genevent-new base
authorcompudj <compudj@04897980-b3bd-0310-b5e0-8ef037075253>
Tue, 8 Nov 2005 17:42:01 +0000 (17:42 +0000)
committercompudj <compudj@04897980-b3bd-0310-b5e0-8ef037075253>
Tue, 8 Nov 2005 17:42:01 +0000 (17:42 +0000)
git-svn-id: http://ltt.polymtl.ca/svn@1308 04897980-b3bd-0310-b5e0-8ef037075253

genevent-new/Makefile [new file with mode: 0644]
genevent-new/README [new file with mode: 0644]
genevent-new/crc32.tab [new file with mode: 0644]
genevent-new/parser.c [new file with mode: 0644]
genevent-new/parser.h [new file with mode: 0644]

diff --git a/genevent-new/Makefile b/genevent-new/Makefile
new file mode 100644 (file)
index 0000000..a1d4c30
--- /dev/null
@@ -0,0 +1,26 @@
+SHELL = /bin/sh
+
+CC = gcc
+
+CFLAGS = -g -o
+
+all: genevent
+
+install: genevent
+       cp -f genevent /usr/local/bin
+
+genevent:genevent.o parser.o 
+       $(CC) $(CFLAGS) $@ genevent.o parser.o
+
+genevent.o: genevent.c genevent.h parser.h
+       $(CC) -g -c genevent.c
+
+parser.o: parser.c parser.h
+       $(CC) -g -c parser.c
+
+.PHONY: clean
+
+clean:
+       rm -rf *.o *~ *# genevent
+
+
diff --git a/genevent-new/README b/genevent-new/README
new file mode 100644 (file)
index 0000000..5c9fb4e
--- /dev/null
@@ -0,0 +1,39 @@
+
+Mathieu Desnoyers -- November 2005
+
+This is a complete rework of genevent.
+
+The 'genevent' program parses event descriptions and generates 
+the inline functions to record events in the kernel.
+
+There are several files in the directory:
+ genevent.c, genevent.h, crc32.tab, parser.c and parser.h
+
+In fact, crc32.tab, parser.c and parser.h are the same files as  
+those in LTT library.
+
+Here is a brief description of how to use genevent.
+
+make
+make install
+
+
+* Add new events to the kernel with genevent
+
+su -
+cd /usr/local/share/LinuxTraceToolkitViewer/facilities
+cp process.xml yourfacility.xml
+  * edit yourfacility.xml to fit your needs.
+cd /tmp
+/usr/local/bin/genevent /usr/local/share/LinuxTraceToolkitViewer/yourfacility.xml
+cp ltt-facility-yourfacility.h ltt-facility-id-yourfacility.h \
+         /usr/src/linux-2.6.12-rc4-mm2-lttng-0.2/include/linux/ltt
+cp ltt-facility-loader-yourfacility.c ltt-facility-loader-yourfacility.h \
+         /usr/src/linux-2.6.12-rc4-mm2-lttng-0.2/ltt
+  * edit the kernel file you want to instrument
+    - Add #include <linux/ltt/ltt-facility-yourfacility.h> at the beginning
+      of the file.
+    - Add a call to the tracing functions. See their names and parameters in
+      /usr/src/linux-2.6.12-rc4-mm2-lttng-0.2/include/linux/ltt/ltt-facility-yourfacility.h
+
+
diff --git a/genevent-new/crc32.tab b/genevent-new/crc32.tab
new file mode 100644 (file)
index 0000000..d0174ad
--- /dev/null
@@ -0,0 +1,52 @@
+  0x00000000U, 0x77073096U, 0xee0e612cU, 0x990951baU, 0x076dc419U,
+  0x706af48fU, 0xe963a535U, 0x9e6495a3U, 0x0edb8832U, 0x79dcb8a4U,
+  0xe0d5e91eU, 0x97d2d988U, 0x09b64c2bU, 0x7eb17cbdU, 0xe7b82d07U,
+  0x90bf1d91U, 0x1db71064U, 0x6ab020f2U, 0xf3b97148U, 0x84be41deU,
+  0x1adad47dU, 0x6ddde4ebU, 0xf4d4b551U, 0x83d385c7U, 0x136c9856U,
+  0x646ba8c0U, 0xfd62f97aU, 0x8a65c9ecU, 0x14015c4fU, 0x63066cd9U,
+  0xfa0f3d63U, 0x8d080df5U, 0x3b6e20c8U, 0x4c69105eU, 0xd56041e4U,
+  0xa2677172U, 0x3c03e4d1U, 0x4b04d447U, 0xd20d85fdU, 0xa50ab56bU,
+  0x35b5a8faU, 0x42b2986cU, 0xdbbbc9d6U, 0xacbcf940U, 0x32d86ce3U,
+  0x45df5c75U, 0xdcd60dcfU, 0xabd13d59U, 0x26d930acU, 0x51de003aU,
+  0xc8d75180U, 0xbfd06116U, 0x21b4f4b5U, 0x56b3c423U, 0xcfba9599U,
+  0xb8bda50fU, 0x2802b89eU, 0x5f058808U, 0xc60cd9b2U, 0xb10be924U,
+  0x2f6f7c87U, 0x58684c11U, 0xc1611dabU, 0xb6662d3dU, 0x76dc4190U,
+  0x01db7106U, 0x98d220bcU, 0xefd5102aU, 0x71b18589U, 0x06b6b51fU,
+  0x9fbfe4a5U, 0xe8b8d433U, 0x7807c9a2U, 0x0f00f934U, 0x9609a88eU,
+  0xe10e9818U, 0x7f6a0dbbU, 0x086d3d2dU, 0x91646c97U, 0xe6635c01U,
+  0x6b6b51f4U, 0x1c6c6162U, 0x856530d8U, 0xf262004eU, 0x6c0695edU,
+  0x1b01a57bU, 0x8208f4c1U, 0xf50fc457U, 0x65b0d9c6U, 0x12b7e950U,
+  0x8bbeb8eaU, 0xfcb9887cU, 0x62dd1ddfU, 0x15da2d49U, 0x8cd37cf3U,
+  0xfbd44c65U, 0x4db26158U, 0x3ab551ceU, 0xa3bc0074U, 0xd4bb30e2U,
+  0x4adfa541U, 0x3dd895d7U, 0xa4d1c46dU, 0xd3d6f4fbU, 0x4369e96aU,
+  0x346ed9fcU, 0xad678846U, 0xda60b8d0U, 0x44042d73U, 0x33031de5U,
+  0xaa0a4c5fU, 0xdd0d7cc9U, 0x5005713cU, 0x270241aaU, 0xbe0b1010U,
+  0xc90c2086U, 0x5768b525U, 0x206f85b3U, 0xb966d409U, 0xce61e49fU,
+  0x5edef90eU, 0x29d9c998U, 0xb0d09822U, 0xc7d7a8b4U, 0x59b33d17U,
+  0x2eb40d81U, 0xb7bd5c3bU, 0xc0ba6cadU, 0xedb88320U, 0x9abfb3b6U,
+  0x03b6e20cU, 0x74b1d29aU, 0xead54739U, 0x9dd277afU, 0x04db2615U,
+  0x73dc1683U, 0xe3630b12U, 0x94643b84U, 0x0d6d6a3eU, 0x7a6a5aa8U,
+  0xe40ecf0bU, 0x9309ff9dU, 0x0a00ae27U, 0x7d079eb1U, 0xf00f9344U,
+  0x8708a3d2U, 0x1e01f268U, 0x6906c2feU, 0xf762575dU, 0x806567cbU,
+  0x196c3671U, 0x6e6b06e7U, 0xfed41b76U, 0x89d32be0U, 0x10da7a5aU,
+  0x67dd4accU, 0xf9b9df6fU, 0x8ebeeff9U, 0x17b7be43U, 0x60b08ed5U,
+  0xd6d6a3e8U, 0xa1d1937eU, 0x38d8c2c4U, 0x4fdff252U, 0xd1bb67f1U,
+  0xa6bc5767U, 0x3fb506ddU, 0x48b2364bU, 0xd80d2bdaU, 0xaf0a1b4cU,
+  0x36034af6U, 0x41047a60U, 0xdf60efc3U, 0xa867df55U, 0x316e8eefU,
+  0x4669be79U, 0xcb61b38cU, 0xbc66831aU, 0x256fd2a0U, 0x5268e236U,
+  0xcc0c7795U, 0xbb0b4703U, 0x220216b9U, 0x5505262fU, 0xc5ba3bbeU,
+  0xb2bd0b28U, 0x2bb45a92U, 0x5cb36a04U, 0xc2d7ffa7U, 0xb5d0cf31U,
+  0x2cd99e8bU, 0x5bdeae1dU, 0x9b64c2b0U, 0xec63f226U, 0x756aa39cU,
+  0x026d930aU, 0x9c0906a9U, 0xeb0e363fU, 0x72076785U, 0x05005713U,
+  0x95bf4a82U, 0xe2b87a14U, 0x7bb12baeU, 0x0cb61b38U, 0x92d28e9bU,
+  0xe5d5be0dU, 0x7cdcefb7U, 0x0bdbdf21U, 0x86d3d2d4U, 0xf1d4e242U,
