+
/* This file is part of the Linux Trace Toolkit viewer
* Copyright (C) 2003-2004 Xiangxiu Yang
* 2005 Mathieu Desnoyers
*
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License Version 2 as
- * published by the Free Software Foundation;
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License Version 2.1 as published by the Free Software Foundation.
*
- * This program is distributed in the hope that it will be useful,
+ * This library 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.
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser 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.
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 02111-1307, USA.
*/
#ifdef HAVE_CONFIG_H
LttType * lookup_named_type(LttFacility *fac, type_descriptor_t * td);
/* construct directed acyclic graph for types, and tree for fields */
-void construct_types_and_fields(LttFacility * fac, type_descriptor_t * td,
- LttField * fld);
+void construct_fields(LttFacility *fac,
+ LttField *field,
+ field_t *fld);
/* generate the facility according to the events belongin to it */
void generateFacility(LttFacility * f, facility_t * fac,
/* functions to release the memory occupied by a facility */
void freeFacility(LttFacility * facility);
void freeEventtype(LttEventType * evType);
-void freeLttType(LttType ** type);
+void freeLttType(LttType * type);
void freeLttField(LttField * fld);
void freeLttNamedType(LttType * type);
+/*****************************************************************************
+ *Function name
+ * parse_fmt : parses the format string
+ *Input params
+ * fmt : format string specified in the xml file
+ * header : points to the extracted header string
+ * separator : points to the extracted separator string
+ * footer : points to the extracted footer string
+ *
+ * Parses the format string in order to extract the header,
+ * the separator and the footer.
+ * The default fmt string is: { %S, %S }
+ * In this case:
+ * The header is: {
+ * The separator is: ,
+ * The footer
+ *
+ ****************************************************************************/
+
+parse_fmt(char *fmt, char **header, char **separator, char **footer){
+ int i;
+ unsigned int num = 0;
+
+ int header_index = 0;//header index
+ int separator_index = 0;//separator index
+ int footer_index = 0;//footer index
+ int fmt_length = strlen(fmt);
+
+ for(i=0; i<fmt_length; i++)
+ {
+ if(fmt[i]=='%')
+ {
+ num++;
+ if(num>=3)
+ g_error("More than 2 '%%' chars were encountered: %s\n",fmt);
+
+ //Detecting identifier
+ if(fmt[++i]!='S')
+ g_error("Unexpected format: %s\n",fmt);
+
+ if(i+1<strlen(fmt))
+ i = i+1;
+ else
+ g_error("Unexpected format: %s\n",fmt);
+
+ //Need to have at least on char after the %S
+ if(fmt[i]=='%')
+ g_error("Missing separator: %s\n",fmt);
+
+ if(separator_index==0)
+ separator_index = i;//index of separator inside the original string.
+ else if(footer_index==0)//the 'else' gives the priority to set the separator index first.
+ footer_index = i;//no break since an error should be generated if more than 2 '%' were encountered
+ }
+ }
+
+
+ //Create header String
+ num = separator_index-header_index-2;//(-2 due to '%S')
+ if(num<0)
+ g_error("Unexpected format");
+ //*header = malloc(num+1);//(+1 for EOS).
+ *header = g_new(gchar,num+1);//(+1 for EOS).
+ strncpy(*header,fmt,num);
+ (*header)[num] = '\0';//need EOS, not handled by strncopy
+
+ //Create seperator String
+ num =footer_index - separator_index - 2;
+ if(num<0)
+ g_error("Unexpected format");
+ //*separator = malloc(num+1);//+1 for EOS
+ *separator = g_new(gchar, num+1);//+1 for EOS
+ strncpy(*separator,fmt+separator_index*sizeof(char),num);
+ (*separator)[num] = '\0';//need EOS, not handled by strncopy
+
+ //Create footer String
+ num = strlen(fmt)-footer_index;
+ //*footer = malloc(num+1);
+ *footer = g_new(gchar, num+1);
+ strncpy(*footer,fmt+footer_index*sizeof(char),num);
+ (*footer)[num] = '\0';
+}
+
+/*****************************************************************************
+ *Function name
+ * verify_fmt_syntax : verifies the syntax of the format string
+ *Input params
+ * fmt : format string specified in the xml file
+ * fmt_type : points to the detected type in the fmt string
+ *
+ * Verifies the syntax of the format string sepcified in the xml file
+ * It allows fmt strings having the following syntax:
+ * "..header text... %12.20d ...footer text..."
+ * It sets the fmt_type accordingly. In the previous example
+ * fmt_type would point to the character 'd' inside fmt.
