rockbox-dev via rockbox-dev | 12 Feb 19:47 2016

Fw: new important message



New message, please read




iRiver H300 bootloader once more

Hello fellow users and devs,

I have a few questions about the iRiver H300-series bootloader, which I post to
both rockbox mailing lists, because this is also related to building the boot

I have understood, that the H300 boot loader is currently broken, and
flashing it will brick the device, am I right?

In order to install the boot loader, I need an original firmware file for the
iRiver, which is currently very hard to find. I found an 1.29EU version, which
is not the newest that has been released (which is 1.30 or 1.31). Does anyone
have the newest firmware released by iRiver folks?

The newest official release of the rockbox boot loader is very old and it has
some issues: no charging and no RTC alarm. These have been implemented in SVN
(nowadays git) and XavierGr said some seven years ago that SVN revision r12547
has a working bootloader.

However, I'm very afraid of trying to build it and merge it with an old iRiver
firmware file. I don't know if the iRiver build's tool chain has been updated,

XavierGr had a working firmware image merged with 1.31K iRiver firmware, but I
don't know if he is around any more.

But if someone has a working firmware image file, I would be very glad to hear
about it.

Otherwise I would like to have some guidance on building my own boot loader.

Thank you for reading this, and for any help you may be able to provide!




Help needed with iBasso DX80 Port

Hi All,

I’m currently working on a port to the iBasso DX80.  Much like the iBasso DX50/DX90, the DX80 runs a 'modified' version of Android.  The DX80 does run a newer version of Android (5.1) as compared to the DX50/DX90 (4.2).

What I have accomplished so far:
- Unpackeing/repacking of
- Enabling ADB
- Getting a build environment setup using modern-ish tools (thanks [Saint]!)
- Building Rockbox and the ‘ibassoboot’ loader as 'Position Independent Executables'
- Pushing the .rockbox directory to /data/.rockbox and the loader application in /system/bin/MangoPlayer

The loader application is fairly simple and crashes for reasons that I understand.  I am confident that I can modify it to work correctly on the DX80 pretty quickly. 

My primary issue at this point is with Rockbox itself.  The initial issue I ran into is that Android 5.1 requires a 'Position Independent Executable’.  I was able to make some changes directly to the Makefile ( — adding the -pie and -fPIE flags with with some help from users in #rockbox. However, now when I go to run the Rockbox binary on my DX80, it instantly gets sigkilled before any useful information can be gathered:

If anyone has any advice or guidance on where to head next I would greatly appreacite it.


Gerrit ssh keys

Hi devs.

It has been brought to my attention that some of you may have ssh-dss keys registered for access to Gerrit and
since a server update a while ago you are now denied access, while getting error messages that give very few clues.

The reason for the problem is that ssh-dss has been found to be a weak key format and thus has been disabled in
the ssh server. You need to delete the dss key and upload an ssh-rsa key to gerrit instead:




Fw: new important message



New message, please read


robby <at>




Harald Welte at the OpenMoko project was kind enough to allocate the Rockbox project a block of 256 USB product IDs ranging from 1d50:5300 to 1d50:53ff.

Now that we have the PIDs, it's time to assign them to targets. The current plan is to allocate all the stable targets and some of the in-development ones (e.g. ipod6g) a unique PID, and all the rest a "generic rockbox device" PID.

See for more information.


Fw: new message



New message, please read


rockbox-bounces <at>


Gerrit is operational

Hi folks,
thanks to Bjorn, Gerrit is now fully operational again:
- OAuth and OpenID login are back, if you have several accounts you can merge them using the UI
- anonymous browing
- checkout/cherry-pick commands are in the Download menu of the patchset
Happy hacking !


Gerrit / git-review

Hi again,
as you may be aware, with the recent version of Gerrit and following the OpenStack trend, git gained a new plugin called git-review which is used to manage patchset and can be used with gerrit. It is included in most distribution (apt-get install git-review) and is very easy ot use.
However, it requires a little configuration named .gitreview which should be committed at the root of the repository with the following content:


One committed, you can setup git-review on your local repository using:
git review -s
It will check everything is all-right and create a gerrit branch. You only need to do it once.
Once setup is done you only need to remember a couple instruction:

To fetch a review numbered 1024:
git review -d 1024
And if you want a specific patchset, say 7, then it becomes:
git review -d 1024,7

You can then work on it as usual, make sure to amend the commit as usual:
git commit --amend

Finally push the change to gerrit:
gerrit review -R

If you want to create a completely new change, simple run:
git review

I am not entirely sure how useful this tool is but some of you might find it easier to work with.



Gerrit is working again

Hi fellows developpers,
thanks to Björn and others, we now have Gerrit working again properly which means that:
* you can authenticate with OpenID
* you can recover your old google account thanks to the automatic OpenID > OAuth migration
To make this work, simply reach at and select "Google OAuth2". It should lead you to a google page where you have to grant gerrit access again and you now have your old account working !

Note: in case you create a new account using OpenID, it should be possible to link your two accounts by reaching and press "Link to another identifty". However, I am unsure how you can login to your OpenID identity using the new sign-in page


POC Fuze+ scrollstrip support

Dear rockboxers,

here you go, Fuze+ with up/down sliding support!
To be honest I wrote this patch to promote my scrollstrip patch (g#786) 
and give it - or rather it's extension - a larger tester base.

The challenge of this code is: while the original patch works quite well 
with up/down movements, the Fuze+ also uses the centre as Select button 
(and other scrollstrip targets as well, like Philips HDD1630). So it's 
mandatory to implement long-button-presses, at least for the centre 
button, to not completely break rockbox' user interface philosophy.

For this reasen this patch allows different strip-button-modes, 
accessible via:
Settings > General Settings > System > Scrollstrip handling
* "Doubleclick mode": a short tap directly followed by a tap-and-hold 
will send a long button press (like your laptop's touchpad).
   Disadvantage: you get a latency on single taps.
* "Long press mode": just stay long enough on the same spot, and you'll 
get repeated button events.
   Disadvantage: you get a big delay for long button presses, and you 
get inadvertent button activity if you accidentally stay too long on the 
same place.
* Both of the above can be set to only work on the centre button.
* "Off": The usual behaviour; Sometimes the best setting. Try playing 
Doom with scrollstrip mode :)

If you like to play with the timing constants: I was too lazy to make 
them configurable via menu, so look out for BUTTON_DOUBLECLICK_TIME and 
BUTTON_LONG_PRESS_TIME. You find them in 

You may also like to play with the Settings > General Settings > System 
 > Scrollstrip speed and afterscroll settings.

I'd appreciate any comments (via mailing list, gerrit or via IRC - I 
read the logs), especialy to these questions:

* I don't think we need all of the above settings. So which setting do 
you like best, and which can be discarded?
* Any Opinions about the doubleclick/long-press timing setting? Should 
it be configurable?

Have fun,