+  0x68ddb3f8U, 0x1fda836eU, 0x81be16cdU, 0xf6b9265bU, 0x6fb077e1U,
+  0x18b74777U, 0x88085ae6U, 0xff0f6a70U, 0x66063bcaU, 0x11010b5cU,
+  0x8f659effU, 0xf862ae69U, 0x616bffd3U, 0x166ccf45U, 0xa00ae278U,
+  0xd70dd2eeU, 0x4e048354U, 0x3903b3c2U, 0xa7672661U, 0xd06016f7U,
+  0x4969474dU, 0x3e6e77dbU, 0xaed16a4aU, 0xd9d65adcU, 0x40df0b66U,
+  0x37d83bf0U, 0xa9bcae53U, 0xdebb9ec5U, 0x47b2cf7fU, 0x30b5ffe9U,
+  0xbdbdf21cU, 0xcabac28aU, 0x53b39330U, 0x24b4a3a6U, 0xbad03605U,
+  0xcdd70693U, 0x54de5729U, 0x23d967bfU, 0xb3667a2eU, 0xc4614ab8U,
+  0x5d681b02U, 0x2a6f2b94U, 0xb40bbe37U, 0xc30c8ea1U, 0x5a05df1bU,
+  0x2d02ef8dU
diff --git a/genevent-new/parser.c b/genevent-new/parser.c
new file mode 100644 (file)
index 0000000..795b3f9
--- /dev/null
@@ -0,0 +1,1407 @@
+/*
+
+parser.c: Generate helper declarations and functions to trace events
+  from an event description file.
+
+Copyright (C) 2002, Xianxiu Yang
+Copyright (C) 2002, Michel Dagenais 
+This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+the Free Software Foundation; version 2 of the License.
+
+This program is distributed in the hope that it will be useful,
+  but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+*/
+
+/* This program reads the ".xml" event definitions input files 
+   and constructs structure for each event.
+   The program uses a very simple tokenizer, called from a hand written
+   recursive descent parser to fill a data structure describing the events.
+   The result is a sequence of events definitions which refer to type
+   definitions.
+
+   A table of named types is maintained to allow refering to types by name
+   when the same type is used at several places. Finally a sequence of
+   all types is maintained to facilitate the freeing of all type 
+   information when the processing of an ".xml" file is finished. */
+
+#include <stdlib.h> 
+#include <string.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <linux/errno.h>  
+#include <assert.h>
+#include <ctype.h>
+
+#include "parser.h"
+
+
+static char *intOutputTypes[] = {
+  "int8_t", "int16_t", "int32_t", "int64_t", "short int", "int", "long int" };
+
+static char *uintOutputTypes[] = {
+  "uint8_t", "uint16_t", "uint32_t", "uint64_t", "unsigned short int", 
+  "unsigned int", "unsigned long int" };
+
+static char *floatOutputTypes[] = {
+  "undef", "undef", "float", "double", "undef", "float", "double" };
+
+
+
+
+/* helper function  */
+void strupper(char *string)
+{
+  char *ptr = string;
+  
+  while(*ptr != '\0') {
+    *ptr = toupper(*ptr);
+               ptr++;
+  }
+}
+
+
+int getSizeindex(int value)
+{ 
+  switch(value) {
+    case 1:
+      return 0;
+    case 2:
+      return 1;
+    case 4:
+      return 2;
+    case 8:
+      return 3;
+    default:
+      printf("Error : unknown value size %d\n", value);
+      exit(-1);
+  }
+}
+
+/*****************************************************************************
+ *Function name
+ *    getSize    : translate from string to integer
+ *Input params 
+ *    in         : input file handle
+ *Return values  
+ *    size                           
+ *****************************************************************************/
+
+int getSize(parse_file_t *in)
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type == NUMBER) {
+    if(strcmp(token,"1") == 0) return 0;
+    else if(strcmp(token,"2") == 0) return 1;
+    else if(strcmp(token,"4") == 0) return 2;
+    else if(strcmp(token,"8") == 0) return 3;
+  }
+  else if(in->type == NAME) {
+    if(strcmp(token,"short") == 0) return 4;
+    else if(strcmp(token,"medium") == 0) return 5;
+    else if(strcmp(token,"long") == 0) return 6;
+  }
+  in->error(in,"incorrect size specification");
+  return -1;
+}
+
+/*****************************************************************************
+ *Function name
+ *    error_callback  : print out error info
+ *Input params
+ *    in              : input file handle
+ *    msg             : message to be printed                  
+ ****************************************************************************/
+
+void error_callback(parse_file_t *in, char *msg)
+{
+  if(in)
+    printf("Error in file %s, line %d: %s\n", in->name, in->lineno, msg);
+  else
+    printf("%s\n",msg);
+  assert(0);
+  exit(1);
+}
+
+/*****************************************************************************
+ *Function name
+ *    memAlloc  : allocate memory                    
+ *Input params 
+ *    size      : required memory size               
+ *return value 
+ *    void *    : pointer to allocate memory or NULL 
+ ****************************************************************************/
+
+void * memAlloc(int size)
+{
+  void * addr;
+  if(size == 0) return NULL;
+  addr = malloc(size);
+  if(!addr){
+    printf("Failed to allocate memory");    
+    exit(1);
+  }
+  return addr;   
+}
+
+/*****************************************************************************
+ *Function name
+ *    allocAndCopy : allocate memory and initialize it  
+ *Input params 
+ *    str          : string to be put in memory         
+ *return value 
+ *    char *       : pointer to allocate memory or NULL
+ ****************************************************************************/
+
+char *allocAndCopy(char *str)
+{
+  char * addr;
+  if(str == NULL) return NULL;
+  addr = (char *)memAlloc(strlen(str)+1);
+  strcpy(addr,str);
+  return addr;
+}
+
+/**************************************************************************
+ * Function :
+ *    getTypeAttributes
+ * Description :
+ *    Read the attribute from the input file.