+ *
+ *returns 1 on success, and -1 or -2 on error.
+ *
+ ****************************************************************************/
+
+int verify_fmt_syntax(char *fmt, char **fmt_type){
+ int i;
+ int dot = 0;//number of dots encountered.
+ unsigned short counter = 0;
+ *fmt_type=NULL;
+
+
+ for(i=0; i<strlen(fmt); i++)
+ {
+ if(fmt[i]=='%')
+ {
+ if(++counter==2)
+ return -1;//should generate an error.
+
+ i = i+1;//go to next character
+
+ if(fmt[i]=='-' && i<(strlen(fmt)-1))
+ i++;//allow this character after the '%'
+
+ if(fmt[i]=='#' && i<(strlen(fmt)-1))
+ i++;//allow this character
+ }
+
+ if(counter ==1 && !isdigit(fmt[i]) && *fmt_type==NULL)
+ switch (fmt[i]){
+ case 'd':
+ case 'i':
+ case 'u':
+ case 'o':
+ case 'x':
+ case 'X':
+ case 'f':
+ case 'e':
+ case 'E':
+ case 's':
+ case 'p':
+ *fmt_type=&fmt[i];
+ break;
+ //do not return yet. may encounter another '%'
+ case '.':
+ if(++dot==2)
+ return -2;//generate error
+ break;
+ default:
+ return -1;
+ }
+
+ }
+
+ return 1;
+
+}
+
+/*****************************************************************************
+ *Function name
+ * append_ll : appends "ll" to the format string
+ *Input params
+ * fmt : format string specified in the xml file
+ * fmt_type : address of the format char inside fmt
+ *
+ * inserts "ll" just before the fmt_type
+ *
+ ****************************************************************************/
+
+void append_ll (char **fmt, char **fmt_type){
+ char *new_fmt;
+ int i;
+ int num;
+
+
+ //the +2 corresponds the the "ll"; the +1 is the \0
+ new_fmt = g_new(gchar, strlen(*fmt)+2+1);
+
+ num = *fmt_type - *fmt;
+
+ for(i=0; i<num;i++)
+ new_fmt[i] =(*fmt)[i];
+
+ new_fmt[i++] = 'l';
+ new_fmt[i++] = 'l';
+
+
+ while(i-2<strlen(*fmt))
+ {
+ new_fmt[i]=(*fmt)[i-2];
+ i++;
+ }
+
+ *fmt = new_fmt;
+ (*fmt)[i] = '\0';
+
+}
+
+
/*****************************************************************************
*Function name
* ltt_facility_open : open facilities
* t : the trace containing the facilities
* pathname : the path name of the facility
*
+ * Open the facility corresponding to the right checksum.
+ *
*returns 0 on success, 1 on error.
****************************************************************************/
int ltt_facility_open(LttFacility *f, LttTrace * t, gchar * pathname)
{
+ int ret = 0;
gchar *token;
parse_file_t in;
- gsize length;
facility_t * fac;
- unsigned long checksum;
- GError * error = NULL;
+ unsigned int checksum;
gchar buffer[BUFFER_SIZE];
+ gboolean generated = FALSE;
in.buffer = &(buffer[0]);
- in.lineno = 0;
+ in.lineno = 1;
in.error = error_callback;
in.name = pathname;
+ in.unget = 0;
- //in.fd = g_open(in.name, O_RDONLY, 0);
- //if(in.fd < 0 ) {
in.fp = fopen(in.name, "r");
- if(in.fp < 0 ) {
+ if(in.fp == NULL) {
g_warning("cannot open facility description file %s",
in.name);
- return 1;
+ ret = 1;
+ goto open_error;
}
- //in.channel = g_io_channel_unix_new(in.fd);
- //in.pos = 0;
-
while(1){
token = getToken(&in);
if(in.type == ENDFILE) break;
+
+ if(g_ascii_strcasecmp(token, "<")) in.error(&in,"not a facility file");
+ token = getName(&in);
+ if(g_ascii_strcasecmp(token, "?")) in.error(&in,"not a facility file");
+ token = getName(&in);
+ if(g_ascii_strcasecmp(token, "xml")) in.error(&in,"not a facility file");
+ token = getName(&in);
+ if(g_ascii_strcasecmp(token, "version")) in.error(&in,"not a facility file");
+ token = getName(&in);
+ if(g_ascii_strcasecmp(token, "=")) in.error(&in,"not a facility file");
+ token = getQuotedString(&in);
+ if(g_ascii_strcasecmp(token, "1.0")) in.error(&in,"not a facility file");
+ token = getName(&in);
+ if(g_ascii_strcasecmp(token, "?")) in.error(&in,"not a facility file");
+ token = getToken(&in);
