+- Kodi Community Forum (https://forum.kodi.tv)
+-- Forum: Discussions (https://forum.kodi.tv/forumdisplay.php?fid=222)
+--- Forum: Hardware (https://forum.kodi.tv/forumdisplay.php?fid=112)
+--- Thread: Hacking the Boxee Box to run XBMC? (/showthread.php?tid=64578)
XBMC on Boxee - lotagu - 2010-12-16
Is anyone working on getting XBMC working on a Boxee? I have 2 Boxee Boxes and 2 Xbox with XBMC running on them...Would love to retire my Xbox and run XBMC on the Boxe Boxes..
Please someone do this
??
- outleradam - 2010-12-16
Any idea on who to contact to persuade intel to write drivers for Linux?
- prae5 - 2010-12-16
outleradam Wrote:Any idea on who to contact to persuade intel to write drivers for Linux?
They are already available, I've seen xbmc running on the ce4100 platform, however afaik they are proprietary drivers and must be licensed, so no hardware acceleration without them.
- poofyhairguy - 2010-12-16
outleradam Wrote:Any idea on who to contact to persuade intel to write drivers for Linux?
The problem is that with that chipset, unlike like previous Intel chipsets, Intel doesn't own the design- it is just licensed from PowerVR (the makers of the Dreamcast GPU).
Until the day Intel realizes that its entire GPU department has been a huge failure and buys PowerVR to take over that task, Intel should be avoided at all cost.
- HwyXingFrog - 2010-12-16
poofyhairguy Wrote:The problem is that with that chipset, unlike like previous Intel chipsets, Intel doesn't own the design- it is just licensed from PowerVR (the makers of the Dreamcast GPU).
Until the day Intel realizes that its entire GPU department has been a huge failure and buys PowerVR to take over that task, Intel should be avoided at all cost.
So, who is giving Boxee the drivers to do GPU Acceleration for 1080p?
The Nvidia drivers that XBMC Live uses aren't open source either, aren't they proprietary. Either way they do work awesome.
- poofyhairguy - 2010-12-16
HwyXingFrog Wrote:So, who is giving Boxee the drivers to do GPU Acceleration for 1080p?
The Nvidia drivers that XBMC Live uses aren't open source either, aren't they proprietary. Either way they do work awesome.
I don't really care that Intel's drivers aren't open sourced (as I LOVE Nvidia), I care that they are closed source AND unavailable. Boxee is able to get Linux drivers from Intel, but they are not available to the general public.
The whole thing for years with Intel (since about GMA 915) was that their drivers sucked, but at least they were open. Sometimes they barely worked, but they came with your favorite Linux distro and they gave you basically functionality without pain.
You are right Nvidia has closed drivers- but they work hard to keep them relevant. Nvidia is constantly worked to keep up with new changes in Xorg with frequent releases. Intel on the other hand dropped a VAAPI driver for one of its licensed PowerVR design once that only works with one version of Ubuntu tweaked the right way. Intel is closed and mostly incomparable, which translates to useless.
- HwyXingFrog - 2010-12-17
poofyhairguy Wrote:I don't really care that Intel's drivers aren't open sourced (as I LOVE Nvidia), I care that they are closed source AND unavailable. Boxee is able to get Linux drivers from Intel, but they are not available to the general public.
The whole thing for years with Intel (since about GMA 915) was that their drivers sucked, but at least they were open. Sometimes they barely worked, but they came with your favorite Linux distro and they gave you basically functionality without pain.
You are right Nvidia has closed drivers- but they work hard to keep them relevant. Nvidia is constantly worked to keep up with new changes in Xorg with frequent releases. Intel on the other hand dropped a VAAPI driver for one of its licensed PowerVR design once that only works with one version of Ubuntu tweaked the right way. Intel is closed and mostly incomparable, which translates to useless.
After all this BS and waiting for the Boxee Box (Which was originally going to be Tegra 2, but I understand the reason for dropping it do to lack of 1080p), I think I'm no longer waiting.
I guess I'll be buying another ION based system for XBMC, since I can't wait any longer to get the Boxee Box to work. And the more I find out, the more I start to think it's not going to be a short wait.
- Philmatic - 2010-12-17
I would be happy with UDF 2.5/2.6 support and bitstream code being backported to XBMC, how about it?