+ *
+ * Parameters :
+ *    in , input file handle.
+ *    t , the type descriptor to fill.
+ *
+ **************************************************************************/
+
+void getTypeAttributes(parse_file_t *in, type_descriptor_t *t)
+{
+  char * token;
+
+  t->fmt = NULL;
+  t->size = -1;
+  t->alignment = 0;
+  
+  while(1) {
+    token = getToken(in); 
+    if(strcmp("/",token) == 0 || strcmp(">",token) == 0){
+      ungetToken(in);
+      break;
+    }
+    
+    if(!strcmp("format",token)) {
+      getEqual(in);
+      t->fmt = allocAndCopy(getQuotedString(in));
+    //} else if(!strcmp("name",token)) {
+     // getEqual(in);
+     // car = seekNextChar(in);
+     // if(car == EOF) in->error(in,"name was expected");
+     // else if(car == '\"') t->type_name = allocAndCopy(getQuotedString(in));
+     // else t->type_name = allocAndCopy(getName(in));
+    } else if(!strcmp("size",token)) {
+      getEqual(in);
+      t->size = getSize(in);
+    } else if(!strcmp("align",token)) {
+      getEqual(in);
+      t->alignment = getNumber(in);
+    }
+  }
+}
+
+/**************************************************************************
+ * Function :
+ *    getEventAttributes
+ * Description :
+ *    Read the attribute from the input file.
+ *
+ * Parameters :
+ *    in , input file handle.
+ *    ev , the event to fill.
+ *
+ **************************************************************************/
+
+void getEventAttributes(parse_file_t *in, event_t *ev)
+{
+  char * token;
+  char car;
+  
+  ev->name = NULL;
+  ev->per_trace = 0;
+  ev->per_tracefile = 0;
+
+  while(1) {
+    token = getToken(in); 
+    if(strcmp("/",token) == 0 || strcmp(">",token) == 0){
+      ungetToken(in);
+      break;
+    }
+
+    if(!strcmp("name",token)) {
+      getEqual(in);
+      car = seekNextChar(in);
+      if(car == EOF) in->error(in,"name was expected");
+      else if(car == '\"') ev->name = allocAndCopy(getQuotedString(in));
+      else ev->name = allocAndCopy(getName(in));
+    } else if(!strcmp("per_trace", token)) {
+      ev->per_trace = 1;
+    } else if(!strcmp("per_tracefile", token)) {
+      ev->per_tracefile = 1;
+    }
+
+  }
+}
+
+/**************************************************************************
+ * Function :
+ *    getFacilityAttributes
+ * Description :
+ *    Read the attribute from the input file.
+ *
+ * Parameters :
+ *    in , input file handle.
+ *    fac , the facility to fill.
+ *
+ **************************************************************************/
+
+void getFacilityAttributes(parse_file_t *in, facility_t *fac)
+{
+  char * token;
+  char car;
+  
+  fac->name = NULL;
+
+  while(1) {
+    token = getToken(in); 
+    if(strcmp("/",token) == 0 || strcmp(">",token) == 0){
+      ungetToken(in);
+      break;
+    }
+
+    if(!strcmp("name",token)) {
+      getEqual(in);
+      car = seekNextChar(in);
+      if(car == EOF) in->error(in,"name was expected");
+      else if(car == '\"') fac->name = allocAndCopy(getQuotedString(in));
+      else fac->name = allocAndCopy(getName(in));
+    }
+  }
+}
+
+/**************************************************************************
+ * Function :
+ *    getFieldAttributes
+ * Description :
+ *    Read the attribute from the input file.
+ *
+ * Parameters :
+ *    in , input file handle.
+ *    f , the field to fill.
+ *
+ **************************************************************************/
+
+void getFieldAttributes(parse_file_t *in, field_t *f)
+{
+  char * token;
+  char car;
+
+  f->name = NULL;
+
+  while(1) {
+    token = getToken(in); 
+    if(strcmp("/",token) == 0 || strcmp(">",token) == 0){
+      ungetToken(in);
+      break;
+    }
+
+    if(!strcmp("name",token)) {
+      getEqual(in);
+      car = seekNextChar(in);
+      if(car == EOF) in->error(in,"name was expected");
+      else if(car == '\"') f->name = allocAndCopy(getQuotedString(in));
+      else f->name = allocAndCopy(getName(in));
+    }
+  }
+}
+
+char *getNameAttribute(parse_file_t *in)
+{
+  char * token;
+  char *name = NULL;
+  char car;
+  
+  while(1) {
+    token = getToken(in); 
+    if(strcmp("/",token) == 0 || strcmp(">",token) == 0){
+      ungetToken(in);
+      break;
+    }
+
+    if(!strcmp("name",token)) {
+      getEqual(in);
+      car = seekNextChar(in);
+      if(car == EOF) in->error(in,"name was expected");
+      else if(car == '\"') name = allocAndCopy(getQuotedString(in));
+      else name = allocAndCopy(getName(in));
+    }
+  }
+  if(name == NULL) in->error(in, "Name was expected");
+  return name;
+  
+}
+
+
+
+//for <label name=label_name value=n format="..."/>, value is an option
+char * getValueStrAttribute(parse_file_t *in)
+{
+  char * token;
+
+  token = getToken(in); 
+  if(strcmp("/",token) == 0){
+    ungetToken(in);
+    return NULL;
+  }
+  
+  if(strcmp("value",token))in->error(in,"value was expected");
+  getEqual(in);
+  token = getToken(in);
+  if(in->type != NUMBER) in->error(in,"number was expected");
+  return token;  
+}
+
+char * getDescription(parse_file_t *in)
+{
+  long int pos;
+  char * token, car, *str;
+
+  pos = ftell(in->fp);
+
+  getLAnglebracket(in);
+  token = getName(in);
+  if(strcmp("description",token)){
+    fseek(in->fp, pos, SEEK_SET);
+    return NULL;
+  }
+  
+  getRAnglebracket(in);
+
+  pos = 0;