+ if(g_ascii_strcasecmp(token, ">")) in.error(&in,"not a facility file");
+
+ token = getToken(&in);
if(g_ascii_strcasecmp(token, "<")) in.error(&in,"not a facility file");
token = getName(&in);
checkNamedTypesImplemented(&fac->named_types);
generateChecksum(fac->name, &checksum, &fac->events);
-
- generateFacility(f, fac, checksum);
+ // FIXME if(checksum == f->checksum) {
+ generateFacility(f, fac, checksum);
+ generated = TRUE;
+ //}
+ if (checksum != f->checksum)
+ g_warning("Facility checksum mismatch for facility %s : kernel 0x%X vs "
+ "XML 0x%X\n", fac->name, f->checksum, checksum);
g_free(fac->name);
+ free(fac->capname);
g_free(fac->description);
freeEvents(&fac->events);
sequence_dispose(&fac->events);
freeTypes(&fac->unnamed_types);
sequence_dispose(&fac->unnamed_types);
g_free(fac);
+ if(generated) break; /* use the first good match */
}
else {
g_warning("facility token was expected in file %s", in.name);
+ ret = 1;
goto parse_error;
}
}
-
+
parse_error:
- //g_io_channel_shutdown(in.channel, FALSE, &error); /* No flush */
- //if(error != NULL) {
fclose(in.fp);
- // g_warning("Can not close file: \n%s\n", error->message);
- // g_error_free(error);
- //}
+open_error:
+
+ if(!generated) {
+ g_warning("Cannot find facility %s, checksum 0x%X",
+ g_quark_to_string(f->name), f->checksum);
+ ret = 1;
+ }
- //g_close(in.fd);
+ return ret;
}
{
char * facilityName = fac->name;
sequence_t * events = &fac->events;
- int i;
- //LttEventType * evType;
- LttEventType * event_type;
- LttField * field;
+ unsigned int i, j;
LttType * type;
+ table_t *named_types = &fac->named_types;
g_assert(f->name == g_quark_from_string(facilityName));
- g_assert(f->checksum == checksum);
+ //g_assert(f->checksum == checksum);
//f->event_number = events->position;
f->events = g_array_set_size(f->events, events->position);
g_datalist_init(&f->events_by_name);
- g_datalist_init(&f->named_types);
-
- //f->named_types_number = fac->named_types.keys.position;
- //f->named_types = g_array_sized_new (FALSE, TRUE, sizeof(LttType),
- // fac->named_types.keys.position);
- //f->named_types = g_new(LttType*, fac->named_types.keys.position);
- //f->named_types = g_array_set_size(f->named_types,
- // fac->named_types.keys.position);
-
- //for each event, construct field tree and type graph
+ // g_datalist_init(&f->named_types);
+#if 0
+ /* The first day, he created the named types */
+
+ for(i=0; i<named_types->keys.position; i++) {
+ GQuark name = g_quark_from_string((char*)named_types->keys.array[i]);
+ type_descriptor_t *td = (type_descriptor_t*)named_types->values.array[i];
+
+ /* Create the type */
+ type = g_new(LttType,1);
+ type->type_name = name;
+ type->type_class = td->type;
+ if(td->fmt) type->fmt = g_strdup(td->fmt);
+ else type->fmt = NULL;
+ type->size = td->size;
+ type->enum_strings = NULL;
+ type->element_type = NULL;
+ type->element_number = 0;
+
+ construct_types_and_fields(type, td, NULL, NULL, ...);
+
+ g_datalist_id_set_data_full(&fac->named_types, name,
+ type, (GDestroyNotify)freeLttNamedType);
+
+ }
+#endif //0
+ /* The second day, he created the event fields and types */
+ //for each event, construct field and type acyclic graph
for(i=0;i<events->position;i++){
- event_type = &g_array_index(f->events, LttEventType, i);
- //evType = g_new(LttEventType,1);
- //f->events[i] = evType;
+ event_t *parser_event = (event_t*)events->array[i];
+ LttEventType *event_type = &g_array_index(f->events, LttEventType, i);
event_type->name =
- g_quark_from_string(((event_t*)(events->array[i]))->name);
+ g_quark_from_string(parser_event->name);
g_datalist_id_set_data(&f->events_by_name, event_type->name,
event_type);
event_type->description =
- g_strdup(((event_t*)(events->array[i]))->description);
