[ydl-gen] PS3 is Hacked!

[Robert Spykerman]

On Wed, Jan 27, 2010 at 3:45 AM, Derick Centeno <dcenteno at ydl.net> wrote:


> On Tue, 26 Jan 2010 22:20:56 +1100
> Robert Spykerman <robert.spykerman at gmail.com> wrote:
>>...
>> Now, I am wondering
>> (1) when we will be able to actually run unsigned code on the ps3
>>...
> In brief as to when the code will be available depends on when he
> releases his work and under what open source license.  If you want to have a
> clue regarding the variety of open source licenses (and gain a better
> understanding of that universe) you can refer here: http://www.opensource.org/

Ah, he's a hacker, quite honestly I doubt that he'd care about
licensing or anything like that.  In the spirit of things, it looks
like he has just realized his hack to the public domain, which I've
just downloaded to take a lookie.

And really, the bottom line here is that he's not really publishing
'code' per se he's describing a 'method'.

So... licensing... Non issue.

Take a look at his blog it's quite interesting.

http://geohotps3.blogspot.com/2010/01/heres-your-silver-platter.html

It seems he boots a linux kernel per normal and then inserts a kernel
module which does some fancy memory footwork that I *STILL* don't
understand and somehow trips something inside. He claims enables him
to inject and run unsigned code.

Given he's actually publicly released the hack, and expects fellow
capable colleagues to replicate it, probabilistically I expect it
would likely work. It sounds crazy enough to.

So I would say, (1) is probably in the bag. Not with any certainty,
but ... all it takes is for a few other people out there to replicated
it. And ... I would wager that quite a few will try.

(That said, I'm not taking a soldering iron to my PS3 till, heck, at
least I have ONE to spare hahaha)

>> (2) if (1) is realized, who will rewrite the kernel (and I also would
>> say drivers of course) to run on an un-hypervisor-abstracted ps3...
>
> Depending under which open source license the work is finally released as a
> variety of projects can participate.  I doubt very much whether Fixstars will
> participate though as they have commercial obligations to adhere to.  Then
> again what Fixstars does or doesn't choose to do has little to do with what
> talented programmers will choose to do on their own anyway, if I was to guess I
> would believe that work would appear within Debian Linux and it's variant
> Ubuntu first.  Slackware Linux and others may move towards this direction and
> others.
>
> Established commercial interests may not offer tools to facilitate this
> independent effort as their business obligations would be threatened or prevent
> them from participating.  IBM for instance, announced that it ceased Cell
> development last December.  So work at a sufficiently sophisticated and useful
> level which interests techncially saavy professionals may be slow to develop if
> at all, since most will be looking towards where IBM is going as opposed to
> Cell projects which IBM has abandoned.

Re: the kernel rewrite IF (1) proves practical and easy (which it
still looks not unfortunately)..

I just hope there is enough interest in the PS3 hacker community to
engineer something and boot linux.

Honestly, I couldn't care who provides us with a better kernel :) Just
that they make it easy enough for the average enthusiast to... er...
deploy. We can worry about stuff like "what distro" later.

Boot first.

Worry later :)

This would be the most CRITICAL point to achieve, I think after this
hack is confirmed (if it is). If this can be achieved... Wow....

>> (3) how easy/practical (1) and in fact (2) actually is...and if it
>> will ever actually happen.
>
> Programming for the Cell isn't easy at all.  Decent discussions involving an in
> depth consideration of various processors was explored with the YDL Board.  One
> such discussion is here:
> http://www.yellowdog-board.com/viewtopic.php?f=19&t=4976

I don't think the Cell is truly the issue here although it is a true
point the cell is hard to program in the sense of truly making use of
its might. All that SPE synergy is not really needed to boot a linux.
All that is required is that the SPE's  not interfere (yes, I mean
that locked-up one especially).

What I perceive the difficulty to be is getting arbitrary unsigned
code running _easily_ and _practically_. At present the way he
describes it appears too inconvenient. But it is a  start.

And then there is the issue of a kernel for effectively what I suspect
is an entirely 'new' machine (I hope I am wrong!), for which ...
manuals are not likely to be available :) So, this is not going to be
easy, I suspect.

And then there is the 'What can Sony do about it?' factor

Interesting stuff. I watch eagerly with anticipation.

Robert Spykerman

-- 
chown -R us ./base