+  while((car = getc(in->fp)) != EOF) {
+    if(car == '<') break;
+    if(car == '\0') continue;
+    in->buffer[pos] = car;
+    pos++;
+  }
+  if(car == EOF)in->error(in,"not a valid description");
+  in->buffer[pos] = '\0';
+
+  str = allocAndCopy(in->buffer);
+
+  getForwardslash(in);
+  token = getName(in);
+  if(strcmp("description", token))in->error(in,"not a valid description");
+  getRAnglebracket(in);
+
+  return str;
+}
+
+/*****************************************************************************
+ *Function name
+ *    parseFacility : generate event list  
+ *Input params 
+ *    in            : input file handle          
+ *    fac           : empty facility
+ *Output params
+ *    fac           : facility filled with event list
+ ****************************************************************************/
+
+void parseFacility(parse_file_t *in, facility_t * fac)
+{
+  char * token;
+  event_t *ev;
+  
+  getFacilityAttributes(in, fac);
+  if(fac->name == NULL) in->error(in, "Attribute not named");
+
+  fac->capname = allocAndCopy(fac->name);
+       strupper(fac->capname);
+  getRAnglebracket(in);    
+  
+  fac->description = getDescription(in);
+  
+  while(1){
+    getLAnglebracket(in);    
+
+    token = getToken(in);
+    if(in->type == ENDFILE)
+      in->error(in,"the definition of the facility is not finished");
+
+    if(strcmp("event",token) == 0){
+      ev = (event_t*) memAlloc(sizeof(event_t));
+      sequence_push(&(fac->events),ev);
+      parseEvent(in,ev, &(fac->unnamed_types), &(fac->named_types));    
+    }else if(strcmp("type",token) == 0){
+      parseTypeDefinition(in, &(fac->unnamed_types), &(fac->named_types));
+    }else if(in->type == FORWARDSLASH){
+      break;
+    }else in->error(in,"event or type token expected\n");
+  }
+
+  token = getName(in);
+  if(strcmp("facility",token)) in->error(in,"not the end of the facility");
+  getRAnglebracket(in); //</facility>
+}
+
+/*****************************************************************************
+ *Function name
+ *    parseEvent    : generate event from event definition 
+ *Input params 
+ *    in            : input file handle          
+ *    ev            : new event                              
+ *    unnamed_types : array of unamed types
+ *    named_types   : array of named types
+ *Output params    
+ *    ev            : new event (parameters are passed to it)   
+ ****************************************************************************/
+
+void parseEvent(parse_file_t *in, event_t * ev, sequence_t * unnamed_types, 
+               table_t * named_types) 
+{
+  char *token;
+
+  //<event name=eventtype_name>
+  getEventAttributes(in, ev);
+  if(ev->name == NULL) in->error(in, "Event not named");
+  getRAnglebracket(in);  
+
+  //<description>...</descriptio>
+  ev->description = getDescription(in); 
+  
+  //event can have STRUCT, TYPEREF or NOTHING
+  getLAnglebracket(in);
+
+  token = getToken(in);
+  if(in->type == FORWARDSLASH){ //</event> NOTHING
+    ev->type = NULL;
+  }else if(in->type == NAME){
+    if(strcmp("struct",token)==0 || strcmp("typeref",token)==0){
+      ungetToken(in);
+      ev->type = parseType(in,NULL, unnamed_types, named_types);
+      if(ev->type->type != STRUCT && ev->type->type != NONE) 
+       in->error(in,"type must be a struct");     
+    }else in->error(in, "not a valid type");
+
+    getLAnglebracket(in);
+    getForwardslash(in);    
+  }else in->error(in,"not a struct type");
+
+  token = getName(in);
+  if(strcmp("event",token))in->error(in,"not an event definition");
+  getRAnglebracket(in);  //</event>
+}
+
+/*****************************************************************************
+ *Function name
+ *    parseField    : get field infomation from buffer 
+ *Input params 
+ *    in            : input file handle
+ *    t             : type descriptor
+ *    unnamed_types : array of unamed types
+ *    named_types   : array of named types
+ ****************************************************************************/
+
+void parseFields(parse_file_t *in, type_descriptor_t *t,
+    sequence_t * unnamed_types,
+               table_t * named_types) 
+{
+  char * token;
+  field_t *f;
+
+  f = (field_t *)memAlloc(sizeof(field_t));
+  sequence_push(&(t->fields),f);
+
+  //<field name=field_name> <description> <type> </field>
+  getFieldAttributes(in, f);
+  if(f->name == NULL) in->error(in, "Field not named");
+  getRAnglebracket(in);
+
+  f->description = getDescription(in);
+
+  //<int size=...>
+  getLAnglebracket(in);
+  f->type = parseType(in,NULL, unnamed_types, named_types);
+
+  getLAnglebracket(in);
+  getForwardslash(in);
+  token = getName(in);
+  if(strcmp("field",token))in->error(in,"not a valid field definition");
+  getRAnglebracket(in); //</field>
+}
+
+
+/*****************************************************************************
+ *Function name
+ *    parseType      : get type information, type can be : 
+ *                     Primitive:
+ *                        int(size,fmt); uint(size,fmt); float(size,fmt); 
+ *                        string(fmt); enum(size,fmt,(label1,label2...))
+ *                     Compound:
+ *                        array(arraySize, type); sequence(lengthSize,type)
+ *                       struct(field(name,type,description)...)