+ g_strdup(parser_event->description);
- field = g_new(LttField, 1);
- event_type->root_field = field;
- event_type->facility = f;
event_type->index = i;
+ event_type->facility = f;
- if(((event_t*)(events->array[i]))->type != NULL){
- // field->field_pos = 0;
- type = lookup_named_type(f,((event_t*)(events->array[i]))->type);
- field->field_type = type;
- field->offset_root = 0;
- field->fixed_root = FIELD_UNKNOWN;
- field->offset_parent = 0;
- field->fixed_parent = FIELD_UNKNOWN;
- // field->base_address = NULL;
- field->field_size = 0;
- field->fixed_size = FIELD_UNKNOWN;
- field->parent = NULL;
- field->child = NULL;
- field->current_element = 0;
-
- //construct field tree and type graph
- construct_types_and_fields(f,((event_t*)(events->array[i]))->type,field);
- }else{
- event_type->root_field = NULL;
- g_free(field);
+ event_type->has_compact_data = parser_event->compact_data;
+
+ event_type->fields = g_array_sized_new(FALSE, TRUE,
+ sizeof(LttField), parser_event->fields.position);
+ event_type->fields =
+ g_array_set_size(event_type->fields, parser_event->fields.position);
+ g_datalist_init(&event_type->fields_by_name);
+
+ for(j=0; j<parser_event->fields.position; j++) {
+ LttField *field = &g_array_index(event_type->fields, LttField, j);
+ field_t *parser_field = (field_t*)parser_event->fields.array[j];
+
+ construct_fields(f, field, parser_field);
+ g_datalist_id_set_data(&event_type->fields_by_name,
+ field->name,
+ field);
}
- }
+ }
+
+ /* What about 2 days weeks ? */
}
* internal recursion function
*Input params
* fac : facility struct
- * td : type descriptor
- * root_field : root field of the event
+ * field : destination lttv field
+ * fld : source parser field
****************************************************************************/
+//DONE
+//make the change for arrays and sequences
+//no more root field. -> change this for an array of fields.
+// Compute the field size here.
+// Flag fields as "VARIABLE OFFSET" or "FIXED OFFSET" : as soon as
+// a field with a variable size is found, all the following fields must
+// be flagged with "VARIABLE OFFSET", this will be done by the offset
+// precomputation.
+
+
+void construct_fields(LttFacility *fac,
+ LttField *field,
+ field_t *fld)
+{
+ char *fmt_type;//gaby
+ guint len;
+ type_descriptor_t *td;
+ LttType *type;
+
+ if(fld->name)
+ field->name = g_quark_from_string(fld->name);
+ else
+ fld->name = 0;
+
+ if(fld->description) {
+ len = strlen(fld->description);
+ field->description = g_new(gchar, len+1);
+ strcpy(field->description, fld->description);
+ }
+ field->dynamic_offsets = NULL;
+ type = &field->field_type;
+ td = fld->type;
+
+ type->enum_map = NULL;
+ type->fields = NULL;
+ type->fields_by_name = NULL;
+ type->network = td->network;
+
+ switch(td->type) {
+ case INT_FIXED:
+ type->type_class = LTT_INT_FIXED;
+ type->size = td->size;
+ break;
+ case UINT_FIXED:
+ type->type_class = LTT_UINT_FIXED;
+ type->size = td->size;
+ break;
+ case POINTER:
+ type->type_class = LTT_POINTER;
+ type->size = fac->pointer_size;
+ break;
+ case CHAR:
+ type->type_class = LTT_CHAR;
+ type->size = td->size;
+ break;
+ case UCHAR:
+ type->type_class = LTT_UCHAR;
+ type->size = td->size;
+ g_assert(type->size != 0);
+ break;
+ case SHORT:
+ type->type_class = LTT_SHORT;
+ type->size = td->size;
+ break;
+ case USHORT:
+ type->type_class = LTT_USHORT;
+ type->size = td->size;
+ break;
+ case INT:
+ type->type_class = LTT_INT;
+ type->size = fac->int_size;
+ break;
+ case UINT:
+ type->type_class = LTT_UINT;
+ type->size = fac->int_size;
+ g_assert(type->size != 0);
+ break;
+ case LONG:
+ type->type_class = LTT_LONG;
+ type->size = fac->long_size;
+ break;
+ case ULONG:
+ type->type_class = LTT_ULONG;
