For 32 bit DSD-over-USB support.

Prepare for removing SAMPLE_FORMAT_DSD_OVER_USB.

Give the caller more control, prepare for DSD-over-USB improvements.

It's a padded 24 bit format.

After a previous refactor, the current code fails on paths that need
expansion (e.g, '~/.mpd/sticker.db'), because we are not passing the correct
path to the sticker database code. Pass the expanded (and previously unused)
string instead of the original string.
Signed-off-by: Dan McGee <dan@archlinux.org>

Implements the dCS suggested standard:

 http://www.dcsltd.co.uk/page/assets/DSDoverUSB.pdf

Fixes build failure on 64 bit.

Make it in_frame_size and out_frame_size, to account for packing.

For simplicity, the MPD core should not have to deal with packing.  It
is rarely used, and those plugins that need it should use the
pcm_export library instead.

Converts padded 24 bit samples to packed 24 bit samples.  Will replace
the packed S24 sample format, which is not used internally.

Improving oss_reopen() by using the very same value that was used
initially.

Let the caller configure the audio_format object.

Merge three functions into one and call get_bitformat() only once.

Delete the now-unused pcm_byteswap library, and optimize the
pcm_export_state object.

Should have been "==", not "!=".

For simplicity, pad the dCS samples to 32 bit.  Packed 24 bit samples
are rarely used.  This patch does not include a real code change,
because there is no user of DSD_OVER_USB yet.

Eliminate support for reverse endian samples from the MPD core.  This
moves a lot of complexity to the plugins that really need it (only
ALSA and CDIO currently).

Don't use audio_format.reverse_endian.

Don't use audio_format.reverse_endian.

Don't use audio_format.reverse_endian.

Halve the sample rate for *:dsdusb:*.

This format is unused since the DSDIFF decoder plugin now reverses the
bit order.

Allow to remove this complexity from the MPD core.

Instead of doing run-time initialisation, use a constant lookup table.