GLAST / LAT > DAQ and FSW > FSW > Doxygen Index> XFC / V0-3-2 > dfc / rhel6-32


Interface   Data Structures   File List   Data Fields   Globals  

DFC.h File Reference

Diagnostic Filter Code, Interface. More...

#include <EFC/EFC_fate.h>
#include <EDS/EBF_siv.h>

Functions

int DFC_sizeof (int cnt)
 Sizes the diagnostic filtering control structure.
int DFC_cfgSizeof (void)
 Sizes the diagnostic filtering configuration as it is used in the actual filtering progam.
int DFC_cfgCompile (DFC_cfg *cfg, const DFC_DB_prms *prms)
 Implements the transformation of the configuration parameters as specified by the user, cfg, into the internal form, ctlCfg.
int DFC_cfgSchemaCompile (DFC_cfg *cfg, const DFC_DB_schema *schema)
 Implements the transformation of the configuration parameters as specified by the user, cfg, into the internal form, ctlCfg.
int DFC_resultSizeof (const DFC *dfc)
 Returns the size, in bytes, of one result vector.
const DFC_cfgDFC_cfgSet (DFC *dfc, DFC_cfg *cfg)
 Makes the specified internal representation of the configuration parameters the current active one.
int DFC_filter (DFC *dfc, unsigned int pktByte, EBF_pkt *pkt, EBF_siv siv, EDS_fwIxb *ixb, unsigned int vetoes, DFC_result *result)
 Determines the fate of 1 event.
unsigned int DFC_fwObjects (void)
 Returns a bit mask of the EDS_fw objects used by the heavy ion filter.
unsigned int DFC_fwNeeds (void)
 Returns a bit mask of the EDS_fw needs.
const EFC_services * DFC_servicesGet (void)


Detailed Description

Diagnostic Filter Code, Interface.

Author:
JJRussell - russell@slac.stanford.edu
    CVS $Id: DFC.h,v 1.3 2012/01/04 20:36:40 russell Exp $

Function Documentation

int DFC_cfgCompile ( DFC_cfg cfg,
const DFC_DB_prms *  prms 
)

Implements the transformation of the configuration parameters as specified by the user, cfg, into the internal form, ctlCfg.

Return values:
0 Successful
-1 Unrecognized or non-backwardly compatiable user configuration version number.
Parameters:
cfg Filled in with the internal form of the configuration parameters
prms The user form of the configuration parameters

References compileCfgPrms(), _DFC_cfg::prms, and _DFC_cfg::titles.

Referenced by DFC_cfgSchemaCompile().

int DFC_cfgSchemaCompile ( DFC_cfg cfg,
const DFC_DB_schema *  schema 
)

Implements the transformation of the configuration parameters as specified by the user, cfg, into the internal form, ctlCfg.

Return values:
0 Successful
-1 Unrecognized or non-backwardly compatiable user configuration version number.
Parameters:
cfg Filled in with the internal form of the configuration parameters
schema The user form of the configuration schema

References DFC_cfgCompile(), and _DFC_cfg::prms.

const DFC_cfg* DFC_cfgSet ( DFC dfc,
DFC_cfg cfg 
)

Makes the specified internal representation of the configuration parameters the current active one.

Returns:
Pointer to the current control configuration parameters
Parameters:
dfc The diagnostic filter control structure
cfg The internal representation of the configuration constants

References _DFC::cfg.

int DFC_cfgSizeof ( void   ) 

Sizes the diagnostic filtering configuration as it is used in the actual filtering progam.

Returns:
Returns the size of the heavy ion filtering configuration control structure
There are two different breeds of the filtering configuration structure. There are two reasons for this. The more familar is the configuration structure that the user fills in. This structure is meant to be convenient for the user, but not for the program. Some of the information in the user structure is transformed into data that the program will find more accessible.
The second reason is that by having a user configuration and an internal configuration, the two can evolve separately. If more parameters are added to the filter, they will find a home in the internal structure. At some time they may be exposed to the user. The correct course of action is to define a new version of the user configuration block. Since the version number is carried along with the configuration constants, the routine transforming the user configuration to its internal representation can evolve and can be made to handle older versions of the user configuration parameters with some degree of grace, at least giving some hope to backward compatibility.
The correct usage model is
       ctlCfg = malloc (DFC_cfgSizeof ());
       status = DFC_cfgCompile (cfg, usrPrms);
       DFC_cfgSet (dfc, cfg);
Note:
Because the 'set routine just seeds a pointer to the user configuration, in the control structure. the user is free to swap it out with any other pre-prepared configuration.

int DFC_filter ( DFC dfc,
unsigned int  pktBytes,
EBF_pkt *  pkt,
EBF_siv  siv,
EDS_fwIxb *  ixb,
unsigned int  vetoes,
DFC_result result 
)

Determines the fate of 1 event.

Return values:
< 0, rejected
==0,undecided 
> 0, accepted
Parameters:
dfc The diagnostic filter control structure
pktBytes The number of bytes in pkt, includes the packet header
pkt The event data packet
siv The packet's state information vector
ixb The information exchange block
vetoes The active vetoes
result The filter result vector
Overview
This acts as a wrapper of the actual filter. This is were the sampling logic is implemented by the use prescalers. A prescaler is defined by a pair of integers; one giving the current countdown value and the other giving the refresh value. When the countdown value reaches 0, the prescaler is said to expire.
Types of prescaling
There are two kinds of samplings allow.
  1. Pass-thru sampling
  2. Individual veto sampling

Pass-Thru Sampling
Conceptually this is the easiest to understand. When the pass-thru prescaler expires, instead of using the normal veto mask, a 0 is used, effectively causing the filter to evaluate all its cuts but not allowing an cut to veto the event. While easy to understand, the event sample leaked through by this technique will be dominated by events that would normally have been rejected by the very early cuts.
Individual Veto Sampling
This prescaler type allows to ignore the reject decision of each veto bit. While a bit harder to understand than the Pass-Thru Sampling, this technique allows one to tap a predetermined percentage of events failing at each cut. The hope is that one will be a statisitically meaningful sample of events that would have normally failed some of the later cuts in the filter.

References dfcFilter().

unsigned int DFC_fwNeeds ( void   ) 

Returns a bit mask of the EDS_fw needs.

Returns:
A bit mask of the EDS_fw needs
Use
This value should be passed to EDS_fwRegister when registering the filter.

Referenced by dfc_construct().

unsigned int DFC_fwObjects ( void   ) 

Returns a bit mask of the EDS_fw objects used by the heavy ion filter.

Returns:
A bit mask of the EDS_fw needs
Use
This value should be passed to EDS_fwRegister when registering the filter.

Referenced by dfc_construct().

int DFC_resultSizeof ( const DFC dfc  ) 

Returns the size, in bytes, of one result vector.

Return values:
The size, in bytes, of one result vector
Parameters:
dfc The configuration control structure

int DFC_sizeof ( int  cnt  ) 

Sizes the diagnostic filtering control structure.

Returns:
Returns the size of the heavy ion filtering control structure
Parameters:
cnt Number of configurations to accommodate


Generated on Fri Nov 30 11:32:10 2012 by  doxygen 1.5.8