Jump to content

ATI AC 97 Sounds (SB400) effort

Andras Kenez

By Andras Kenez,
in General Discussion

 Share
696 posts in this topic

Recommended Posts

EdFlyer
EdFlyer
Posted
Posted

Has anyone gotten a KP (kernel panic) with their hardware and the sound drivers in linux? We should make up a list of all the hardware configurations that work in linux with what build (live cd's) and then maybe we'll have to write custom versions for each version. It appears that even though these all share the same device ID, they are in fact wired differently.

 

Errrrrrrrr to KP's and their ability to not leave debugging information.

 

-ed

burty89
burty89
Posted
Posted

OK I know my previous efforts haven't done very much in terms of solving the problem, but I think that the problem is getDMAEngineHardwarePointer returning an invlid value.

 

I've changed the method to the following and it seems that the kernel panics have stopped, but I get audio repeating itself a few times after it should have stopped, so its still not right...

 

IOByteCount CLASS::getDMAEngineHardwarePointer(IOAC97DMAEngineID engine)
{
IOByteCount position;

CHECK_DMA_ENGINE_ID(engine, 0);
ATIIXPDMADesc *dma = dma_channels[engine];

// Get physical address of sample buffer
UInt32 tmp;
IOPhysicalAddress buf_addr = dma->sampleMemory->getPhysicalSegment(0, &tmp);

// The DMA current pointer returns a physical memory address
position = regs->read(dma->currentPointerOffset);

// Subtract the sample buffer base address to get an offset
IOByteCount ret = position - buf_addr;

if (ret < 0)
	ret = 0;
if (ret > dma->ringBufferSize)
	ret = dma->ringBufferSize;

return ret;
}

 

As you can see at the end of the code I made it check that the return value is between 0 and dma->ringBufferSize and if not then to change it so it is. I'm not sure exactly what the bounds for the return value should be, I just guessed these from reading through the code (I remind you again this isn't my programming language...).

 

If you try the above code you should see what I mean by the audio repeating itself. I think therefore that the problem has to be in getDMAEngineHardwarePointer as I no longer get kernel panics/system lockups.

 

As always if I'm completely wrong I apologize.

thedopefishlives
thedopefishlives
Posted
Posted

You may well be right, since I didn't do any error checking on that value (I just took it out because this seemed to be correct). Give it a good, long burn-in test and see if it works. I have a couple of other things to try as well, and I'll throw my original error-checking code in too, and if it all works out I'll release 1.0.4.

thedopefishlives
thedopefishlives
Posted
Posted

Hey Burty, to fix the sound-repeating problem, you need to increase the buffer size in one of the activateConfiguration functions, I can't remember which offhand. The correct ringBufferSize is the full size of the sample buffer, which is bufferBytes * ATI_MAX_DESCRIPTORS.

OldSkoolGamer
OldSkoolGamer
Posted
Posted

Hey Ed,

 

Did you ever get the sound to work on your laptop ? I have the same issue as you with it showing as line out, no audio through headphone or speaker jack though, but I can hear a small pop when a sound is supposed to play on the internal speakers.

SkylerF
SkylerF
Posted
Posted

I've had my laptop on for over a day now running 1.0.3 with no problems, both video and music, other than the video is a little jumpy (but I think that's a video driver issue). I had to use Tariq's build, since thedopefishlives's always gets a link/load error for me.

thedopefishlives
thedopefishlives
Posted
Posted
I've had my laptop on for over a day now running 1.0.3 with no problems, both video and music, other than the video is a little jumpy (but I think that's a video driver issue). I had to use Tariq's build, since thedopefishlives's always gets a link/load error for me.

You have a better record with this driver than I do, man. Thanks for the encouragement. I'm going to try to replicate Tariq's build by doing exactly what he did. I think on one of my builds I ran a "clean all" that may be causing all the problems.

zelda
zelda
Posted
Posted

I just updated to 10.4.7 (working great, thanks Jas/Dax)

and changed this in AppleAC97Audio.kext/Contents/Info.plist

 

<key>com.apple.iokit.IOAudioFamily</key>

<string>1.1d25</string>

 

working fine...

 

Dope I have also noticed that if I change any code in any other directory (like IOAC97Family) I always get link/load error... try the compile command I gave earlier.. it works fine with you package.

Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...