRM, so saying that DRM prevents that just makes the argument look wrong to semi-informed users.
- falcolas 12y ago> saying that DRM prevents that just makes the argument look wrong to semi-informed users. It does prevent these uses, however, unless the content provider explicitly provides it. The US law says that the content provider does not need to explicitly allow it; it's a right granted to the user by copyright law.
- walterbell 12y agoDRM and sandboxing will also be using hardware enclaves within the x86 processor, https://news.ycombinator.com/item?id=8425178 https://news.ycombinator.com/item?id=8425178
- copsarebastards 12y ago> Chrome and Firefox's sandboxes, meanwhile, are both open-source. You can inspect what powers the EME module might possibly have, and know that it can't gain any more. Yeah, that's not true[1]. Sandbox security is not perfect. Sandboxed plugins are definitely are an improvement in security over non-sandboxed ones, but running untrusted code in a sandbox is still running untrusted code on your machine. At the very least that has to be considered one level of privilege escalation that's just being handed to malicious coders for free. Sandboxes are great. I'd love to see open source code run inside sandboxes, because even if the code has been audited, because even if I know the code isn't trying to do anything bad, it might have vulnerabilities that let other people do bad things. Running closed-source EME code inside a sandbox is still an unnecessary risk, though, because we have no way of verifying whether that code is secure and in fact that code might be an attacker itself. [1] http://www.zdnet.com/article/pwn2own-2012-google-chrome-browser-sandbox-first-to-fall/ http://www.zdnet.com/article/pwn2own-2012-google-chrome-brow...
- vertex-four 12y ago> running untrusted code in a sandbox is still running untrusted code on your machine asm.js and various obfuscators allow websites to do this in the first place with closed-source code, with a far larger attack service than the EME sandbox. In not a lot of time from now, they quite reasonably could.
- comex 12y agoThe JavaScript VM is also a type of sandbox, whose security in every browser is compromised on a depressingly regular schedule. It's true that the EME sandbox is an additional risk, but it's only supposed to run a single piece of software from (in Firefox's case) Adobe, which is probably not malware, and would thus have to be exploited before the strength of the sandbox made a difference. It probably can be exploited, but considering that attackers can run arbitrary evil JavaScript directly... EME is only a small addition to the overall risk profile. What I think is more important, from a security perspective, is that EME defeats (or attempts to defeat) attempts to get by without JavaScript or the complexity of modern browsers altogether. YouTube doesn't work as is without JavaScript anyway, but one determined not to use it could just download all their videos with youtube-dl; until someone comes out with a reverse engineered Widevine or Adobe Primetime or whatever other long-i-festooned DRM systems these sites support, this plan won't work for DRMed video.
- zobzu 12y agoIts more about freedom than security tho. Really if anything, EME is about security for vendors, not users. Really, not much to do with the fact that it's sandboxed. Its not like if proprietary, freedom-restricting software was suddenly ok because there's some kind of a sandbox around it.. as if the only issue with these was that they're poorly coded and supported.
- vertex-four 12y ago> as if the only issue with these was that they're poorly coded and supported A major issue is that arbitrary code can do arbitrary things, up to and including installing rootkits on your machine (for example, the Sony rootkit debacle). The sandbox means it can only do what the sandbox allows it to - in the case of Firefox's sandbox, that's very little. "I should be able to watch whatever content I want on my system without installing proprietary software, even if that software is highly sandboxed and can't even see anything on my system and can't cause any permanent changes" is a political and moral issue more than a technical one.
- AnkhMorporkian 12y agoNo user is being forced into using this. I haven't really followed this argument from the beginning. Philosophically I despise DRM, but pragmatically as long as it isn't forced on users as the only option, then I don't see any issue with an extension having that option being available.
- FeeTinesAMady 12y agoIt will be forced on the users, though. If it becomes widespread, the choices are to support DRM or be locked out of the web.
- nsgi 12y agoNot being able to play protected content (which you couldn't in the first place without installing Flash or another proprietary plugin) is a far cry from being locked out of the web.
- eridius 12y agoI would be shocked if Safari did not also sandbox EME just as tightly as Chrome and Firefox. Of course, Safari isn't open source, but a sufficiently knowledgeable person could locate the background process Safari uses for EME (it probably has a very distinctive name) and check what the sandbox it uses is.
- SixSigma 12y agoWhat does Theo think? > You are absolutely deluded, if not stupid, if you think that a worldwide collection of software engineers who can't write operating systems or applications without security holes, can then turn around and suddenly write virtualization layers without security holes.
- dredmorbius 12y agoThe virtualization layer becomes a single point of concentration for both security and attack. That offers opportunities to reduce the attack surface and opportunity. Net net, given the choice of extensive plug-ins vs. virtualization, I'll go with the latter.
- fnordfnordfnord 12y agoAgree that eating less shit sandwich is better than eating more. But it's still a shit sandwich.