WD TV Live Very Limited Public Beta Version 1.03.42

RoofingGuy wrote:

Well, you saw how well WD’s “please don’t distribute this…” has worked with the other firmwares… they were posted (and re-hosted) everywhere within minutes of WD emailing the links out.

 

People appear to just be so desperate to have the mkvmerge issue “fixed”, that they’ll do anything… well, anything aside from avoiding piracy or re-muxing the affected files themselves, it appears.

 

Which does open a large security hole.  As you say, anyone with malicious intent can start posting new “firmwares,” and people will load them and distribute them without even giving it a moment’s thought.

 

There’s still the infamous instance of a legitimate WD update bricking boxes… I’d wait for a confirmed stable release before I just start clicking and installing the first thing I see.  But I don’t think checksums (whether they end up as valid or not) will stop those who are desperate to have the latest beta – they’re not going to run the check anyways.  They’ll just find the firmware somewhere and install it, and hope for the best.

Pretty much in agreement on all of the above.  Couple of things:

Fixing the mkvmerge issues is undoubtedly part of the problem as relates to backchannel spread of beta firmware.  However, my gut feeling is that a large part of it also comes down to WD’s lack of regular updates for the devices: people are basically anxious enough to get anything newer than firmware from last March that they’re taking a greater interest in the beta releases.  The thing is, WD’s release and distribution methods aren’t conducive to receiving them in a controlled manner, which ultimately lead to the creation of the firmware ‘black market’ we’re seeing today.  In a sense, it’s a vicious circle.

While some of this may be down to idle curiosity or early adopters doing what early adopters do, there is functionality that has been added in the betas over the last (almost) seven months which has never been packaged into a stable release.  Using my position as an example, support for DVD menus is something that I definitely require in order to be able to view much of my media - without it, I can only see the first feature on the disc, which is a huge issue on discs that contain multiple features.  But the only way to get that functionality at this time is to use beta firmware.

Related to checksums and distribution of malicious firmware: I completely agree that it will not completely stop the problem of someone releasing one that has been altered to be malicious; all the necessary ingredients are in place for one to be concocted and distributed.  However, it will give an avenue by which integrity can be verified - so at the very least, a baseline as to what is an official release and what isn’t can be established.  As things stand now, we can’t even reliably do that with firmware downloaded directly from WD: there’s no way to verify that, say, the 1.02.21 firmware downloaded today is the same as when it was posted on the 30th of March this year.

So yes, while I agree that the introduction of checksums into the distribution process is not the overall solution to the problems we’re seeing here, it will at least introduce a level of sanity to the process that is presently lacking.  It’s definitely just one piece of a much larger puzzle, but a crucial one.

Your comment about the 1.01.12 update bricking units has me doing some tangental thinking, however.  I’m almost tempted to start tearing into the firmware again and looking at the traffic generated during an update session to see if there’s a way to introduce a bogus (read: malicious) firmware to the unit…  But I’m not overly-keen on doing my day job in an unpaid capacity during my free time.  Maybe I’ll poke around in it a little, but the motivation to do so is quite honestly low.

Ok from now on…can we request that WD make available MD5 and/or SHA1 and/or CRC32 checksums of the zipped firmwares?

I have been using this freeware hash calculator for years now. Its small and requires no installation.

http://keir.net/hash.html

Yeah, because we all know how often firmware coming from the legitimate site gets hacked…  ;)  (Not just WD, I don’t think ANY firmware from any vendor I’ve done business with does that…)

Even Cisco only does it for their enterprise products…

oh cmon, firmwares dont get released everyday :wink:

it only takes what…5 secs to generate those hashes…

and 5 secs to copy/paste it somewhere so that those that want to can refer to it…

…and yea just because other instituitions dont do it doesnt mean WD should not lead in terms of best practices…imo

TonyPh12345 wrote:

Yeah, because we all know how often firmware coming from the legitimate site gets hacked…  ;)  (Not just WD, I don’t think ANY firmware from any vendor I’ve done business with does that…)

 

Even Cisco only does it for their enterprise products…

Okay, I’ll clarify why this is necessary.  Again.  As succinctly as possible.

