Rockbox.org home
release
dev builds
extras
themes manual
wiki
device status forums
mailing lists
IRC bugs
patches
dev guide



Rockbox mail archive

Subject: Re: Cowon D2 EABI

Re: Cowon D2 EABI

From: Thomas Martitz <thomas.martitz_at_student.htw-berlin.de>
Date: Mon, 14 Jun 2010 00:54:24 +0200

Am 13.06.2010 23:04, schrieb Rob Purchase:
> Hi all,
>
> I noticed today that the Cowon D2 build has been switched to using the
> EABI compiler. I'm sure this would be good thing if it actually worked :(

I'm sorry that it doesn't work :(

>
> While Rockbox boots and I can play music, numerous things are broken:
> the touchscreen does not work at all, backlight fading does not work,
> and the RTC displays --:--. The common factor in all these things is
> the PCF50606 PMU driver, so it looks like the new compiler has broken
> something in that area (most likely the timing of the bit-banging, at
> a guess).
>
> At the moment I do not have the new compiler installed, nor the time
> to do so and investigate a fix. So can I ask that this change is
> reverted until it is tested and working properly, and also that in
> future significant changes like this aren't committed until somebody
> has (even briefly) tested that it works? Obviously I know I haven't
> been around here too much lately, but I'm certainly not the only
> person on this list with access to a D2.

I'm sure it can be reverted, since it has shown bad effects. But I would
also like to ask you to sort out the problems (once you have some time)
since we want the same gcc for all ARM targets, and this "dual-mode" is
supposed to exist just for the transition phase.

Also, as you just said you haven't been around lately, it's hard for our
project, considering the growing number of targets, to wait for a port
maintainer to actually show up and do some work. If the maintainer is
not around, then the port is effectively not maintained. It can get
broken if it remains unmaintained.

>
> I realise the D2 is an "unstable" target, but there are quite a
> significant number of users who rely on the current builds to actually
> work, and I think this is something we should aim for at all times
> (obviously within reason).
>

We've always told that current builds may actually be not functional and
that you should never rely on current builds (they are not dependable).
That's one of the reasons every target should aim for a stable status.

I think it was a good idea to simply switch the gcc for all non-stable
targets.

Best regards,
Thomas Martitz
Received on 2010-06-14


Page was last modified "Jan 10 2012" The Rockbox Crew
aaa