Clean up EVS ports implementation

[skliper]

Fix a few minor issues with EVS ports implementation:

1. Although a macro is used when checking bits of the "OutputPort" mask, the value is still effectively hard coded with the shift, so the macro cannot change without also breaking these checks (violates the spirit of using the macro to begin with). A simplification of the conditional will allow them to change independently (just mask and non-zero):
   

   cFE/modules/evs/fsw/src/cfe_evs_task.c

   Lines 831 to 847 in 73c338d

   	/* Process command data */
   	if (((CmdPtr->BitMask & CFE_EVS_PORT1_BIT) >> 0) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort |= CFE_EVS_PORT1_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT2_BIT) >> 1) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort |= CFE_EVS_PORT2_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT3_BIT) >> 2) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort |= CFE_EVS_PORT3_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT4_BIT) >> 3) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort |= CFE_EVS_PORT4_BIT;
   	}

   
   

   cFE/modules/evs/fsw/src/cfe_evs_task.c

   Lines 885 to 901 in 73c338d

   	/* Process command data */
   	if (((CmdPtr->BitMask & CFE_EVS_PORT1_BIT) >> 0) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort &= ~CFE_EVS_PORT1_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT2_BIT) >> 1) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort &= ~CFE_EVS_PORT2_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT3_BIT) >> 2) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort &= ~CFE_EVS_PORT3_BIT;
   	}
   	if (((CmdPtr->BitMask & CFE_EVS_PORT4_BIT) >> 3) == true)
   	{
   	CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort &= ~CFE_EVS_PORT4_BIT;
   	}

   
   

   cFE/modules/evs/fsw/src/cfe_evs_utils.c

   Lines 469 to 502 in 73c338d

   	if (((CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort & CFE_EVS_PORT1_BIT) >> 0) == true)
   	{
   	/* Copy event message to string format */
   	snprintf(PortMessage, CFE_EVS_MAX_PORT_MSG_LENGTH, "EVS Port1 %u/%u/%s %u: %s",
   	(unsigned int)EVS_PktPtr->Payload.PacketID.SpacecraftID,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.ProcessorID, EVS_PktPtr->Payload.PacketID.AppName,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.EventID, EVS_PktPtr->Payload.Message);
   	/* Send string event out port #1 */
   	EVS_OutputPort1(PortMessage);
   	}
   	if (((CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort & CFE_EVS_PORT2_BIT) >> 1) == true)
   	{
   	/* Copy event message to string format */
   	snprintf(PortMessage, CFE_EVS_MAX_PORT_MSG_LENGTH, "EVS Port2 %u/%u/%s %u: %s",
   	(unsigned int)EVS_PktPtr->Payload.PacketID.SpacecraftID,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.ProcessorID, EVS_PktPtr->Payload.PacketID.AppName,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.EventID, EVS_PktPtr->Payload.Message);
   	/* Send string event out port #2 */
   	EVS_OutputPort2(PortMessage);
   	}
   	if (((CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort & CFE_EVS_PORT3_BIT) >> 2) == true)
   	{
   	/* Copy event message to string format */
   	snprintf(PortMessage, CFE_EVS_MAX_PORT_MSG_LENGTH, "EVS Port3 %u/%u/%s %u: %s",
   	(unsigned int)EVS_PktPtr->Payload.PacketID.SpacecraftID,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.ProcessorID, EVS_PktPtr->Payload.PacketID.AppName,
   	(unsigned int)EVS_PktPtr->Payload.PacketID.EventID, EVS_PktPtr->Payload.Message);
   	/* Send string event out port #3 */
   	EVS_OutputPort3(PortMessage);
   	}
   	if (((CFE_EVS_Global.EVS_TlmPkt.Payload.OutputPort & CFE_EVS_PORT4_BIT) >> 3) == true)

2. When sending to multiple ports, the (essentially) same snprintf() is done for each one and only one digit in the string changes. Would be way more efficient to construct the later half of the string once, and change the port number.

The first issue should be resolved prior to integration with EDS as those macros become part of the EDS. Therefore they should be changeable without breaking the implementation that uses them. The second issue is just a performance/size enhancement.

[skliper]

Imported from trac issue 62. Created by jphickey on 2015-06-04T12:13:01, last modified: 2019-04-03T17:41:37

[skliper]

Trac comment by jhageman on 2019-04-03 17:41:37:

CCB 4/3/19: See also #94, coordinate solution with implementation of a PSP SendMessageToPort.