+ *                     type name:
+ *                        type(name,type)
+ *Input params 
+ *    in               : input file handle
+ *    inType           : a type descriptor          
+ *    unnamed_types    : array of unamed types
+ *    named_types      : array of named types
+ *Return values  
+ *    type_descriptor* : a type descriptor             
+ ****************************************************************************/
+
+type_descriptor_t *parseType(parse_file_t *in, type_descriptor_t *inType, 
+                          sequence_t * unnamed_types, table_t * named_types) 
+{
+  char *token;
+  type_descriptor_t *t;
+
+  if(inType == NULL) {
+    t = (type_descriptor_t *) memAlloc(sizeof(type_descriptor_t));
+    t->type_name = NULL;
+    t->type = NONE;
+    t->fmt = NULL;
+    sequence_push(unnamed_types,t);
+  }
+  else t = inType;
+
+  token = getName(in);
+
+  if(strcmp(token,"struct") == 0) {
+    t->type = STRUCT;
+    getTypeAttributes(in, t);
+    getRAnglebracket(in); //<struct>
+    getLAnglebracket(in); //<field name=..>
+    token = getToken(in);
+    sequence_init(&(t->fields));
+    while(strcmp("field",token) == 0){
+      parseFields(in,t, unnamed_types, named_types);
+      
+      //next field
+      getLAnglebracket(in);
+      token = getToken(in);    
+    }
+    if(strcmp("/",token))in->error(in,"not a valid structure definition");
+    token = getName(in);
+    if(strcmp("struct",token)!=0)
+      in->error(in,"not a valid structure definition");
+    getRAnglebracket(in); //</struct>
+  }
+  else if(strcmp(token,"union") == 0) {
+    t->type = UNION;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "Union has empty size");
+    getRAnglebracket(in); //<union typecodesize=isize>
+
+    getLAnglebracket(in); //<field name=..>
+    token = getToken(in);
+    sequence_init(&(t->fields));
+    while(strcmp("field",token) == 0){
+      parseFields(in,t, unnamed_types, named_types);
+      
+      //next field
+      getLAnglebracket(in);
+      token = getToken(in);    
+    }
+    if(strcmp("/",token))in->error(in,"not a valid union definition");
+    token = getName(in);
+    if(strcmp("union",token)!=0)
+      in->error(in,"not a valid union definition");        
+    getRAnglebracket(in); //</union>
+  }
+  else if(strcmp(token,"array") == 0) {
+    t->type = ARRAY;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "Array has empty size");
+    getRAnglebracket(in); //<array size=n>
+
+    getLAnglebracket(in); //<type struct> 
+    t->nested_type = parseType(in, NULL, unnamed_types, named_types);
+
+    getLAnglebracket(in); //</array>
+    getForwardslash(in);
+    token = getName(in);
+    if(strcmp("array",token))in->error(in,"not a valid array definition");
+    getRAnglebracket(in);  //</array>
+  }
+  else if(strcmp(token,"sequence") == 0) {
+    t->type = SEQUENCE;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "Sequence has empty size");
+    getRAnglebracket(in); //<array lengthsize=isize>
+
+    getLAnglebracket(in); //<type struct> 
+    t->nested_type = parseType(in,NULL, unnamed_types, named_types);
+
+    getLAnglebracket(in); //</sequence>
+    getForwardslash(in);
+    token = getName(in);
+    if(strcmp("sequence",token))in->error(in,"not a valid sequence definition");
+    getRAnglebracket(in); //</sequence>
+  }
+  else if(strcmp(token,"enum") == 0) {
+    char * str, *str1;
+    t->type = ENUM;
+    sequence_init(&(t->labels));
+    sequence_init(&(t->labels_description));
+               t->already_printed = 0;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "Sequence has empty size");
+    getRAnglebracket(in);
+
+    //<label name=label1 value=n/>
+    getLAnglebracket(in);
+    token = getToken(in); //"label" or "/"
+    while(strcmp("label",token) == 0){
+      str   = allocAndCopy(getNameAttribute(in));      
+      token = getValueStrAttribute(in);
+      if(token){
+       str1 = appendString(str,"=");
+       free(str);
+       str = appendString(str1,token);
+       free(str1);
+       sequence_push(&(t->labels),str);
+      }
+      else
+       sequence_push(&(t->labels),str);
+
+      getForwardslash(in);
+      getRAnglebracket(in);
+      
+      //read description if any. May be NULL.
+      str = allocAndCopy(getDescription(in));
+                       sequence_push(&(t->labels_description),str);
+                             
+      //next label definition
+      getLAnglebracket(in);
+      token = getToken(in); //"label" or "/"      
+    }
+    if(strcmp("/",token))in->error(in, "not a valid enum definition");
+    token = getName(in);
+    if(strcmp("enum",token))in->error(in, "not a valid enum definition");
+      getRAnglebracket(in); //</label>
+  }
+  else if(strcmp(token,"int") == 0) {
+    t->type = INT;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "int has empty size");
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"uint") == 0) {
+    t->type = UINT;
+    getTypeAttributes(in, t);
+    if(t->size == -1) in->error(in, "uint has empty size");
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"pointer") == 0) {
+    t->type = POINTER;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"long") == 0) {
+    t->type = LONG;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"ulong") == 0) {
+    t->type = ULONG;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"size_t") == 0) {
+    t->type = SIZE_T;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"ssize_t") == 0) {
+    t->type = SSIZE_T;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"off_t") == 0) {
+    t->type = OFF_T;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"float") == 0) {
+    t->type = FLOAT;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"string") == 0) {
+    t->type = STRING;
+    getTypeAttributes(in, t);
+    getForwardslash(in);
+    getRAnglebracket(in); 
+  }
+  else if(strcmp(token,"typeref") == 0){
+    // Must be a named type
+    if(inType != NULL) 
+      in->error(in,"Named type cannot refer to a named type");
+    else {
+      free(t);
+      sequence_pop(unnamed_types);
+      token = getNameAttribute(in);
+      t = find_named_type(token, named_types);
+      getForwardslash(in);  //<typeref name=type_name/>
+      getRAnglebracket(in);
+      return t;
+    }
+  }else in->error(in,"not a valid type");
+
+  return t;
+}    
+
+/*****************************************************************************
+ *Function name
+ *    find_named_type     : find a named type from hash table 
+ *Input params 
+ *    name                : type name          
+ *    named_types         : array of named types
+ *Return values  
+ *    type_descriptor *   : a type descriptor                       
+ *****************************************************************************/
+
+type_descriptor_t * find_named_type(char *name, table_t * named_types)
+{ 
+  type_descriptor_t *t;
+
+  t = table_find(named_types,name);
+  if(t == NULL) {
+    t = (type_descriptor_t *)memAlloc(sizeof(type_descriptor_t));
+    t->type_name = allocAndCopy(name);
+    t->type = NONE;
+    t->fmt = NULL;
+    table_insert(named_types,t->type_name,t);
+    //    table_insert(named_types,allocAndCopy(name),t);
+  }
+  return t;
+}  
+
+/*****************************************************************************
+ *Function name
+ *    parseTypeDefinition : get type information from type definition 
+ *Input params 
+ *    in                  : input file handle          
+ *    unnamed_types       : array of unamed types
+ *    named_types         : array of named types
+ *****************************************************************************/
+
+void parseTypeDefinition(parse_file_t * in, sequence_t * unnamed_types,
+                        table_t * named_types)
+{
+  char *token;
+  type_descriptor_t *t;
+
+  token = getNameAttribute(in);
+  if(token == NULL) in->error(in, "Type has empty name");
+  t = find_named_type(token, named_types);
+
+  if(t->type != NONE) in->error(in,"redefinition of named type");
+  getRAnglebracket(in); //<type name=type_name>
+  getLAnglebracket(in); //<
+  token = getName(in);
+  //MD ??if(strcmp("struct",token))in->error(in,"not a valid type definition");
+  ungetToken(in);
+  parseType(in,t, unnamed_types, named_types);
+  
+  //</type>
+  getLAnglebracket(in);
+  getForwardslash(in);
+  token = getName(in);
+  if(strcmp("type",token))in->error(in,"not a valid type definition");  
+  getRAnglebracket(in); //</type>
+}
+
+/**************************************************************************
+ * Function :
+ *    getComa, getName, getNumber, getEqual
+ * Description :
+ *    Read a token from the input file, check its type, return it scontent.