There is no way at this time to verify that the contents of those packages is what WD intended to distribute and have not been altered by a third party.  Given that this applies to every firmware release - beta or stable - they have ever had for these devices, the slipshod manner in which these releases have been handled, the ease with which the firmware can potentially be maliciously altered, and the fact that WD has pretty much lost control of the situation at this point, some baseline level of sanity needs to be brought back into the process.  This would be a step in that direction for them.

You may not see this as being a problem.  However, that doesn’t mean that it isn’t one.

What Cisco - or any other vendor - does or does not do is not relevant here.

I tried the 1.03.42 (Russian forum) download link earlier in this thread - basically I like to live on the brim of the lip of the cusp of the rim of the precipice. I installed it and found no functional differences for the limited amount of time I’ve used it. When a new firmware comes out it will simply be overwritten.

Nobody died after this greatest of all risk taking firmware update ventures (apart from a couple of Georgian’s). I will also be publishing my memoirs of this particular event in time for the Christmas prezzy rush, for those interested about my adventures in the fast lane.

I may eat some sand tomorrow, but you’ll have to buy the book to find out what happens.

casm23 wrote:

 

Okay, I’ll clarify why this is necessary.  Again.  As succinctly as possible.

 

There is no way at this time to verify that the contents of those packages is what WD intended to distribute and have not been altered by a third party.  

 

I have no problem with anyone who wants to spend the time and effort posting MD5SUM for product.   It just won’t address the concern you have.

If someone can modify the online firmware, they can also very likely modify the online MD5SUM.

Meaning, if someone can actually modify the code that is deployed via official channels, they can also modify the official MD5SUM.  Thus, you’re posited problem isn’t solved.

    • *> TonyPh12345 wrote:> Meaning, if someone can actually modify the code that is deployed via official channels, they can also modify the official MD5SUM.  Thus, you’re posited problem isn’t solved.> * * *

Well, yes - this is correct.  It’s also not the scenario I was addressing with any of this.  What I am referring to is firmware that is downloaded from official channels, then subsequently modified and distributed outside of those same official channels.

Example: at some point, someone downloaded the 1.03.42 package from WD and placed it on a third-party hosting service, the URL to which was passed around on various forums.  Everyone who is downloading that package is assuming that it has not been modified since it was obtained from WD.  With WD providing no md5 values or similar, there is no way for an end user to verify that the firmware they received from that third-party source is unaltered from what WD originally provided.

In the scenario you’re putting forward, an attacker has completely demolished the chain of trust between WD and the end user by placing a malicious package at its primary point of distribution.  Of course any md5 values generated from a malicious update provided in this manner would not prevent it from appearing valid on comparison - checksums are simply a method of establishing integrity of the software, not intent.  It comes back to the age-old problem of determining if your source is trustworthy or not: if firmware is being distributed from a compromised host maintained by an otherwise trusted entity (as in the scenario you have outlined), then by definition anything obtained from that source is now untrustworthy.  However, this a problem that is well outside the scope of what is being sought to be accomplished with making checksums available for verification purposes.

Look at md5 values as being equivalent to needle exchanges for junkies: they won’t eliminate the problem entirely, but at least they curtail some of the more unpleasant aspects of it.

Are md5 checksums necessary for testing file integrity when downloading from WDC servers? Won’t testing the archive be sufficient to determine a successful / safe download? Not sure WD has any obligation or motive to provide md5 checksums just to help reassure those who jump the gun and download leaked firmwares from suspicious sources.

casm23 wrote:

 

Look at md5 values as being equivalent to needle exchanges for junkies: they won’t eliminate the problem entirely, but at least they curtail some of the more unpleasant aspects of it.

 

Wow.   An interesting parallel!   :smileyvery-happy:

I agree with your conclusions based on YOUR premise.

However, I don’t agree with your premise.  :)     There’s no reason to get it from third parties when WD’s distribution is ample.   The only reason people are pulling it off third party sites is because WD REMOVES it from their own.   

… and they do that for a reason.   

