Files
asterisk/include/asterisk/stringfields.h

647 lines
22 KiB
C
Raw Normal View History

/*
* Asterisk -- An open source telephony toolkit.
*
* Copyright (C) 2006, Digium, Inc.
*
* Kevin P. Fleming <kpfleming@digium.com>
*
* See http://www.asterisk.org for more information about
* the Asterisk project. Please do not directly contact
* any of the maintainers of this project for assistance;
* the project provides a web site, mailing lists and IRC
* channels for your use.
*
* This program is free software, distributed under the terms of
* the GNU General Public License Version 2. See the LICENSE file
* at the top of the source tree.
*/
/*! \file
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
\page Stringfields String Fields
\brief String fields in structures
This file contains objects and macros used to manage string
fields in structures without requiring them to be allocated
as fixed-size buffers or requiring individual allocations for
for each field.
Using this functionality is quite simple. An example structure
with three fields is defined like this:
\code
struct sample_fields {
int x1;
AST_DECLARE_STRING_FIELDS(
AST_STRING_FIELD(foo);
AST_STRING_FIELD(bar);
AST_STRING_FIELD(blah);
);
long x2;
};
\endcode
When an instance of this structure is allocated (either statically or
dynamically), the fields and the pool of storage for them must be
initialized:
\code
struct sample_fields *x;
x = ast_calloc(1, sizeof(*x));
if (x == NULL || ast_string_field_init(x, 252)) {
if (x)
ast_free(x);
x = NULL;
... handle error
}
\endcode
Fields will default to pointing to an empty string, and will revert to
that when ast_string_field_set() is called with a NULL argument.
A string field will \b never contain NULL.
ast_string_field_init(x, 0) will reset fields to the
initial value while keeping the pool allocated.
Reading the fields is much like using 'const char * const' fields in the
structure: you cannot write to the field or to the memory it points to.
Writing to the fields must be done using the wrapper macros listed below;
and assignments are always by value (i.e. strings are copied):
* ast_string_field_set() stores a simple value;
* ast_string_field_build() builds the string using a printf-style format;
* ast_string_field_build_va() is the varargs version of the above;
* variants of these function allow passing a pointer to the field
as an argument.
\code
ast_string_field_set(x, foo, "infinite loop");
ast_string_field_set(x, foo, NULL); // set to an empty string
ast_string_field_ptr_set(x, &x->bar, "right way");
ast_string_field_build(x, blah, "%d %s", zipcode, city);
ast_string_field_ptr_build(x, &x->blah, "%d %s", zipcode, city);
ast_string_field_build_va(x, bar, fmt, args)
ast_string_field_ptr_build_va(x, &x->bar, fmt, args)
\endcode
When the structure instance is no longer needed, the fields
and their storage pool must be freed:
\code
ast_string_field_free_memory(x);
ast_free(x);
\endcode
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
A new feature "Extended String Fields" has been added in 13.9.0.
An extended field is one that is declared outside the AST_DECLARE_STRING_FIELDS
block but still inside the parent structure. It's most useful for extending
structures where adding a new string field to an existing AST_DECLARE_STRING_FIELDS
block would break ABI compatibility.
Example:
\code
struct original_structure_version {
AST_DECLARE_STRING_FIELDS(
AST_STRING_FIELD(foo);
AST_STRING_FIELD(bar);
);
int x1;
int x2;
};
\endcode
Adding "blah" to the existing string fields breaks ABI compatibility because it changes
the offsets of x1 and x2.
\code
struct new_structure_version {
AST_DECLARE_STRING_FIELDS(
AST_STRING_FIELD(foo);
AST_STRING_FIELD(bar);
AST_STRING_FIELD(blah);
);
int x1;
int x2;
};
\endcode
However, adding "blah" as an extended string field to the end of the structure doesn't break
ABI compatibility but still allows the use of the existing pool.
\code
struct new_structure_version {
AST_DECLARE_STRING_FIELDS(
AST_STRING_FIELD(foo);
AST_STRING_FIELD(bar);
);
int x1;
int x2;
AST_STRING_FIELD_EXTENDED(blah);
};
\endcode
The only additional step required is to call ast_string_field_init_extended so the
pool knows about the new field. It must be called AFTER ast_string_field_init or
ast_calloc_with_stringfields. Although ast_calloc_with_stringfields is used in the
sample below, it's not necessary for extended string fields.
\code
struct new_structure_version *x = ast_calloc_with_stringfields(1, struct new_structure_version, 252);
if (!x) {
return;
}
ast_string_field_init_extended(x, blah);
\endcode
The new field can now be treated just like any other string field and it's storage will
be released with the rest of the string fields.
\code
ast_string_field_set(x, foo, "infinite loop");
ast_stringfield_free_memory(x);
ast_free(x);
\endcode
This completes the API description.
*/
#ifndef _ASTERISK_STRINGFIELDS_H
#define _ASTERISK_STRINGFIELDS_H
#include "asterisk/inline_api.h"
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
#include "asterisk/vector.h"
/*!
\internal
\brief An opaque type for managed string fields in structures
Don't declare instances of this type directly; use the AST_STRING_FIELD()
macro instead.
In addition to the string itself, the amount of space allocated for the
field is stored in the two bytes immediately preceding it.
*/
typedef const char * ast_string_field;
/* the type of storage used to track how many bytes were allocated for a field */
typedef uint16_t ast_string_field_allocation;
/*!
\internal
\brief A constant empty string used for fields that have no other value
*/
extern const char *__ast_string_field_empty;
/*!
\internal
\brief Structure used to hold a pool of space for string fields
\note base is aligned so base+used can stay aligned by incrementing used with
aligned numbers only
*/
struct ast_string_field_pool {
struct ast_string_field_pool *prev; /*!< pointer to the previous pool, if any */
size_t size; /*!< the total size of the pool */
size_t used; /*!< the space used in the pool */
size_t active; /*!< the amount of space actively in use by fields */
char base[0] __attribute__((aligned(__alignof__(ast_string_field_allocation)))); /*!< storage space for the fields */
};
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
/*!
\internal
\brief The definition for the string field vector used for compare and copy
\since 13.9.0
*/
AST_VECTOR(ast_string_field_vector, const char **);
/*!
\internal
\brief Structure used to hold a pointer to the embedded pool and the field vector
\since 13.9.0
*/
struct ast_string_field_header {
struct ast_string_field_pool *embedded_pool; /*!< pointer to the embedded pool, if any */
struct ast_string_field_vector string_fields; /*!< field vector for compare and copy */
};
/*!
\internal
\brief Structure used to manage the storage for a set of string fields.
*/
struct ast_string_field_mgr {
ast_string_field last_alloc; /*!< the last field allocated */
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
struct ast_string_field_header *header; /*!< pointer to the header */
#if defined(__AST_DEBUG_MALLOC)
const char *owner_file; /*!< filename of owner */
const char *owner_func; /*!< function name of owner */
int owner_line; /*!< line number of owner */
#endif
};
/*!
\internal
\brief Attempt to 'grow' an already allocated field to a larger size
\param mgr Pointer to the pool manager structure
\param needed Amount of space needed for this field
\param ptr Pointer to a field within the structure
\return 0 on success, non-zero on failure
This function will attempt to increase the amount of space allocated to
an existing field to the amount requested; this is only possible if the
field was the last field allocated from the current storage pool and
the pool has enough space available. If so, the additional space will be
allocated to this field and the field's address will not be changed.
*/
int __ast_string_field_ptr_grow(struct ast_string_field_mgr *mgr,
struct ast_string_field_pool **pool_head, size_t needed,
const ast_string_field *ptr);
/*!
\internal
\brief Allocate space for a field
\param mgr Pointer to the pool manager structure
\param needed Amount of space needed for this field
\param fields Pointer to the first entry of the field array
\return NULL on failure, an address for the field on success.
This function will allocate the requested amount of space from
the field pool. If the requested amount of space is not available,
an additional pool will be allocated.
*/
ast_string_field __ast_string_field_alloc_space(struct ast_string_field_mgr *mgr,
struct ast_string_field_pool **pool_head, size_t needed);
/*!
\internal
\brief Set a field to a complex (built) value
\param mgr Pointer to the pool manager structure
\param pool_head Pointer to the current pool
\param ptr Pointer to a field within the structure
\param format printf-style format string
\return nothing
*/
void __ast_string_field_ptr_build(struct ast_string_field_mgr *mgr,
struct ast_string_field_pool **pool_head,
ast_string_field *ptr, const char *format, ...) __attribute__((format(printf, 4, 5)));
/*!
\internal
\brief Set a field to a complex (built) value
\param mgr Pointer to the pool manager structure
\param pool_head Pointer to the current pool
\param ptr Pointer to a field within the structure
\param format printf-style format string
\param args va_list of the args for the format_string
\return nothing
*/
void __ast_string_field_ptr_build_va(struct ast_string_field_mgr *mgr,
struct ast_string_field_pool **pool_head,
ast_string_field *ptr, const char *format, va_list ap) __attribute__((format(printf, 4, 0)));
/*!
\brief Declare a string field
\param name The field name
*/
#define AST_STRING_FIELD(name) const ast_string_field name
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
/*!
\brief Declare an extended string field
\since 13.9.0
\param name The field name
*/
#define AST_STRING_FIELD_EXTENDED(name) AST_STRING_FIELD(name)
enum ast_stringfield_cleanup_type {
/*!
* Reset all string fields and free all extra pools that may have been created
* The allocation or structure can be reused as is.
*/
AST_STRINGFIELD_RESET = 0,
/*!
* Reset all string fields and free all pools.
* If the pointer was returned by ast_calloc_with_stringfields, it can NOT be reused
* and should be immediately freed. Otherwise, you must call ast_string_field_init
* again if you want to reuse it.
*/
AST_STRINGFIELD_DESTROY = -1,
};
/*!
\brief Declare the fields needed in a structure
\param field_list The list of fields to declare, using AST_STRING_FIELD() for each one.
Internally, string fields are stored as a pointer to the head of the pool,
followed by individual string fields, and then a struct ast_string_field_mgr
which describes the space allocated.
We split the two variables so they can be used as markers around the
field_list, and this allows us to determine how many entries are in
the field, and play with them.
In particular, for writing to the fields, we rely on __field_mgr_pool to be
a non-const pointer, so we know it has the same size as ast_string_field,
and we can use it to locate the fields.
*/
#define AST_DECLARE_STRING_FIELDS(field_list) \
struct ast_string_field_pool *__field_mgr_pool; \
field_list \
struct ast_string_field_mgr __field_mgr
/*!
\brief Initialize a field pool and fields
\param x Pointer to a structure containing fields
\param size Amount of storage to allocate.
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
Use AST_STRINGFIELD_RESET to reset fields to the default value,
and release all but the most recent pool.
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
AST_STRINGFIELD_DESTROY (used internally) means free all pools which is
equivalent to calling ast_string_field_free_memory.
\return 0 on success, non-zero on failure
*/
#define ast_string_field_init(x, size) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__res__ = __ast_string_field_init(&(x)->__field_mgr, &(x)->__field_mgr_pool, size, __FILE__, __LINE__, __PRETTY_FUNCTION__); \
} \
__res__ ; \
})
/*!
* \brief free all memory - to be called before destroying the object
*
* \param x
*
*/
#define ast_string_field_free_memory(x) \
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__res__ = __ast_string_field_free_memory(&(x)->__field_mgr, &(x)->__field_mgr_pool, \
AST_STRINGFIELD_DESTROY, __FILE__, __LINE__, __PRETTY_FUNCTION__); \
} \
__res__; \
})
int __ast_string_field_free_memory(struct ast_string_field_mgr *mgr,
struct ast_string_field_pool **pool_head, enum ast_stringfield_cleanup_type cleanup_type,
const char *file, int lineno, const char *func);
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
/*!
* \brief Initialize an extended string field
* \since 13.9.0
*
* \param x Pointer to a structure containing the field
* \param field The extended field to initialize
* \retval zero on success
* \retval non-zero on error
*
* \note
* This macro must be called on ALL fields defined with AST_STRING_FIELD_EXTENDED after
* ast_string_field_init has been called.
*/
#define ast_string_field_init_extended(x, field) \
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL && (x)->__field_mgr.header != NULL) { \
ast_string_field *non_const = (ast_string_field *)&(x)->field; \
*non_const = __ast_string_field_empty; \
__res__ = AST_VECTOR_APPEND(&(x)->__field_mgr.header->string_fields, non_const); \
} \
__res__; \
})
/*!
* \internal
* \brief internal version of ast_string_field_init
*/
int __ast_string_field_init(struct ast_string_field_mgr *mgr, struct ast_string_field_pool **pool_head,
int needed, const char *file, int lineno, const char *func);
/*!
* \brief Allocate a structure with embedded stringfields in a single allocation
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
* \param n Current imlementation only allows 1 structure to be allocated
* \param type The type of structure to allocate
* \param size The number of bytes of space (minimum) to allocate for stringfields to use
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
* in each structure
*
* This function will allocate memory for one or more structures that use stringfields, and
* also allocate space for the stringfields and initialize the stringfield management
* structure embedded in the outer structure.