+ *
+ * Parameters :
+ *    in , input file handle.
+ *
+ * Return values :
+ *    address of token content.
+ *
+ **************************************************************************/
+
+char *getName(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != NAME) in->error(in,"Name token was expected");
+  return token;
+}
+
+int getNumber(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != NUMBER) in->error(in, "Number token was expected");
+  return atoi(token);
+}
+
+char *getForwardslash(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != FORWARDSLASH) in->error(in, "forward slash token was expected");
+  return token;
+}
+
+char *getLAnglebracket(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != LANGLEBRACKET) in->error(in, "Left angle bracket was expected");
+  return token;
+}
+
+char *getRAnglebracket(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != RANGLEBRACKET) in->error(in, "Right angle bracket was expected");
+  return token;
+}
+
+char *getQuotedString(parse_file_t * in) 
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != QUOTEDSTRING) in->error(in, "quoted string was expected");
+  return token;
+}
+
+char * getEqual(parse_file_t *in)
+{
+  char *token;
+
+  token = getToken(in);
+  if(in->type != EQUAL) in->error(in, "equal was expected");
+  return token;
+}
+
+char seekNextChar(parse_file_t *in)
+{
+  char car;
+  while((car = getc(in->fp)) != EOF) {
+    if(!isspace(car)){
+      ungetc(car,in->fp);
+      return car;
+    }
+  }  
+  return EOF;
+}
+
+/******************************************************************
+ * Function :
+ *    getToken, ungetToken
+ * Description :
+ *    Read a token from the input file and return its type and content.
+ *    Line numbers are accounted for and whitespace/comments are skipped.
+ *
+ * Parameters :
+ *    in, input file handle.
+ *
+ * Return values :
+ *    address of token content.
+ *
+ ******************************************************************/
+
+void ungetToken(parse_file_t * in)
+{
+  in->unget = 1;
+}
+
+char *getToken(parse_file_t * in)
+{
+  FILE *fp = in->fp;
+  char car, car1;
+  int pos = 0, escaped;
+
+  if(in->unget == 1) {
+    in->unget = 0;
+    return in->buffer;
+  }
+
+  /* skip whitespace and comments */
+
+  while((car = getc(fp)) != EOF) {
+    if(car == '/') {
+      car1 = getc(fp); 
+      if(car1 == '*') skipComment(in);
+      else if(car1 == '/') skipEOL(in);
+      else { 
+        car1 = ungetc(car1,fp);
+        break;
+      }
+    }
+    else if(car == '\n') in->lineno++;
+    else if(!isspace(car)) break;
+  }
+
+  switch(car) {
+    case EOF:
+      in->type = ENDFILE;
+      break;
+    case '/':
+      in->type = FORWARDSLASH;
+      in->buffer[pos] = car;
+      pos++;
+      break;
+    case '<':
+      in->type = LANGLEBRACKET;
+      in->buffer[pos] = car;
+      pos++;
+      break;
+    case '>':
+      in->type = RANGLEBRACKET;
+      in->buffer[pos] = car;
+      pos++;
+      break;
+    case '=':
+      in->type = EQUAL;
+      in->buffer[pos] = car;
+      pos++;
+      break;
+    case '"':
+      escaped = 0;
+      while((car = getc(fp)) != EOF && pos < BUFFER_SIZE) {
+        if(car == '\\' && escaped == 0) {
+         in->buffer[pos] = car;
+         pos++;
+          escaped = 1;
+          continue;
+        }
+        if(car == '"' && escaped == 0) break;
+        if(car == '\n' && escaped == 0) {
+          in->error(in, "non escaped newline inside quoted string");
+        }
+        if(car == '\n') in->lineno++;
+        in->buffer[pos] = car;
+        pos++;
+        escaped = 0;
+      }
+      if(car == EOF) in->error(in,"no ending quotemark");
+      if(pos == BUFFER_SIZE) in->error(in, "quoted string token too large");
+      in->type = QUOTEDSTRING;
+      break;
+    default:
+      if(isdigit(car)) {
+        in->buffer[pos] = car;
+        pos++;
+        while((car = getc(fp)) != EOF && pos < BUFFER_SIZE) {
+          if(!isdigit(car)) {
+            ungetc(car,fp);
+            break;
+          }
+          in->buffer[pos] = car;
+          pos++;
+        }
+       if(car == EOF) ungetc(car,fp);
+        if(pos == BUFFER_SIZE) in->error(in, "number token too large");
+        in->type = NUMBER;
+      }    
+      else if(isalpha(car)) {
+        in->buffer[0] = car;
+        pos = 1;
+        while((car = getc(fp)) != EOF && pos < BUFFER_SIZE) {
+          if(!(isalnum(car) || car == '_')) {
+            ungetc(car,fp);
+            break;
+          }
+          in->buffer[pos] = car;
+          pos++;
+        }
+       if(car == EOF) ungetc(car,fp);
+        if(pos == BUFFER_SIZE) in->error(in, "name token too large");
+        in->type = NAME;
+      }
+      else in->error(in, "invalid character, unrecognized token");
+  }
+  in->buffer[pos] = 0;
+  return in->buffer;
+}
+
+void skipComment(parse_file_t * in)
+{
+  char car;
+  while((car = getc(in->fp)) != EOF) {
+    if(car == '\n') in->lineno++;
+    else if(car == '*') {
+      car = getc(in->fp);
+      if(car ==EOF) break;
+      if(car == '/') return;
+      ungetc(car,in->fp);
+    }
+  }
+  if(car == EOF) in->error(in,"comment begining with '/*' has no ending '*/'");
+}
+
+void skipEOL(parse_file_t * in)
+{
+  char car;
+  while((car = getc(in->fp)) != EOF) {
+    if(car == '\n') {
+      ungetc(car,in->fp);
+      break;
+    }
+  }
+  if(car == EOF)ungetc(car, in->fp);
+}
+
+/*****************************************************************************
+ *Function name
+ *    checkNamedTypesImplemented : check if all named types have definition
+ ****************************************************************************/
+
+void checkNamedTypesImplemented(table_t * named_types)
+{
+  type_descriptor_t *t;
+  int pos;
+  char str[256];
+
+  for(pos = 0 ; pos < named_types->values.position; pos++) {
+    t = (type_descriptor_t *) named_types->values.array[pos];
+    if(t->type == NONE){
+      sprintf(str,"named type '%s' has no definition",
+          (char*)named_types->keys.array[pos]);
+      error_callback(NULL,str);   
+    }
+  }
+}
+
