Nano 2G

[mctubster]

Looks like the new Nano 2G is not based on the PortalPlayer chips

http://www.eet.com/news/latest/showArticle.jhtml?articleID=193000601

from the article ...

According to the report, the design contrasts with the first-generation Nano, where there were no Apple-marked chips. Wedbush Morgan Securities (Los Angeles) believes one of the Apple-marked chips is Samsung Electronics Co. Ltd.'s S5L8701B05 ARM processor (part #337S3291-8701), occupying the socket formerly supplied by PortalPlayer.

The second mysterious chip is the audio driver and codec (part #338S0310), a socket formerly owned by Wolfson Microelectronics plc. NXP, formerly Philips Semiconductors, provided the power-management chip in the older-generation Nano.

[scorche]

Llorean shared this on IRC, and it seems relevant to put here.  Sadly, they are more stylish photos than high-res scans, but they do the job (somewhat)

iPod nano 2G Disassembly pics

This (we believe) is the processor:
http://gallery.ilounge.com/ipod/displayimage.php?album=61&pos=29

Sadly, it might be a while before Rockbox will work under these new nanos.

EDIT: I have been searching on Samsung's site for more information (or what little bits I can scrounge up).  The chip is not listed under either their ARM7 or ARM9 product list.  From this page, we know this about the chip so far:
Large Classification: MOS
Small Classification: L
Feature/Serial Number: 8701
Version: B
Mask Option: 05

EDIT #2: Found dissection instructions.  Will be useful if someone (anyone) who reads this and has their hands on one (and the bravery to take it apart) in order to get us some hi-res scans :-)
Link here

[Potter]

Jus picked 1 of these up..how long does anyone it may concern suspect it will be before there's a rockbox release for it? I don't dare try since the current builds since the "Compatible Ipods" on this site only lists Nano 1st Edition.

[scorche]

Your post was in the wrong place.  I have merged it with the correct place for this.  Also, please search the forums/read the posting guidelines before posting.  

That being said, it appears that the 2nd gen Nano uses a Samsung processor, so Rockbox will require some tweaking to get things working.  It is not worth trying a current Rockbox build on it.

If you wish to push along the port effort, we still need some high-resolution scans of the internals (disassembly instructions are in my previous post).  All that is required is a decent scanner, a shared curiosity of what is inside with us, and the mettle to take apart your new toy (this will not break it provided you are careful).  If you are not willing to do this, and if you have a decent knowledge of C, you may have to wait a while, but you can most likely still help.  If not, you can sit on the sidelines and wait.  

[linuxstb]

That being said, it appears that the 2nd gen Nano uses a Samsung processor, so Rockbox will require some tweaking to get things working.  It is not worth trying a current Rockbox build on it.

It's going to be a little more than "some tweaking" - it's basically a completely new port.

Interestingly, the eetimes article has changed since I read it yesteday.  In the first version of the article, they quoted Wedbush Morgan Securities as speculating that the processor was an "SA5B450X03" - which I assumed should have been the SA58450X03, the same processor that is used in most of the current iriver players.  They've now edited the article to claim it's an S5L8701B05.  

Not that it really makes much difference - both seem to lack any publicly available datasheets, but at least the SA58450X03 has a little bit of information available on Samsung's website:

http://www.samsung.com/products/semiconductor/SystemLSI/DigitalMedia/OpticalASSP/OpticalPlayer/MP3/SA58450/SA58450.htm

So the usual rules for new ports are going to apply:

1) Physically disassemble a player and identify as many chips as possible

2) Disassemble the original firmware in order to find information about the hardware and how to control it.

3) Work out the boot process on the player, crack any encryption that may be in use, and find a safe way to run your own code without risk of bricking.

4) Port Rockbox...

[Llorean]

The Apple chip actually has the number "8701" on it, which according to Samsung's site for decoding the "S5L8701B05" style numbers is the "Serial Number or Function" or something like that (my memory is slightly fuzzy, it was quite a few hours ago I looked there) so my bet is that it is actually the S5L8701B05 chip.

[saratoga]

On the upside, the 940 is basically a classic 5 stage RISC processor, so performnace should be a lot better running the same code as the pp chips.  It should also clock higher.  Additionally, it seems we now have a hardware MAC instruction too, which should be nice for codecs (at least according to the samsung link above).

[linuxstb]

I couldn't resist the shiny aluminium case and picked up a 2nd gen Nano today.

On the outside, it appears the same as any other ipod - the usual key combinations all work (MENU+SELECT for reboot, SELECT+PLAY for disk mode, SELECT+LEFT for diagnostics mode), and the disk is organised the same as the previous ipods - a relatively small (100MB) firmware partition, and a FAT32 paritition taking up the rest of the disk with the usual iPod_Control folder.

As an aside, the diagnostics mode features a "line-in recording" test - so it seems Apple have included recording functionality in the new Nano.  Although it's not accessible via Apple's firmware...

The firmware partition is very similar to the v1.2 Nano firmware partition - it contains an "osos" image with the main firmware, an "rsrc" image containing the Nike-related files, and an "aupd" update image (this is the image that gets flashed).

For those that don't know the Apple firmware partition layout, details of the previous firmware partitions are in the IPL wiki: http://ipodlinux.org/Firmware

But there are some differences.  Firstly, the images are now prefixed with string "DNAN" instead of "!ATA", and secondly both the main firmware and the aupd image appear to be encrypted.

In the PortalPlayer based ipods, the main firmware was unencrypted (but protected by a simple checksum), and only the aupd image was encrypted.

So the first step towards a port would be to understand this encryption - so we can both unencrypt the Apple firmware image for disassembly and also encrypt our own code in order to run it on the ipod.

[Bagder]

But there are some differences.  Firstly, the images are now prefixed with string "DNAN" instead of "!ATA", and secondly both the main firmware and the aupd image appear to be encrypted.

A guess is that "NAND" instead of "ATA!" could imply that they don't access the flash using ATA.

[iconoclast]

I ordered one of these when they were first announced.  From the announcement, it seemed the changes were mostly the case, the battery, and software,  so I'd hoped the internals would be similar enough to the first gen nano that rockbox would easily run.  Now of course we know that's not the case.  Anyway, mine arrived in the mail today, and if there's anything I can do to help out in the porting, please let me know.  I don't have any current experience at the sort of low-level coding the firmware folks are doing, but if you still need scans of the internals for example, or something similar, just tell me what you need.

[Bagder]

We do need scans, but we probably even more need to start working on figuring out the encryption.

[linuxstb]

Another (better) set of photos of the Nano's internals is here:

http://pc.watch.impress.co.jp/docs/2006/0915/apple.htm

[iconoclast]

Another tiny change...  It seems they aren't using the SysInfo file anymore.  On the Nano 2G, the file iPod_Control/Device/SysInfo exists, but is zero bytes long.

[dmd]

I don't know if it's much worth, but here's some close-ups of the chips:
http://arstechnica.com/reviews/hardware/ipod-2g.ars/4

btw doesn't seem to be very durable player, one drop and the screen died  

[ssx]

So the first step towards a port would be to understand this encryption - so we can both unencrypt the Apple firmware image for disassembly and also encrypt our own code in order to run it on the ipod.

is it bootloader who decrypts the code?
someone was able to get it from the new nano?