- topfs2 - 2010-12-17
Philmatic Wrote:I would be happy with UDF 2.5/2.6 support and bitstream code being backported to XBMC, how about it?
The code is available, how about you fetch us some patches..
- outleradam - 2010-12-17
Hrm... Will that code take care of the issue? Is Boxee actually running hardware verified signed code?
If that be the case, maybe Team Xecuter could help once again?
- topfs2 - 2010-12-17
outleradam Wrote:Hrm... Will that code take care of the issue? Is Boxee actually running hardware verified signed code?
If that be the case, maybe Team Xecuter could help once again?
If they can hack SSL public key.... which takes years with brute force.
- outleradam - 2010-12-17
Yeah I know about that. I'm studying for security+.
We would have to bypass the current firmware completely. Basically the proprietariness would have to be completely bypassed. The way team xecuter did their magic on XBOX was to Bypass the XBOX bios. This allowed XBOX to run homebrew software like XBMC. Which is why we are here in the first place.
Assuming the keys are encrypted in the bios there is no reason we cannot bypass that with some expert help. If we leave the Boxee filesystem in-tact and install some custom firmware then Boxee should still be able to run and execute proprietary Boxee code while leaving the door opened for unencrypted x86 code.
The main goal should be to find a way to allow Boxee to execute unsigned code from a USB boot drive right? I think we need to start looking at hardware and firmware so we can avoid the "SSL"
When you say SSL are you actually talking about Secure Socket Layer or are you referring to some sort of encryption? If so then it should be possible to obtain the key when it is used by looking at the hardware. Are you sure it's not like AES or bitlocker or something?
If anyone has a boxes I'd like to play with it a bit. I'd like to telnet in and see what I can see.
- topfs2 - 2010-12-17
outleradam Wrote:Yeah I know about that. I'm studying for security+.
We would have to bypass the current firmware completely. Basically the proprietariness would have to be completely bypassed. The way team xecuter did their magic on XBOX was to Bypass the XBOX bios. This allowed XBOX to run homebrew software like XBMC. Which is why we are here in the first place.
Assuming the keys are encrypted in the bios there is no reason we cannot bypass that with some expert help. If we leave the Boxee filesystem in-tact and install some custom firmware then Boxee should still be able to run and execute proprietary Boxee code while leaving the door opened for unencrypted x86 code.
The main goal should be to find a way to allow Boxee to execute unsigned code from a USB boot drive right? I think we need to start looking at hardware and firmware so we can avoid the "SSL"
When you say SSL are you actually talking about Secure Socket Layer or are you referring to some sort of encryption? If so then it should be possible to obtain the key when it is used by looking at the hardware. Are you sure it's not like AES or bitlocker or something?
If anyone has a boxes I'd like to play with it a bit. I'd like to telnet in and see what I can see.
Sorry, I didn't mean SSL. its RSA or something like that, it do look like open key stuff though. I'll admit its not my main area of expertise, I know a bit of security stuff but not enough to remember stuff off the top of my head
It does't really use BIOS in the normal sense of the word, and it is perhaps possible to work around but for that you need the SDK, which is not open :S And likely you need either root or serial access to it, the latter not really end user friendly.
Reading out the key from the hardware would be interesting, not sure how easy it is though. If we have that key it would be a big step forward.
- teaguecl - 2010-12-17
topfs2 Wrote:Reading out the key from the hardware would be interesting, not sure how easy it is though. If we have that key it would be a big step forward.
If there is a key in the firmware of the BB, it won't be of much use - it will be the public key. That doesn't help, you need the private key. If they've left the private key somewhere in the firmware, then somebody should be fired
- Hannes The Hun - 2010-12-17
the CE4100 SoC incorporates a dedicated security processor with on-die fuses. if they actually use this baby, good luck trying to run anything on that thing other than the original firmware. as I said before, the situation is very similar to the fucking signed bootloaders motorola uses on the milestone (europe's droid). if they want, they can even blow the whole box (figuratively speaking...) when somebody messes with the system.
d-link, boxee and of course intel needed to make sure that the box could be locked as tightly as possible due to contracts with the media overlords and their love for everything DRM. heck, in theory the boxee box could even get a bluray license. this could very well be one of the major points why they switched from tegra2 to intel in the last minute (and not only the decoding limitations). my best guess would be that an android-based tegra2 box would have been far more hacker-friendly, but I'm no expert here.