+
+/*****************************************************************************
+ *Function name
+ *    generateChecksum  : generate checksum for the facility
+ *Input Params
+ *    facName           : name of facility
+ *Output Params
+ *    checksum          : checksum for the facility
+ ****************************************************************************/
+
+void generateChecksum(char* facName,
+    unsigned long * checksum, sequence_t * events)
+{
+  unsigned long crc ;
+  int pos;
+  event_t * ev;
+  char str[256];
+
+  crc = crc32(facName);
+  for(pos = 0; pos < events->position; pos++){
+    ev = (event_t *)(events->array[pos]);
+    crc = partial_crc32(ev->name,crc);    
+    if(!ev->type) continue; //event without type
+    if(ev->type->type != STRUCT){
+      sprintf(str,"event '%s' has a type other than STRUCT",ev->name);
+      error_callback(NULL, str);
+    }
+    crc = getTypeChecksum(crc, ev->type);
+  }
+  *checksum = crc;
+}
+
+/*****************************************************************************
+ *Function name
+ *   getTypeChecksum    : generate checksum by type info
+ *Input Params
+ *    crc               : checksum generated so far
+ *    type              : type descriptor containing type info
+ *Return value          
+ *    unsigned long     : checksum 
+ *****************************************************************************/
+
+unsigned long getTypeChecksum(unsigned long aCrc, type_descriptor_t * type)
+{
+  unsigned long crc = aCrc;
+  char * str = NULL, buf[16];
+  int flag = 0, pos;
+  field_t * fld;
+
+  switch(type->type){
+    case INT:
+      str = intOutputTypes[type->size];
+      break;
+    case UINT:
+      str = uintOutputTypes[type->size];
+      break;
+    case POINTER:
+      str = allocAndCopy("void *");
+                       flag = 1;
+      break;
+    case LONG:
+      str = allocAndCopy("long");
+                       flag = 1;
+      break;
+    case ULONG:
+      str = allocAndCopy("unsigned long");
+                       flag = 1;
+      break;
+    case SIZE_T:
+      str = allocAndCopy("size_t");
+                       flag = 1;
+      break;
+    case SSIZE_T:
+      str = allocAndCopy("ssize_t");
+                       flag = 1;
+      break;
+    case OFF_T:
+      str = allocAndCopy("off_t");
+                       flag = 1;
+      break;
+    case FLOAT:
+      str = floatOutputTypes[type->size];
+      break;
+    case STRING:
+      str = allocAndCopy("string");
+      flag = 1;
+      break;
+    case ENUM:
+      str = appendString("enum ", uintOutputTypes[type->size]);
+      flag = 1;
+      break;
+    case ARRAY:
+      sprintf(buf,"%d",type->size);
+      str = appendString("array ",buf);
+      flag = 1;
+      break;
+    case SEQUENCE:
+      sprintf(buf,"%d",type->size);
+      str = appendString("sequence ",buf);
+      flag = 1;
+      break;
+    case STRUCT:
+      str = allocAndCopy("struct");
+      flag = 1;
+      break;
+    case UNION:
+      str = allocAndCopy("union");
+      flag = 1;
+      break;
+    default:
+      error_callback(NULL, "named type has no definition");
+      break;
+  }
+
+  crc = partial_crc32(str,crc);
+  if(flag) free(str);
+
+  if(type->fmt) crc = partial_crc32(type->fmt,crc);
+    
+  if(type->type == ARRAY || type->type == SEQUENCE){
+    crc = getTypeChecksum(crc,type->nested_type);
+  }else if(type->type == STRUCT || type->type == UNION){
+    for(pos =0; pos < type->fields.position; pos++){
+      fld = (field_t *) type->fields.array[pos];
+      crc = partial_crc32(fld->name,crc);
+      crc = getTypeChecksum(crc, fld->type);
+    }    
+  }else if(type->type == ENUM){
+    for(pos = 0; pos < type->labels.position; pos++)
+      crc = partial_crc32((char*)type->labels.array[pos],crc);
+  }
+
+  return crc;
+}
+
+
+/* Event type descriptors */
+void freeType(type_descriptor_t * tp)
+{
+  int pos2;
+  field_t *f;
+
+  if(tp->fmt != NULL) free(tp->fmt);
+  if(tp->type == ENUM) {
+    for(pos2 = 0; pos2 < tp->labels.position; pos2++) {
+      free(tp->labels.array[pos2]);
+    }
+    sequence_dispose(&(tp->labels));
+  }
+  if(tp->type == STRUCT) {
+    for(pos2 = 0; pos2 < tp->fields.position; pos2++) {
+      f = (field_t *) tp->fields.array[pos2];
+      free(f->name);
+      free(f->description);
+      free(f);
+    }
+    sequence_dispose(&(tp->fields));
+  }
+}
+
+void freeNamedType(table_t * t)
+{
+  int pos;
+  type_descriptor_t * td;
+
+  for(pos = 0 ; pos < t->keys.position; pos++) {
+    free((char *)t->keys.array[pos]);
+    td = (type_descriptor_t*)t->values.array[pos];
+    freeType(td);
+    free(td);
+  }
+}
+
+void freeTypes(sequence_t *t) 
+{
+  int pos;
+  type_descriptor_t *tp;
+
+  for(pos = 0 ; pos < t->position; pos++) {
+    tp = (type_descriptor_t *)t->array[pos];
+    freeType(tp);
+    free(tp);
+  }
+}
+
+void freeEvents(sequence_t *t) 
+{
+  int pos;
+  event_t *ev;
+
+  for(pos = 0 ; pos < t->position; pos++) {
+    ev = (event_t *) t->array[pos];
+    free(ev->name);
+    free(ev->description);
+    free(ev);
+  }
+
+}
+
+
+/* Extensible array */
+
+void sequence_init(sequence_t *t) 
+{
+  t->size = 10;
+  t->position = 0;
+  t->array = (void **)memAlloc(t->size * sizeof(void *));
+}
+
+void sequence_dispose(sequence_t *t) 
+{
+  t->size = 0;
+  free(t->array);
+  t->array = NULL;
+}
+
+void sequence_push(sequence_t *t, void *elem) 
+{
+  void **tmp;
+
+  if(t->position >= t->size) {
+    tmp = t->array;
+    t->array = (void **)memAlloc(t->size * 2 * sizeof(void *));
+    memcpy(t->array, tmp, t->size * sizeof(void *));
+    t->size = t->size * 2;
+    free(tmp);
+  }
+  t->array[t->position] = elem;
+  t->position++;
+}
+
+void *sequence_pop(sequence_t *t) 
+{
+  return t->array[t->position--];
+}
+
+
+/* Hash table API, implementation is just linear search for now */
+
+void table_init(table_t *t) 
+{
+  sequence_init(&(t->keys));
+  sequence_init(&(t->values));
+}
+
+void table_dispose(table_t *t) 
+{
+  sequence_dispose(&(t->keys));
+  sequence_dispose(&(t->values));
+}
+
+void table_insert(table_t *t, char *key, void *value) 