+ type->size = fac->long_size;
+ break;
+ case SIZE_T:
+ type->type_class = LTT_SIZE_T;
+ type->size = fac->size_t_size;
+ break;
+ case SSIZE_T:
+ type->type_class = LTT_SSIZE_T;
+ type->size = fac->size_t_size;
+ break;
+ case OFF_T:
+ type->type_class = LTT_OFF_T;
+ type->size = fac->size_t_size;
+ break;
+ case FLOAT:
+ type->type_class = LTT_FLOAT;
+ type->size = td->size;
+ break;
+ case STRING:
+ type->type_class = LTT_STRING;
+ type->size = 0;
+ break;
+ case ENUM:
+ type->type_class = LTT_ENUM;
+ type->size = fac->int_size;
+ {
+ guint i;
+ type->enum_map = g_hash_table_new(g_direct_hash, g_direct_equal);
+ type->lowest_value = G_MAXINT32;
+ type->highest_value = G_MININT32;
+ for(i=0; i<td->labels.position; i++) {
+ GQuark value = g_quark_from_string((char*)td->labels.array[i]);
+ gint key = *(int*)td->labels_values.array[i];
+ g_hash_table_insert(type->enum_map, (gpointer)key, (gpointer)value);
+ type->highest_value = max(key, type->highest_value);
+ type->lowest_value = min(key, type->lowest_value);
+ }
+ }
+ g_assert(type->size != 0);
+ break;
+ case ARRAY:
+ type->type_class = LTT_ARRAY;
+ type->size = td->size;
+ type->fields = g_array_sized_new(FALSE, TRUE, sizeof(LttField),
+ td->fields.position);
+ type->fields = g_array_set_size(type->fields, td->fields.position);
+ {
+ guint i;
+
+ for(i=0; i<td->fields.position; i++) {
+ field_t *schild = (field_t*)td->fields.array[i];
+ LttField *dchild = &g_array_index(type->fields, LttField, i);
+
+ construct_fields(fac, dchild, schild);
+ }
+ }
+ break;
+ case SEQUENCE:
+ type->type_class = LTT_SEQUENCE;
+ type->size = 0;
+ type->fields = g_array_sized_new(FALSE, TRUE, sizeof(LttField),
+ td->fields.position);
+ type->fields = g_array_set_size(type->fields, td->fields.position);
+ {
+ guint i;
+
+ for(i=0; i<td->fields.position; i++) {
+ field_t *schild = (field_t*)td->fields.array[i];
+ LttField *dchild = &g_array_index(type->fields, LttField, i);
+
+ construct_fields(fac, dchild, schild);
+ }
+ }
+ break;
+ case STRUCT:
+ type->type_class = LTT_STRUCT;
+ type->size = 0; // Size not calculated by the parser.
+ type->fields = g_array_sized_new(FALSE, TRUE, sizeof(LttField),
+ td->fields.position);
+ type->fields = g_array_set_size(type->fields, td->fields.position);
+ g_datalist_init(&type->fields_by_name);
+ {
+ guint i;
+
+ for(i=0; i<td->fields.position; i++) {
+ field_t *schild = (field_t*)td->fields.array[i];
+ LttField *dchild = &g_array_index(type->fields, LttField, i);
+
+ construct_fields(fac, dchild, schild);
+ g_datalist_id_set_data(&type->fields_by_name,
+ dchild->name,
+ dchild);
+ }
+ }
+ break;
+ case UNION:
+ type->type_class = LTT_UNION;
+ type->size = 0; // Size not calculated by the parser.
+ type->fields = g_array_sized_new(FALSE, TRUE, sizeof(LttField),
+ td->fields.position);
+ type->fields = g_array_set_size(type->fields, td->fields.position);
+ g_datalist_init(&type->fields_by_name);
+ {
+ guint i;
+
+ for(i=0; i<td->fields.position; i++) {
+ field_t *schild = (field_t*)td->fields.array[i];
+ LttField *dchild = &g_array_index(type->fields, LttField, i);
+
+ construct_fields(fac, dchild, schild);
+ g_datalist_id_set_data(&type->fields_by_name,
+ dchild->name,
+ dchild);
+ }
+ }
+ break;
+ case NONE:
+ default:
+ g_error("construct_fields : unknown type");
+ }
+
+ field->field_size = type->size;
+
+ /* Put the fields as "variable" offset to root first. Then,
+ * the offset precomputation will only have to set the FIELD_FIXED until
+ * it reaches the first variable length field, then stop.
+ */
+ field->fixed_root = FIELD_VARIABLE;
+
+ if(td->fmt) {
+ len = strlen(td->fmt);
+ type->fmt = g_new(gchar, len+1);
+ strcpy(type->fmt, td->fmt);
+ }
+ //here I should verify syntax based on type.
+ //if type is array or sequence or enum, parse_fmt.
+ //if type is basic, verify syntax, and allow or not the type format.
+ //the code can be integrated in the above code (after testing)
+
+
+
+ switch (type->type_class){
+ case LTT_ARRAY:
+ case LTT_UNION:
+ case LTT_STRUCT:
+ case LTT_SEQUENCE:
+ if(type->fmt==NULL)
+ {
+ //Assign a default format for these complex types
+ //type->fmt = malloc(11*sizeof(char));
+ type->fmt = g_new(gchar, 11);
+ type->fmt = g_strdup("{ %S, %S }");//set a default value for fmt. can directly set header and footer, but kept this way on purpose.
+ }
+ //Parse the fmt string in order to extract header, separator and footer
+ parse_fmt(type->fmt,&(type->header),&(type->separator),&(type->footer));
+
+ break;
+ case LTT_SHORT:
+ case LTT_INT:
+ case LTT_LONG:
+ case LTT_SSIZE_T:
+ case LTT_INT_FIXED:
+
+ if(type->fmt == NULL)
+ {
+ //Assign a default format string
+ //type->fmt = malloc(5*sizeof(char));
+ type->fmt = g_new(gchar, 5);
+ type->fmt=g_strdup("%lld");
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>0)
+ switch(fmt_type[0]){
+ case 'd':
+ case 'i':
+ case 'x':
+ case 'X':
+ append_ll(&(type->fmt),&fmt_type);//append 'll' to fmt
+ break;
+ default:
+ g_error("Format type '%c' not supported\n",fmt_type[0]);
+ break;
+ }
+ break;
+
+ case LTT_USHORT:
+ case LTT_UINT:
+ case LTT_ULONG:
+ case LTT_SIZE_T:
+ case LTT_OFF_T:
+ case LTT_UINT_FIXED:
+ if(type->fmt == NULL)
+ {
+ //Assign a default format string
+ //type->fmt= malloc(5*sizeof(char));
+ type->fmt = g_new(gchar, 5);
+ type->fmt=g_strdup("%lld");
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>0)
+ switch(fmt_type[0]){
+ case 'd':
+ case 'u':
+ case 'o':
+ case 'x':
+ case 'X':
+ append_ll(&(type->fmt),&fmt_type);
+ break;
+ default:
+ g_error("Format type '%c' not supported\n",fmt_type[0]);
+ }
+ break;
+
+ case LTT_CHAR:
+ case LTT_UCHAR:
+ if(type->fmt == NULL)
+ {
+ //Assign a default format string
+ //type->fmt = malloc(3*sizeof(char));
+ type->fmt = g_new(gchar, 3);
+ type->fmt = g_strdup("%c");
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>1)
+ switch(fmt_type[0]){
+ case 'c':
+ case 'd':
+ case 'u':
+ case 'x':
+ case 'X':
+ case 'o':
+ break;
+ default:
+ g_error("Format type '%c' not supported\n",fmt_type[0]);
+ }
+ break;
+
+ case LTT_FLOAT:
+ if(type->fmt == NULL)
+ {
+ //Assign a default format string
+ //type->fmt = malloc(3*sizeof(char));
+ type->fmt = g_new(gchar, 3);
+ type->fmt = g_strdup("%g");
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>0)
+ switch(fmt_type[0]){
+ case 'f':
+ case 'g':
+ case 'G':
+ case 'e':
+ case 'E':
+ break;
+ default:
+ g_error("Format type '%c' not supported\n",fmt_type[0]);
+ }
+ break;
+
+ case LTT_POINTER:
+ if(type->fmt == NULL)
+ {
+ //Assign a default format string
+ //type->fmt = malloc(7*sizeof(char));
+ type->fmt = g_new(gchar, 7);
+ type->fmt = g_strdup("0x%llx");
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>0)
+ switch(fmt_type[0]){
+ case 'p':
+ //type->fmt = malloc(7*sizeof(char));
+ type->fmt = g_new(gchar, 7);
+ type->fmt = g_strdup("0x%llx");
+ break;
+ case 'x':
+ case 'X':
+ case 'd':
+ append_ll(&(type->fmt),&fmt_type);
+ break;
+ default:
+ g_error("Format type '%c' not supported\n",fmt_type[0]);
+ }
+ break;
+
+ case LTT_STRING:
+ if(type->fmt == NULL)
+ {
+ //type->fmt = malloc(7*sizeof(char));
+ type->fmt = g_new(gchar, 5);
+ type->fmt = g_strdup("\"%s\"");//default value for fmt.
+ break;
+ }
+ else
+ if(verify_fmt_syntax((type->fmt),&fmt_type)>0)
+ switch(fmt_type[0]){
+ case 's':
+ break;
+ default:
+ g_error("Format type '%c' not supported\n", fmt_type[0]);
+ }
+ break;
+
+ default:
+ //missing enum
+ break;
+ }
+
+
+}
+#if 0
void construct_types_and_fields(LttFacility * fac, type_descriptor_t * td,
LttField * fld)
{
- int i, flag;
+ int i;
type_descriptor_t * tmpTd;
switch(td->type) {
- case LTT_ENUM:
- fld->field_type->element_number = td->labels.position;
- fld->field_type->enum_strings = g_new(GQuark,td->labels.position);
- for(i=0;i<td->labels.position;i++){
- fld->field_type->enum_strings[i]
- = g_quark_from_string(((char*)(td->labels.array[i])));
- }
- break;
- case LTT_ARRAY:
- fld->field_type->element_number = (unsigned)td->size;
- case LTT_SEQUENCE:
+ case INT:
+ case UINT:
+ case FLOAT:
+ fld->field_type->size = td->size;
+ break;
+ case POINTER:
+ case LONG:
+ case ULONG:
+ case SIZE_T:
+ case SSIZE_T:
+ case OFF_T:
+ fld->field_type->size = 0;
+ break;
+ case STRING:
+ fld->field_type->size = 0;
+ break;
+ case ENUM:
+ fld->field_type->element_number = td->labels.position;
+ fld->field_type->enum_strings = g_new(GQuark,td->labels.position);
+ for(i=0;i<td->labels.position;i++){
+ fld->field_type->enum_strings[i]
+ = g_quark_from_string(((char*)(td->labels.array[i])));
+ }
+ fld->field_type->size = td->size;
+ break;
+
+ case ARRAY:
+ fld->field_type->element_number = (unsigned)td->size;
+ case SEQUENCE:
fld->field_type->element_type = g_new(LttType*,1);
tmpTd = td->nested_type;
fld->field_type->element_type[0] = lookup_named_type(fac, tmpTd);
fld->child[0]->current_element = 0;
construct_types_and_fields(fac, tmpTd, fld->child[0]);
break;
- case LTT_STRUCT:
- case LTT_UNION:
+
+ case STRUCT:
+ case UNION:
fld->field_type->element_number = td->fields.position;
g_assert(fld->field_type->element_type == NULL);
for(i=0;i<td->fields.position;i++){
tmpTd = ((field_t*)(td->fields.array[i]))->type;
- fld->field_type->element_type[i] = lookup_named_type(fac, tmpTd);
+ fld->field_type->element_type[i] = lookup_named_type(fac, tmpTd);
fld->child[i] = g_new(LttField,1);
// fld->child[i]->field_pos = i;
fld->child[i]->field_type = fld->field_type->element_type[i];
fld->child[i]->field_type->element_name
- = g_quark_from_string(((field_t*)(td->fields.array[i]))->name);
+ = g_quark_from_string(((field_t*)(td->fields.array[i]))->name);
fld->child[i]->offset_root = 0;
fld->child[i]->fixed_root = FIELD_UNKNOWN;
fld->child[i]->current_element = 0;
construct_types_and_fields(fac, tmpTd, fld->child[i]);
}
-
break;
+
default:
g_error("construct_types_and_fields : unknown type");
}
}
-
+#endif //0
#if 0
void construct_types_and_fields(LttFacility * fac, type_descriptor * td,
tmpTd = ((type_fields*)(td->fields.array[i]))->type;
if(flag)
- fld->field_type->element_type[i] = lookup_named_type(fac, tmpTd);
+ fld->field_type->element_type[i] = lookup_named_type(fac, tmpTd);
fld->child[i] = g_new(LttField,1);
fld->child[i]->field_pos = i;
fld->child[i]->field_type = fld->field_type->element_type[i];
if(flag){
- fld->child[i]->field_type->element_name
- = g_quark_from_string(((type_fields*)(td->fields.array[i]))->name);
+ fld->child[i]->field_type->element_name
+ = g_quark_from_string(((type_fields*)(td->fields.array[i]))->name);
}
fld->child[i]->offset_root = -1;
}
#endif //0
+#if 0
/*****************************************************************************
*Function name
* lookup_named_type: search named type in the table
* internal function
*Input params
* fac : facility struct
- * td : type descriptor
+ * name : type name
*Return value
* : either find the named type, or create a new LttType
****************************************************************************/
-LttType * lookup_named_type(LttFacility *fac, type_descriptor_t * td)
