compress: Remove hardcoded limit on length of poll() wait

For best power-saving we want to sleep on poll() for as long as
possible, we don't want to wake unnecessarily for arbitrary time
limits. Adds function compress_set_max_poll_wait() so that
client can configure the poll() timeout.

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Fix incorrect error message on compress_resume()

Was reporting compress_resume() errors as "cannot pause"

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Change running from bitfield to plain int

Declaring running as a single-bit bitfield does not save any space
or give any coding advantages, but does leave potential opportunities
for the compiler to be inefficient in how it handles bitfields compared
to how it would handle a machine int.

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Remove unused buffer_size member from struct compress

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Change compress_write() to take buffer as const void *

The input buffer to compress_write() should be const, and is
better defined as void * so the caller doesn't have to cast.

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Add function to get timestamp in samples

The compress_get_hpointer() converts the timestamp into actual
time. But Android needs it in samples. To avoid the inefficiency
of using compress_get_hpointer() and converting into time and then
back into samples, this change adds compress_get_tstamp() which
returns the raw sample count.

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: fix arithmetic exception in compress_get_hpointer

In compress_get_hpointer, check for invalid sample
rate to prevent devide-by-zero exceptions

Signed-off-by: Xuemin Su <xuemin.su@intel.com>
Signed-off-by: He Bo <bo.he@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: cache the config values passed

The library should not rely on users pointer for config data, so cache it for
use afterwards

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: remove unused variables

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: make variable 'running' as unsigned

as we never expect this to be negative

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Do not put newlines on error messages

In the oops function text is appended onto the end of the supplied error
message, so newlines appear in the middle of error messages. This patch
removes all newlines from supplied error messages.

Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Return error messages correctly from compress_open

The allocated compress object will be freed and bad_compress will be
returned, so error messages written to the allocated compress struct
will be lost. This patch writes error messages to bad_compress for error
paths in compress_open, ensuring that error messages are returned
correctly.

Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Return error when failing to allocate compress struct

Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: Extend limit on error messages

32 characters is a very short size for the error message, extend this
to 128 characters to prevent clipping on the majority of messages.

Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>

compress: add compress_resume

Signed-off-by: Namarta Kohli <namartax.kohli@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>

compress: write when we have a fragment to write

this needs fix when we do the drain, while draining we should write anyway
whatever we have

Signed-off-by: Namarta Kohli <namartax.kohli@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>

compress: fix the timestamp calculation

use u64 for extended range and also ensure that seconds and nano seconds are
claculated seprately

Signed-off-by: Namarta Kohli <namartax.kohli@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>

initial creation

Add the initial files of tinycompress and utility player cplay

Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>

Initial test commit...