Whether that's been accidentally left in or not is not clear. It is hard to see how it could be useful on a phone.
But the screenshot mentions "/SourceCache/zfs/zfs-138.0.1/src/cmd/zfs/zfs_main.c" - how does that version number compare with the various ZFS code drops from Apple?
Chris
> According to
> <http://www.imore.com/2012/02/03/zfs-references-latest-ios-51-beta/> there
> is ZFS code in the latest iOS 5.1 beta.
Good eye, Chris. I would never have imagined such a thing, and I'll update the FAQ ;) I originally added an entry about whether ZFS existed on iOS, as a half joke just because people literally have asked that more than once. And to me, that constitutes "frequently-asked" ;)
This part perplexes me a bit:
"Even without official Apple support, we’ve seen OS X ZFS ports crop up since then. What sets this iOS implementation apart from some of the very original, official, OS X ports of ZFS is that it already contains the ZFS utilities (zpool, zdump, etc). By having the utilities already installed it means that the port is in a more complete stage of development and has full read + write access (unlike the original OS X ports)."
So for some reason, he's referring to the public release of Apple's original port circa Mac OS 10.5. But he seems to be saying that the fact that they ported some binaries inherently means that it's fully functional. If that's what he's saying, then I'd have to disagree with that assumption unless it's proven. I don't want to encourage anyone to violate their NDA by talking about it on a public mailing list, but I guess a person who did have that beta wouldn't be able to know anything unless they jailbroke it and examined the 'zpool' binary with 'strings' or something like that. I don't know.
>
>
>> According to
>> <http://www.imore.com/2012/02/03/zfs-references-latest-ios-51-beta/> there
>> is ZFS code in the latest iOS 5.1 beta.
>
>
> Good eye, Chris. I would never have imagined such a thing, and I'll update the FAQ ;) I originally added an entry about whether ZFS existed on iOS, as a half joke just because people literally have asked that more than once. And to me, that constitutes "frequently-asked" ;)
As Adam Leventhal (ex-Sun kernel/dtrace/ZFS developer, now at Delphix) noted, iPhone 4S hardware is pretty much as powerful as desktop hardware was when ZFS was released 5 or so years ago.
That might be stretching it a little, but he's probably not so far off.
>
> This part perplexes me a bit:
>
> "Even without official Apple support, we’ve seen OS X ZFS ports crop up since then. What sets this iOS implementation apart from some of the very original, official, OS X ports of ZFS is that it already contains the ZFS utilities (zpool, zdump, etc). By having the utilities already installed it means that the port is in a more complete stage of development and has full read + write access (unlike the original OS X ports)."
Apple's original ZFS bits for OS X came with zfs and zpool commands, yet were read-only IIRC. They just send commands down to the kernel, so the kernel bits are what makes it read/write.
Chris
Would be quite interested in a more substantial article delineating contextual facts that would allow readers to make their own conclusions based on context.
Thanks much!
-George
> Hi Jack, the information on the screen just appears to be just a bunch of
> information cobbled together vs anything real contextual. Was this acquired
> from a jailbroken release? Does the source have something more substantial in
> terms of actual screenshots from a shell on the device or some more substantial
> strings from the image?
Yeah, Jack, if ZFS is in iOS, then as a recipient of the binary, you're legally entitled to request its source code. Do it! I'd request it but I don't have the iOS developer subscription; only for Mac OS.
I sure don't see it here! http://opensource.apple.com/release/ios-50/
I doubt there's any ZFS components in IOS. For one, ZFS is very heavily processor oriented and using that as a filesystem would result in a higher computation (and thus power drain). There's also relatively little benefit for the phone itself; for the user partition (where the application data resides) the data can mutate quite quickly. If you've got quickly mutating data on a ZFS system, and you snapshot, you quickly run out of space.
(It's one of the reasons why e.g. /var and /tmp tends to have its own filesystem when you're doing it on boot-from-zfs devices.)
Secondly, I don't see the phone ever needing to snapshot. The only reason to snapshot is to either backup off-line (which doesn't happen) or to restore (which also doesn't happen). And with iCloud being a major part of Apple's strategy as a means to backup to entirely new devices (which, I should remind readers of this list, is done on a per-app basis and not at the filesystem level) it doesn't make sense to view ZFS as a restoration point.
About the only benefit I can see is so that the firmware/os (root partition) can enable a failback in the case of upgrades going wrong. But to do that you'd need (up to) double the OS space (one to hold prior copy) to install the new image. And Apple (and TeVo) never managed the boot-from-zfs problem, which again flags this up as suspicious.
Finally, upgrades fail regularly on iDevices. (I hit a failure 1/3 or similar, but I have a fair few and I upgrade a lot.) All that happens is it goes into drive restore mode, from where you can lay down the image again.
So all that to solve an occasional problem at device upgrade time? I really doubt it. Even the upgrade OTA probably works by just laying down a filesystem delta, which is basically what Apple's system updates have done from day one.
But then, you can all link back to this post if it turns out to be true in a few years time :)
Alex
I don't primarily care whether it's a sound commercial product strategy. My primary concern is whether or not ZFS technology has been delivered by Apple to anyone whatsoever outside of Apple, and whether they will request and receive the source code for it. Then, we'll ponder the commercial ramifications for dear old Apple! ^_^
> Hi there Alex! While I can nod agreeably through everything you just said, which may or may not be relevant to a cutting edge megacorporation with virtually unlimited financial resources and forward-reaching technological scope, I am still puzzled by the apparent evidence put forth by the gentleman.
My understanding of the evidence us that is a screenshot purporting to be of a next gen OS in the lab.
I have no direct knowledge of the accuracy of that evidence, and it is not like Apple are going to confirm the existence of it. Nor is it possible to independently obtain the evidence suggested.
On balance of probabilities (both in the likelihood of ZFS existing in situ and of screenshots of same making their way past Apple legal/security) I have to assume that the evidence presented is faked.
That's not to say I have proof of faking nor that I have evidence to the contrary. It is entirely possible that it is true. However it is much more probable that it is false.
Alex
> My understanding of the evidence us that is a screenshot purporting to be of a
> next gen OS in the lab.
Ohhhhh, okay. So I guess if it's not the iOS 5.1 beta, then it wouldn't have been externally released and thus the source code is not required to be distributed. I guess we'll see! ;)
Well then I'll leave you with the following movie. This is me, except I'm all like, "Daaaaaang, beautiful. Can I have your source code? Can I have your source code? Can I have it? Can I? Can I have it? Will you please relinquish and upload unto my storage medium, the sequence of ones and zeroes, which, when sent through a compilational toolchain, will lead me to the execution of your beautiful executables? No? Oh, you just all up into the ponderization of proprietariful computer sciences and secret algorithmical theoriness in the laboratory and WHAT NOT. OkAY. okAY. Okay, Dennis Ritchie. mmHMMMM. Daaaaaang."
http://www.youtube.com/watch?v=kTFZyl7hfBw
(no encouragementizing of the infringementalness of copyright is intendified by the distribunation of the aforementioned video frame visualizations)
> According to <http://www.imore.com/2012/02/03/zfs-references-latest-ios-51-beta/> there is ZFS code in the latest iOS 5.1 beta.
Now that iOS 5.1 is released, do we know if there's actually ZFS code in there?
Chris
Jason
Sent from my iPad