Re: Why sound isn't working on the TP2.
 

Cool, I'll try it out in an about an hour or so.

Re: Why sound isn't working on the TP2.
 

this has been very interesting. thanks to all that have helped so far. at this point the windows driver running under android is pretty much in the same shape as the native driver. ie it complains about never getting an interrupt.

i believe this is caused by one of the things we are not doing. probably klt or spm. the native dll's for these can probably be used as well. i will start trying to get those loaded up soon.

for those wanting to help. wrapper could be written for microp spm and adsprtos. collecting data on these dll's will help us get further along. i will write these wrappers if nobody else does. i am moving to the caribean right now. arriving on the 10th. so things are a bit hectic.

Re: Why sound isn't working on the TP2.
 

thanks for all ur hard work guys. please keep updating us on the progress. will keep checking regularly if u need any testers.

Re: Why sound isn't working on the TP2.
 

Could somebody post a recmodded pm.dll from a WM6.0 gsm rhodium?

Re: Why sound isn't working on the TP2.
 

As far as I know, Rhod only shipped with 6.1 Pro, there was not an option for WM6.0 Std. I'm downloading the original AT&T shipped rom right now. I'll post the recmodded OEMDrivers and XIP in a few minutes. (File is still downloading)

Re: Why sound isn't working on the TP2.
 

http://www.4shared.com/file/JA3DWd32...ATT_stuff.html

This is the WM6.1 (earliest I could find) GSM RHOD OEMDrivers and XIP. PM.dll is in the XIP.zip file.

I'm looking through the nk.exe again, and I see it does Interrupt Masking, could this be causing the issues for sound and 3d? Here's the exact text line from the nk - '[K][GuardDog][Interrupt_MASK] owner proc: %s, cur proc: %s, thrd: %08X, int_id: %08X, masked?'. It seems to mask a few other things as well.

Re: Why sound isn't working on the TP2.
 

Thats just a text line i think. If you see the address 0xc0000010 being modified then that might be worth a look.

Check out the wiki on the topaz section. We have a dump there of the whole VIC section there. If the asm you see matches some of those values, let me know and we can take a look.

Do you have a link for 6.1 CDMA rhod files? i'm interested in comparing the nk.exe to the 6.5

Re: Why sound isn't working on the TP2.
 

Here is the link to the Dzo's code that i wanted you to look at. The irq.c is modified a bit and there are references to the dsp irq.

http://androidhtc.git.sourceforge.ne....git;a=summary

Re: Why sound isn't working on the TP2.
 

[acl]. are you talking about irq.c from mach-msm? i diffed it with ours and see very few differences and no references to adsp.

wozzer, thanks a bunch.

I thought pm.dll would have stuff like SPM_Open and SPM_IOControl, instead it has only one export, start, and no imports. very weird. I wish there was an easy way to scan through all of the dlls for an export.

EDIT: i found the export in HtcPM.dll

Re: Why sound isn't working on the TP2.
 

sorry i meant vogueirq.c. I know its not our device, but for it looks like the vic types are preset to the same vals that are on wince. We actually set them all to 0. just figured it was worth a look. Also look for the comment "fix for adsp handler".

Once you give me what your pmem.c is set to for sounds i can start testing these changes.

Re: Why sound isn't working on the TP2.
 

[ACL], I'm downloading what I think is a 6.1 CDMA shipped rom right now, I'll post the XIP and OEMDrivers for you once it's done.

Here is the WM6.1 Sprint OEMDrivers and XIP

Re: Why sound isn't working on the TP2.
 

thanks man.. ill check it out. im hoping it's really different ..

Re: Why sound isn't working on the TP2.
 

I checked the Sprint 6.1 NK, and it still has the guarddog, but I don't see anything about masking the interrupt. So that is probably not causing the issue.

Re: Why sound isn't working on the TP2.
 

[acl]. i don't see what the point of pmem.c is. Looks like it is there so userland applications can have access to physmem that might move around between devices. Basically the same thing and wce_pmem.c. For sound, the only thing that would use it is htc_acoustic, which we don't have. qdsp5 driver finds its own memory and doesn't use pmem.

wce_pmem and wavedev are mapping in *lots* of phys mem. I count at least 5 16mb regions.

my technique of mmapping /dev/mem, doesn't require it. pmem is probably a prettier solution for when it works.

