CFE_SB_ReceiveBuffer() -- switch timeout to int32

[CDKnightNASA]

Describe the bug
Per discussion at the Jan. 6, 2020 CCB, @jphickey suggested that the timeout parameter should be int32, not uint32. This should be changed.

See also #1063

Reporter Info
[email protected]

[jphickey]

CFE_SB_ReceiveBuffer() is already int32 (see

cFE/fsw/cfe-core/src/sb/cfe_sb_api.c

Lines 1470 to 1473 in 1ede295

	int32 CFE_SB_ReceiveBuffer(CFE_SB_Buffer_t **BufPtr,
	CFE_SB_PipeId_t PipeId,
	int32 TimeOut)
	{

)

But CFE_SB_ReadBuffer (internal function) is not.

Unless there is a good reason to "hide" OSAL definitions here (and I can't think of a good reason, because everything already depends on osapi.h anyway), I'd rather simplify and just make it so we have:

#define CFE_SB_PEND_FOREVER   OS_PEND
#define CFE_SB_POLL           OS_CHECK

... rather than redefining these values.

Right now we use the same underlying values but redefine them, so to be correct/safe the ReadQueue needs to translate. With them defined this way we still have symbol independence (so SB can diverge from OSAL if it needs to, though I don't know why it would). Then the ReadQueue function doesn't need to translate this value, so long as it shares the same definition.

[jphickey]

Actually, just assign this one to me. As I'm going through code looking for race conditions in issue #1073 the entire design of CFE_SB_ReadQueue() is broken - it is accessing PipeDscPtr without the SB lock. I will replace this with something that works.