Cowon D2

[Llorean]

Why are you hinging everything on one person?

At any point, any of you can pick up and continue working on things. Those who wish to "help" with NAND can attempt to work on it on their own, see if they have any breakthroughs themselves, etc.

Instead of talking about how to "figure a way to keep the port developing", why not just develop?

[3amsleep]

Why are you hinging everything on one person?

At any point, any of you can pick up and continue working on things. Those who wish to "help" with NAND can attempt to work on it on their own, see if they have any breakthroughs themselves, etc.

Instead of talking about how to "figure a way to keep the port developing", why not just develop?

i understand what you mean, and I didn't want to hinge everything in shotofadds, but i think it would be nice if he can share some knowledge on the NAND driver so other people can catch up and do some work there. It seems to me that many people is counting on shotofadds to finish that driver, so if he says that he dosn't want to expend much time doing that, people will start working on that.

Personally i'd love to just go and finish the driver myself, but I have next to no skills/knowledge in that area, in fact many times I tried to understand how it works/what needs to be done, but i've failed miserably.

 I think comunication and organization have been missing in this project for the last year or so, I think it would be really helpful if shotofadds beign the project main developer states who is working in what, and where help is needed how.

[shotofadds]

Apologies in advance for the long post..

Now I don't mean to be rude, but please realise that everyone contributes to Rockbox in their spare time. That is, the time left in the week after working, eating (this is a big one for me!), sleeping, socialising, spending time with my friends/family and all those other mundane time-consuming things like trying to find a new place to live.

If there's some time left in the week after that, and I choose to spend it on Rockbox, then great. If not, then I really don't need to explain myself.

Also, don't assume that no posts here means that no work is going on (for example, I actually read your post while waiting for a D2 build to compile). Tasks like disassembling and fully understanding complex routines in the original firmware are long, slow jobs where you potentially won't see the results for months, and far more of that will be needed if we are to ever have read/write NAND capability.

Having said that...

It's quite obvious that the D2 port hasn't progressed as far as it could since it got audio playback and a reliable read-only filesystem last year. The reason for that is simple, and it's exactly as Paul says: as far as I know, no-one else has expressed any serious interest in working on low-level aspects such as the NAND driver or SD card driver. These are the two most important obstacles to usability of the D2 port, far moreso than issues with the touchscreen UI (imho).

Don't get me wrong, people have contributed all sorts of improvements and tweaks over time, and these are hugely appreciated. Unfortunately the fact remains that no-one seems to want to tackle low-level issues. The other Rockbox devs who own D2s (there are a few) have enough on their plate already, before worrying about "new" ports like the D2.

As far as the SD/MMC driver is concerned, the necessary specifications are freely available, links to the TCC7801 controller specs have been leaked several times on here and elsewhere, and the Cowon firmware is available for reverse engineering. Drivers for other players have been written using far less information than that in the past, but so far no-one has decided to work on it...

The NAND driver is a bit of a special case because it's rather complicated and there is no documentation whatsoever. I'm part way through documenting in plain English everything needed to read a logical sector of data (ie. what the driver does already). When this is complete I'll add this to the wiki in the hope of attracting some attention from others. Expanding the driver to a read/write capability may be possible one day, but only if a very significant amount of time and effort is spent on it.

I think comunication and organization have been missing in this project for the last year or so, I think it would be really helpful if shotofadds beign the project main developer states who is working in what, and where help is needed how.

The CowonD2Info wiki page already has sections listing Known Problems and Work To Be Done, and since I try to keep the page up to date with any significant changes, those are still current. What would you like to see beyond that?

Finally, let me just say I definitely haven't lost interest (remember I originally bought my D2 just to work on this!). I just don't have as much spare time to spend on it at the moment.

[GodEater]

If there's some time left in the week after that, and I choose to spend it on Rockbox, then great. If not, then I really don't need to explain myself.

Too right mate. You don't owe anyone anything.