To track your parallel:  The Food and Drug Administration (for those not in the USA:  That’s the governmental agency responsible for the approval for the manufacture, distribution and sale of almost ALL medical products in the USA) orders a product recall ANY product containing PPA, as they did a few years back.  Who in their right mind would go trolling the streets for “grey market” drugs containing PPA?   :slight_smile:

mabuk wrote:

I tried the 1.03.42 (Russian forum) download link earlier in this thread - basically I like to live on the brim of the lip of the cusp of the rim of the precipice. I installed it and found no functional differences for the limited amount of time I’ve used it. When a new firmware comes out it will simply be overwritten.

 

Nobody died after this greatest of all risk taking firmware update ventures (apart from a couple of Georgian’s). I will also be publishing my memoirs of this particular event in time for the Christmas prezzy rush, for those interested about my adventures in the fast lane.

 

I may eat some sand tomorrow, but you’ll have to buy the book to find out what happens.

Well lets hope you always stay lucky.

Out of interest it appears that eating sand can be healthy, its probably better than downloading dubious firmware from the net.

http://news.softpedia.com/news/Eat-Sand-to-Stay-Fit-and-Preserve-Your-Health-37135.shtml

I guess the beta was limited to zero.

TonyPh12345 wrote:

 


casm23 wrote:

 

Look at md5 values as being equivalent to needle exchanges for junkies: they won’t eliminate the problem entirely, but at least they curtail some of the more unpleasant aspects of it.

 


Wow.   An interesting parallel!   :smileyvery-happy:

 

I agree with your conclusions based on YOUR premise.

 

However, I don’t agree with your premise.  :)     There’s no reason to get it from third parties when WD’s distribution is ample.   The only reason people are pulling it off third party sites is because WD REMOVES it from their own.   

 

… and they do that for a reason.   

 

To track your parallel:  The Food and Drug Administration (for those not in the USA:  That’s the governmental agency responsible for the approval for the manufacture, distribution and sale of almost ALL medical products in the USA) orders a product recall ANY product containing PPA, as they did a few years back.  Who in their right mind would go trolling the streets for “grey market” drugs containing PPA?   :slight_smile:

 

 

Glad you enjoyed the metaphor :slight_smile:

I’m actually in agreement with you regarding the availability of the firmware from WD - but only to a point.  They can definitely get the firmware out there when they want to, without a doubt.  But their methodology in distributing it is flawed, and that is where my concern arises.

Using your FDA example for a moment, I completely agree that this is a case of ‘who in their right mind would get it from the grey market’.  But WD is creating a situation whereby that grey market exists (which I believe I alluded to in one of my earlier posts), and people are getting it from there.  The problem then becomes one of quality control: if people are obtaining the firmware from sources other than WD directly, they don’t know what they’re getting.  Using a drug-based allusion again, there’s basically no way to know what it’s been cut with - it might just get you high, but it might also stop your heart on the spot.

Just to clarify: I have no issue with the removal of firmware if there’s a compelling reason to do so, be it if a showstopper bug is found after release, or it’s just superseded in the natural course of development; either of those would both be completely understandable reasons.

Really, this is something I would be far less concerned about if WD appeared to be taking an approach to development and releases that wasn’t as haphazard as it evidently is now.  The manner in which they handled the 1.03.42 release was extremely poor all around, even in comparison with their less-than-stellar recent track record in that department.  Couple this with the fact that everything necessary to start doing Nasty Stuff with their firmware is already out there and it sets alarm bells ringing - after all, devices like these have traditionally been considered low-risk, but given that they run OSes capable of acting as staging points for LAN- or WAN-based attacks of various flavours, it would be good to have some ability to at least categorise updates as safe or not.

Anyway, it seems like we’re 90% in agreement on this - we’re just haggling over the merit of the other 10%.  My opinion is that that remaining 10% is largely down to WD’s overall approach and can be mitigated by some effort on their behalf to improve this.  Whether they actually do that or not remains to be seen, but let me say that I’m not holding my breath on the likelihood of them actually taking those steps :wink:

. It’s alive……   :smileyvery-happy:

1.03.42 is now live. Sorry about the confusion earlier.

And with that this thread is closed. :stuck_out_tongue: