JustKernel

some code snippets to explain the audio implementation in Virtualbox and how the code changed from earlier QEMU based monolithic blob.

Taking an example of AC97 Device: DevIchAc97.cpp

PPDMGSTVOICEIN voice_pi[2]; // corresponds to 2 LUNS in audio in direction (recording)
+ /** PCM out */
+ PPDMGSTVOICEOUT voice_po[2]; //corresponds to 2 LUNs in audio out direction (playing)

for (lun = 0; lun < 2; lun++)

{

//call intemediate audio driver (DrvAudio) to write the audio to the actual hardware for playing for both the LUNs simultaneously
copied = pThis->pDrv[lun]->pfnWrite(pThis->pDrv[lun], pThis->voice_po[lun], tmpbuf, to_copy);
}

Intermediate audio driver: DrvAudio.c

drvAudioWrite() // implementatio of pfnWrite

{

//calls the host backend to write the audio to the actual audio hardware attached to the host.

bytes = audio_pcm_sw_write(pThis, sw, buf, size);

}

audio_pcm_sw_write function implemented in the host audio driver (host backend) does the task of actually writing the samples on the audio hardware on the host.

mail_to : anshul_makkar@justkernel.com

Tags: audio