+LttType * lookup_named_type(LttFacility *fac, GQuark type_name)
{
- GQuark name = g_quark_from_string(td->type_name);
+ LttType *type = NULL;
- LttType *type = g_datalist_id_get_data(&fac->named_types, name);
+ /* Named type */
+ type = g_datalist_id_get_data(&fac->named_types, name);
+ g_assert(type != NULL);
+#if 0
if(type == NULL){
+ /* Create the type */
type = g_new(LttType,1);
type->type_name = name;
- g_datalist_id_set_data_full(&fac->named_types, name,
- type, (GDestroyNotify)freeLttNamedType);
type->type_class = td->type;
if(td->fmt) type->fmt = g_strdup(td->fmt);
else type->fmt = NULL;
type->enum_strings = NULL;
type->element_type = NULL;
type->element_number = 0;
+
+ if(td->type_name != NULL)
+ g_datalist_id_set_data_full(&fac->named_types, name,
+ type, (GDestroyNotify)freeLttNamedType);
}
-
+#endif //0
return type;
}
-
+#endif //0
/*****************************************************************************
*Function name
}
g_array_free(fac->events, TRUE);
- g_datalist_clear(&fac->named_types);
+ g_datalist_clear(&fac->events_by_name);
+ // g_datalist_clear(&fac->named_types);
}
void freeEventtype(LttEventType * evType)
{
+ unsigned int i;
LttType * root_type;
if(evType->description)
- g_free(evType->description);
- if(evType->root_field){
- root_type = evType->root_field->field_type;
- freeLttField(evType->root_field);
- freeLttType(&root_type);
+ g_free(evType->description);
+
+ for(i=0; i<evType->fields->len;i++) {
+ LttField *field = &g_array_index(evType->fields, LttField, i);
+ freeLttField(field);
}
+ g_array_free(evType->fields, TRUE);
+ g_datalist_clear(&evType->fields_by_name);
}
-void freeLttNamedType(LttType * type)
-{
- freeLttType(&type);
-}
-
-void freeLttType(LttType ** type)
+void freeLttType(LttType * type)
{
unsigned int i;
- if(*type == NULL) return;
- //if((*type)->type_name){
- // return; //this is a named type
- //}
- if((*type)->fmt)
- g_free((*type)->fmt);
- if((*type)->enum_strings){
- g_free((*type)->enum_strings);
- }
- if((*type)->element_type){
- for(i=0;i<(*type)->element_number;i++)
- freeLttType(&((*type)->element_type[i]));
- g_free((*type)->element_type);
- }
- g_free(*type);
- *type = NULL;
-}
+ if(type->fmt)
+ g_free(type->fmt);
-void freeLttField(LttField * fld)
-{
- int i;
- int size = 0;
-
- if(fld->field_type){
- if(fld->field_type->type_class == LTT_ARRAY ||
- fld->field_type->type_class == LTT_SEQUENCE){
- size = 1;
- }else if(fld->field_type->type_class == LTT_STRUCT){
- size = fld->field_type->element_number;
- }
- }
+ if(type->enum_map)
+ g_hash_table_destroy(type->enum_map);
- if(fld->child){
- for(i=0; i<size; i++){
- if(fld->child[i])freeLttField(fld->child[i]);
+ if(type->fields) {
+ for(i=0; i<type->fields->len; i++) {
+ freeLttField(&g_array_index(type->fields, LttField, i));
}
- g_free(fld->child);
+ g_array_free(type->fields, TRUE);
}
- g_free(fld);
+ if(type->fields_by_name)
+ g_datalist_clear(&type->fields_by_name);
+
+ if(type->header)
+ g_free(type->header);//no need for condition? if(type->header)
+ if(type->separator)
+ g_free(type->separator);
+ if(type->footer)
+ g_free(type->footer);
+}
+
+void freeLttNamedType(LttType * type)
+{
+ freeLttType(type);
+}
+
+void freeLttField(LttField * field)
+{
+ if(field->description)
+ g_free(field->description);
+ if(field->dynamic_offsets)
+ g_array_free(field->dynamic_offsets, TRUE);
+ freeLttType(&field->field_type);
}
/*****************************************************************************
LttEventType *ltt_facility_eventtype_get_by_name(LttFacility *f, GQuark name)
{
- LttEventType *et = g_datalist_get_data(&f->events_by_name, name);
+ LttEventType *et = g_datalist_id_get_data(&f->events_by_name, name);
+ return et;
}