Re: Why sound isn't working on the TP2.
 

[acl]. i tried removing powermanagement support since i am going to be loading the wince one. Just having it removed gives me a dsp related int somewhere, although it still does not work. Also it is crashing my wce_pmem somehow. That is going to be tough to figure out.

Re: Why sound isn't working on the TP2.
 

ur guys pockets will be LOADED when u get everything working ahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh hhhhhhhhhhhhhhhhhhhhh

Re: Why sound isn't working on the TP2.
 

jonpry : i tried the dzo irq changes. It doesnt break the build but it didnt help with 3d either. Replicating the VIC settings dont work too well either since our code disables them anyways. For example i left the 3d bit enabled but once hw3d.c loads, the bit gets disabled during init. I'm not sure if thats the case for all the devices.

Re: Why sound isn't working on the TP2.
 

[acl]. i don't think we should let anything enable or disable interrupts. We should just setup VIC just like WinCE and then ignore any calls to irq_enable or irq_disable other than hooking up the handlers. Presumably all the linux drivers will be requesting interrupts to get setup the way they already are anyways.

Re: Why sound isn't working on the TP2.
 

Jonpry: There was some progress with the nand boot. Wozzer was able to boot but we are crapping out during the usb probe like expected.

Re: Why sound isn't working on the TP2.
 

As relatively quiet as this thread has been (stuck to the topic very well!) I find it necessary to state the obvious and just say a HUGE thank you to everyone working so hard on all of this. Not just the sound, but the guys trying to install to NAND for us, the original devs that first got android to boot goofy touchscreen and all. Its guys (and girls) like you that make computers truly interesting. It's because of you that these devices are able to reach their true potential and I think its an absolute shame if you all don't end up in some very high-end programming or systems engineering job. I speak as someone who fixes computers frequently, servers, pcs, you name it, and am regarded as somewhat of a 'guru' around my friends and family, and I know full well your skills far FAR surpass mine.

Just figured a big thank you was in order. With talent like this, I know we'll get it eventually if it CAN be gotten.

Re: Why sound isn't working on the TP2.
 

Dood where the f*ck are u? lol. WisTilt2 claims to have irq's working.

Re: Why sound isn't working on the TP2.
 

whooooooooooooooooooooooooooooooooooooooooo i dont know what acr means? but im sure its good news whoooooooooooooooooooooooooooooooooooooooooooooooo ooo:headbang:=D>

Re: Why sound isn't working on the TP2.
 

He's moving to the Caribbean.

Re: Why sound isn't working on the TP2.
 

can i move with him. you got a extra room hit me up. lol

Re: Why sound isn't working on the TP2.
 

what can we test?! :)

Re: Why sound isn't working on the TP2.
 

Not sound... just IRQ's

Re: Why sound isn't working on the TP2.
 

Probably Been said/tested before but hasnt the hero almost the same hardware? (cpu atleast)
Maybe the hero drivers would work?

Re: Why sound isn't working on the TP2.
 

if any testing needs to be done on hero i might be able to help.. i have both hero and tp2

Re: Why sound isn't working on the TP2.
 

from my following this thread for some time it comes down to the fact that the tp2, the topaz , blackstone have some kind of very different architecture which is why even though the port works great for other phones it doenst work ( atleast with sound and 3d ) on ours.

And i think the problems in the kernel level not the driver level ( correct me if im wrong as my programming and reverse engineering skills are limited to cheering the actual developers on).

Re: Why sound isn't working on the TP2.
 

They'll figure it out eventually

Re: Why sound isn't working on the TP2.
 

ditto, i run a TP2 and my wife a hero

Re: Why sound isn't working on the TP2.
 

Hi All. I've been hacking away here. Still no interrupts. But I am fairly close to being able to use wince DEX,SPM,RPC,SHM,KLT. which hopefully will get things moving along in the right way.