+{
+  sequence_push(&(t->keys),key);
+  sequence_push(&(t->values),value);
+}
+
+void *table_find(table_t *t, char *key) 
+{
+  int pos;
+  for(pos = 0 ; pos < t->keys.position; pos++) {
+    if(strcmp((char *)key,(char *)t->keys.array[pos]) == 0)
+      return(t->values.array[pos]);
+  }
+  return NULL;
+}
+
+void table_insert_int(table_t *t, int *key, void *value)
+{
+  sequence_push(&(t->keys),key);
+  sequence_push(&(t->values),value);
+}
+
+void *table_find_int(table_t *t, int *key)
+{
+  int pos;
+  for(pos = 0 ; pos < t->keys.position; pos++) {
+    if(*key == *(int *)t->keys.array[pos])
+      return(t->values.array[pos]);
+  }
+  return NULL;
+}
+
+
+/* Concatenate strings */
+
+char *appendString(char *s, char *suffix) 
+{
+  char *tmp;
+  if(suffix == NULL) return s;
+
+  tmp = (char *)memAlloc(strlen(s) + strlen(suffix) + 1);
+  strcpy(tmp,s);
+  strcat(tmp,suffix);  
+  return tmp;
+}
+
+
diff --git a/genevent-new/parser.h b/genevent-new/parser.h
new file mode 100644 (file)
index 0000000..c846697
--- /dev/null
@@ -0,0 +1,209 @@
+#ifndef PARSER_H
+#define PARSER_H
+
+/* Extensible array container */
+
+typedef struct _sequence {
+  int size;
+  int position;
+  void **array;
+} sequence_t;
+
+void sequence_init(sequence_t *t);
+void sequence_dispose(sequence_t *t);
+void sequence_push(sequence_t *t, void *elem);
+void *sequence_pop(sequence_t *t);
+
+
+/* Hash table */
+
+typedef struct _table {
+  sequence_t keys;
+  sequence_t values;
+} table_t;
+
+void table_init(table_t *t);
+void table_dispose(table_t *t);
+void table_insert(table_t *t, char *key, void *value);
+void *table_find(table_t *t, char *key);
+void table_insert_int(table_t *t, int *key, void *value);
+void *table_find_int(table_t *t, int *key);
+
+
+/* Token types */
+
+typedef enum _token_type {
+  ENDFILE,
+  FORWARDSLASH,
+  LANGLEBRACKET,
+  RANGLEBRACKET,
+  EQUAL,
+  QUOTEDSTRING,
+  NUMBER,
+  NAME
+} token_type_t;
+
+
+/* State associated with a file being parsed */
+typedef struct _parse_file {
+  char *name;
+  FILE * fp;
+  int lineno;
+  char *buffer;
+  token_type_t type; 
+  int unget;
+  void (*error) (struct _parse_file *, char *);
+} parse_file_t;
+
+void ungetToken(parse_file_t * in);
+char *getToken(parse_file_t *in);
+char *getForwardslash(parse_file_t *in);
+char *getLAnglebracket(parse_file_t *in);
+char *getRAnglebracket(parse_file_t *in);
+char *getQuotedString(parse_file_t *in);
+char *getName(parse_file_t *in);
+int   getNumber(parse_file_t *in);
+char *getEqual(parse_file_t *in);
+char  seekNextChar(parse_file_t *in);
+
+void skipComment(parse_file_t * in);
+void skipEOL(parse_file_t * in);
+
+/* Some constants */
+
+static const int BUFFER_SIZE = 1024;
+
+
+/* Events data types */
+
+typedef enum _data_type {
+  INT,
+  UINT,
+       POINTER,
+       LONG,
+       ULONG,
+       SIZE_T,
+       SSIZE_T,
+       OFF_T,
+  FLOAT,
+  STRING,
+  ENUM,
+  ARRAY,
+  SEQUENCE,
+  STRUCT,
+  UNION,
+  NONE
+} data_type_t;
+
+/* Event type descriptors */
+
+typedef struct _type_descriptor {
+  char * type_name; //used for named type
+  data_type_t type;
+  char *fmt;
+  int size;
+  sequence_t labels; // for enumeration
+       sequence_t labels_description;
+       int     already_printed;
+  sequence_t fields; // for structure
+  struct _type_descriptor *nested_type; // for array and sequence 
+  int alignment;
+} type_descriptor_t;
+
+
+/* Fields within types */
+
+typedef struct _field{
+  char *name;
+  char *description;
+  type_descriptor_t *type;
+} field_t;
+
+
+/* Events definitions */
+
+typedef struct _event {  
+  char *name;
+  char *description;
+  type_descriptor_t *type; 
+  int  per_trace;   /* Is the event able to be logged to a specific trace ? */
+  int  per_tracefile;  /* Must we log this event in a specific tracefile ? */
+} event_t;
+
+typedef struct _facility {
+  char * name;
+       char * capname;
+  char * description;
+  sequence_t events;
+  sequence_t unnamed_types;
+  table_t named_types;
+} facility_t;
+
+int getSize(parse_file_t *in);
+unsigned long getTypeChecksum(unsigned long aCrc, type_descriptor_t * type);
+
+void parseFacility(parse_file_t *in, facility_t * fac);
+void parseEvent(parse_file_t *in, event_t *ev, sequence_t * unnamed_types,
+    table_t * named_types);
+void parseTypeDefinition(parse_file_t *in,
+    sequence_t * unnamed_types, table_t * named_types);
+type_descriptor_t *parseType(parse_file_t *in,
+    type_descriptor_t *t, sequence_t * unnamed_types, table_t * named_types);
+void parseFields(parse_file_t *in, type_descriptor_t *t,
+    sequence_t * unnamed_types, table_t * named_types);
+void checkNamedTypesImplemented(table_t * namedTypes);
+type_descriptor_t * find_named_type(char *name, table_t * named_types);
+void generateChecksum(char * facName,
+    unsigned long * checksum, sequence_t * events);
+
+
+/* get attributes */
+char * getNameAttribute(parse_file_t *in);
+char * getFormatAttribute(parse_file_t *in);
+int    getSizeAttribute(parse_file_t *in);
+int    getValueAttribute(parse_file_t *in);
+char * getValueStrAttribute(parse_file_t *in);
+
+char * getDescription(parse_file_t *in);
+
+
+/* Dynamic memory allocation and freeing */
+
+void * memAlloc(int size);
+char *allocAndCopy(char * str);
+char *appendString(char *s, char *suffix);
+void freeTypes(sequence_t *t);
+void freeType(type_descriptor_t * td);
+void freeEvents(sequence_t *t);
+void freeNamedType(table_t * t);
+void error_callback(parse_file_t *in, char *msg);
+
+
+//checksum part
+static const unsigned int crctab32[] =
+{
+#include "crc32.tab"
+};
+
+static inline unsigned long
+partial_crc32_one(unsigned char c, unsigned long crc)
+{
+  return crctab32[(crc ^ c) & 0xff] ^ (crc >> 8);
+}
+
+static inline unsigned long
+partial_crc32(const char *s, unsigned long crc)
+{
+  while (*s)
+    crc = partial_crc32_one(*s++, crc);
+  return crc;
+}
+
+static inline unsigned long
+crc32(const char *s)
+{
+  return partial_crc32(s, 0xffffffff) ^ 0xffffffff;
+}
+
+
+#endif // PARSER_H
This page took 0.039766 seconds and 4 git commands to generate.