[grantmasterflash]

This is probably a terrible time for a question like this, so first I'll say thanks to shotofadds and everyone that works on the Rockbox port.  Also, I realize the D2 port is far from finished, so it's a bit early for feature requests.

With all that said...

I see in some of the released Rockbox ports that there is an option to flip the display upside down.  Would there be a similar way to make an additional option in Rockbox for D2 for moving the display orientation from landscape to portrait mode (like a 90 degree and 270 degree option, so that the headphone port could be oriented at the top or bottom of the device)?

I'm guessing the answer is: Yes, get the code and do it.  If that is the case, I'll simply say that I bet there are much more capable programmers here than me, so if you get an inclination to this, I'll bet it would be very much appreciated by the D2 Rockbox community.

[shotofadds]

I don't think flipping the display orientation at runtime would be possible without major changes. More realistically, you could potentially make a build that uses a 240x320 resolution instead of 320x240, "just" by changing the LCD_WIDTH and LCD_HEIGHT defines and making the LCD driver draw things sideways... I believe some other ports can be built this way already.

But yes, like everything else, it needs an interested programmer to implement it! I'm not saying I'm not interested, but there are more fundamental things to fix first...

[grantmasterflash]

I guess the touchscreen grid would have to be reoriented as well, so that taps would correspond with how the screen layout has changed.

Flipping during runtime wouldn't be necessary, but I think it would be cool in the future to have another build or just instructions on what to change in the code so you could compile your own vertical Rockbox for D2.

Just an idea to throw out there in the idea pile.

[kugel.]

shotofadds, just a small remark regarding r21465, I don't think you actually need USEC_TIMER. The AMS Sansas are doing it with plain current_tick (each 5 ticks 1 yield), so 77x could use this too.

[shotofadds]

shotofadds, just a small remark regarding r21465, I don't think you actually need USEC_TIMER. The AMS Sansas are doing it with plain current_tick (each 5 ticks 1 yield), so 77x could use this too.

That would work too.

I just copied another driver (I forget which) with the view that in the long run 77x would have USEC_TIMER too. 77x is somewhat unfinished in the timer department, eg. it currently doesn't have udelay() either. That could be fixed easily using the same method I used for 78x.

I think I'll leave it how it is, and implement the usec timer on 77x.

@everyone else, you'll be pleased to know that the "delay before playback starts" bug is fixed in r21465. I'm also working on a possible fix for the "Max NAND write caches reached" panic that some people reported.

[andiator]

The new bootloader doesn't work for me. It hangs in bootscreen showing rev number. The first boot was funny: The screen had a distortion like an old analog tv set with poor receiving.
I have no idea what's wrong.

[shotofadds]

The new bootloader doesn't work for me. It hangs in bootscreen showing rev number. The first boot was funny: The screen had a distortion like an old analog tv set with poor receiving.
I have no idea what's wrong.

* shotofadds slaps his forehead.

Yes, you're quite right. That change was never going to work for bootloader builds. The D2 doesn't have interrupts/threading in the bootloader, so adding calls to yield() is a recipe for disaster...

Just use the old bootloader for now, and I'll fix it as soon as I can.

[andiator]

Yes, I already have installed the old one to see if it still works.
But hey, that's a nice improvement, the big delay is no more 

[shotofadds]

As a quick update: I've managed to get interrupts/threading to work in the bootloader now, so a fix will be coming soon.

Unfortunately my SVN tree is now in such a mess it'll take a little while to clean things up and commit. Hopefully I'll have it fixed tomorrow.

[shotofadds]

The above bootloader fix, and a fix for the "Max NAND write caches reached" panic that some users were seeing, are now in SVN (r21486).

Also, it seems that JPEG album art has become broken in recent versions (although this works in the sim, and the JPEG viewer plugin itself still works on target). I assume one of the recent optimisations isn't playing nicely on the D2 - I'll look into it soon.

EDIT: The JPEG problem is fixed now, too.

[buk]

Thanks for your work on FS#10415 shotofadds !
Work well on my D2 with a 4GB Kingston SDHC.

It is a really good replacement for NAND write support.