I think there's an ascii art layout of the buffer somewhere in the code,
but yours is pretty close to correct.

The reset_buffers method and pcmbuf_init methods give this information
after a careful enough reading :-P

Here's an edited version to the best of my knowledge
--- audiobuf (initially)
tagcache (optional, size varies)

----
dircache (optional, size varies)
----
voice thumbnail buffer (optional, size MAX_THUMBNAIL_BUFSIZE)
---
voice file (optional, size of voice file)
---- audiobuf (after initialisation)
codec malloc buffer (size MALLOC_BUFSIZE)
---- filebuf (audiobuf[MALLOC_BUFSIZE])
compressed codec data + guard buffer to allow guaranteed block size reads
--- filebuf + filebuflen
audio codec IRAM (size CODEC_IRAM_SIZE)
---
voice codec IRAM (optional, size CODEC_IRAM_SIZE)
---
audio code DRAM (size CODEC_SIZE)
---
voice codec DRAM (optional, size CODEC_SIZE)
--- audiobufend - pcmbuf_size + pcmbufdesc_size
PCM buffers, including descriptors
--- audiobufend
Thanks for working on this code in my absence,
Brandon Low
On 2006-08-16 (Wed) at 18:55:38 +0100, Steve Bavin wrote:
> Hi,
> 
> This is a call to all developers who know and understand playback.c. ;-)
> 
> Can somebody produce, or point me in the direction of, a basic memory map
> showing how the audio buffer is used on SWCODEC targets.
> 
> I have a rough, and very probably incorrect outline:
> 
> --- audiobuf (initially)
> tagcache (optional, size varies)
> ----
> dircache (optional, size varies)
> ----
> voice thumbnail buffer (optional, size MAX_THUMBNAIL_BUFSIZE)
> ---
> voice file (optional, size of voice file)
> ---- audiobuf (after initialisation)
> codec malloc buffer (size MALLOC_BUFSIZE)
> ---- filebuf
> PCM buffers, including guard buffers/descriptors
> --- filebuf + filebuflen
> audio codec IRAM (size CODEC_IRAM_SIZE)
> ---
> voice codec IRAM (optional, size CODEC_IRAM_SIZE)
> ---
> audio code DRAM (size CODEC_SIZE)
> ---
> voice codec DRAM (optional, size CODEC_SIZE)
> --- audiobufend
> 
> I intend to put this in the wiki if this information doesn't already exist
> in there (and if it does, I can't find it!)
> 
> Thanks,
> Steve Bavin
> 
> 
>