If there are any linux experts out there. I am having trouble mmap'ing contiguos regions. Like mmap(0xfu,0x1000,somephys), then mmap(0xfu+0x1000,0x1000,someotherphys)

Seems like it needs at least 3 pages for it to work. I *need* it to be contiguous though. Anything including hacking the kernel would be fine. I'm just not sure what the trouble is.

Re: Why sound isn't working on the TP2.
 

Have you been able to replicate the half second of sound that was reported early this week?

Re: Why sound isn't working on the TP2.
 

I haven't got that half second of sound. I've talked to the guy, but he is pretty closed lipped about exactly how they got it to work.

Re: Why sound isn't working on the TP2.
 

been silently watching this post mainly because im an electrical engineer who was never into programing or software dev in general but the fact that the rhodium port has been in the works since february and that people are holding back is just crazy and frustrating, specially when the HD2 went up and running in no time.

Re: Why sound isn't working on the TP2.
 

Agreed. There's a reason most professional development teams have more than 1 single person. Everyone has different skills and ideas that should all be used together to solve complex issues like this. Your work will be greatly appreciated by many, whether you do it by yourself or with the aide of other like and different minded people other than yourself.

Regardless, thank you everyone involved and keep up the good work!

Re: Why sound isn't working on the TP2.
 

That is the first I have heard about any sound. Thats sounds like a buffer issue.

Re: Why sound isn't working on the TP2.
 

Lol.. Cotulla himself said the tp2 is very difficult.

Frustrating yes.. but imagine how we feel. I been banging my head on irq issues for months now. Out of desperation i decided to make a nand version which we got booting and we are in the same boat.

Re: Why sound isn't working on the TP2.
 

ACL i know how you feel and YES the TP2 is a nightmare, if only i ever run into the engineer behind it, that would be interesting lol

sometimes when your working on something else is when you figure out what you were working on initially.
how is it on nand ? compared to a 16gb class 2?
im sure you guys might have heard this before but i was recently tinkering with a friend's HTC Dream/G1 and i gota admit, huge hardware similarities.
oh well i shall wait but wish i could help and sorry if i took this thread off course

Re: Why sound isn't working on the TP2.
 

I found this post on a different thread and figured I would share it in case it helped. It is obviously about the Rhodium sound problem.

"Something just occured to me... I'm a sound engineer... so i know all about feedback, sends/returns etc in a live environment, which is basically the same wherever you have a live microphone near a speaker....

In my industry there are 2 devices you can insert into the "sound chain"...

1. If an artist screams into a microphone, you insert a compressor, which essentially brings the input gain down, preventing the microphone from distorting....
This is practically achieved by "routing" the microphone VIA a compressor 1st, the compressor does it's processing, which you set manually, consisting of:
- Threshold (Input Gain)
- Ratio (How much of the threshold to process)
- Output Gain (How loud the output is)

2. If your microphone environment is too noisy, in otherwords if you are scared of feedback occuring, you can insert a GATE into the sound chain. This basically keeps the microphone "off" (or gated) waiting for input signal to be loud enough, which then switches the microphone on. Practically a simple example would be on a drumkit, where the drum A must not pick up drum B when the drummer hits drum B... the microphone on drum A is "Gated" and only opens when it is hit and the signal is loud enough (or close enough to the microphone)

You also set gates manually, consisting mainly of:
- Threshold (Input Gain)
- Ratio (How much of the threshold to process)
- Attack time (How quickly to process)
- Decay time (How quickly to release)
- Output Gain (How loud the output is)


In our industry, it is a common mistake where new engineers under stress forgets to take the "insert button" out of the chain... so they struggle to get any sound going to the speakers until they take out the insert button.

So with example a (compressors) and example b (gates), either of those 2 components may cause the speaker to stop working, if the device is inserted.

If you say that everything is being routed to the 3.5mm jack, what are the chances the designers have some sort of gating/compressor system in the sound chain, to prevent feedback/echo from happening... and the default settings are simply set to "off" or "maximum"... and its inserted by default.

...like i said - these devices are found everywhere... just in my industry it's something inserted physically as a separate device. So maybe everything works... it's just something in the chain which doesn't....

...hope it makes sense to someone"


Now, this makes me think that the problem has to do with the dual microphones the rhodium uses for noise cancellation. I hope this isn't old news and maybe it is helpful to someone...