*
* \since 1.8
*/
#define ast_calloc_with_stringfields(n, type, size) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
__ast_calloc_with_stringfields(n, sizeof(type), offsetof(type, __field_mgr), \
offsetof(type, __field_mgr_pool), size, __FILE__, __LINE__, __PRETTY_FUNCTION__)
/*!
* \internal
* \brief internal version of ast_calloc_with_stringfields
*/
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
void * attribute_malloc __ast_calloc_with_stringfields(unsigned int num_structs,
size_t struct_size, size_t field_mgr_offset, size_t field_mgr_pool_offset, size_t pool_size,
const char *file, int lineno, const char *func);
/*!
\internal
\brief Release a field's allocation from a pool
\param pool_head Pointer to the current pool
\param ptr Field to be released
\return nothing
This function will search the pool list to find the pool that contains
the allocation for the specified field, then remove the field's allocation
from that pool's 'active' count. If the pool's active count reaches zero,
and it is not the current pool, then it will be freed.
*/
void __ast_string_field_release_active(struct ast_string_field_pool *pool_head,
const ast_string_field ptr);
/*!
\brief Macro to provide access to the allocation field that lives immediately in front of a string field
\param x Pointer to the string field
Note that x must be a pointer to a byte-sized type -- normally (char *) -- or this calculation
would break horribly
*/
#define AST_STRING_FIELD_ALLOCATION(x) *((ast_string_field_allocation *) (x - __alignof__(ast_string_field_allocation)))
/*!
\brief Set a field to a simple string value
\param x Pointer to a structure containing fields
\param ptr Pointer to a field within the structure
\param data String value to be copied into the field
\retval zero on success
\retval non-zero on error
*/
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
#define ast_string_field_ptr_set(x, ptr, data) \
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__res__ = ast_string_field_ptr_set_by_fields((x)->__field_mgr_pool, (x)->__field_mgr, ptr, data); \
} \
__res__; \
})
#define ast_string_field_ptr_set_by_fields(field_mgr_pool, field_mgr, ptr, data) \
({ \
int __res__ = 0; \
const char *__d__ = (data); \
size_t __dlen__ = (__d__) ? strlen(__d__) + 1 : 1; \
ast_string_field *__p__ = (ast_string_field *) (ptr); \
ast_string_field target = *__p__; \
if (__dlen__ == 1) { \
__ast_string_field_release_active(field_mgr_pool, *__p__); \
*__p__ = __ast_string_field_empty; \
} else if ((__dlen__ <= AST_STRING_FIELD_ALLOCATION(*__p__)) || \
(!__ast_string_field_ptr_grow(&field_mgr, &field_mgr_pool, __dlen__, __p__)) || \
(target = __ast_string_field_alloc_space(&field_mgr, &field_mgr_pool, __dlen__))) { \
if (target != *__p__) { \
__ast_string_field_release_active(field_mgr_pool, *__p__); \
*__p__ = target; \
} \
memcpy(* (void **) __p__, __d__, __dlen__); \
} else { \
__res__ = -1; \
} \
__res__; \
})
/*!
\brief Set a field to a simple string value
\param x Pointer to a structure containing fields
\param field Name of the field to set
\param data String value to be copied into the field
\retval zero on success
\retval non-zero on error
*/
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
#define ast_string_field_set(x, field, data) \
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__res__ = ast_string_field_ptr_set(x, &(x)->field, data); \
} \
__res__; \
})
/*!
\brief Set a field to a complex (built) value
\param x Pointer to a structure containing fields
\param ptr Pointer to a field within the structure
\param fmt printf-style format string
\param args Arguments for format string
\return nothing
*/
#define ast_string_field_ptr_build(x, ptr, fmt, args...) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__ast_string_field_ptr_build(&(x)->__field_mgr, &(x)->__field_mgr_pool, (ast_string_field *) ptr, fmt, args); \
__res__ = 0; \
} \
__res__; \
})
/*!
\brief Set a field to a complex (built) value
\param x Pointer to a structure containing fields
\param field Name of the field to set
\param fmt printf-style format string
\param args Arguments for format string
\return nothing
*/
#define ast_string_field_build(x, field, fmt, args...) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__ast_string_field_ptr_build(&(x)->__field_mgr, &(x)->__field_mgr_pool, (ast_string_field *) &(x)->field, fmt, args); \
__res__ = 0; \
} \
__res__; \
})
/*!
\brief Set a field to a complex (built) value with prebuilt va_lists.
\param x Pointer to a structure containing fields
\param ptr Pointer to a field within the structure
\param fmt printf-style format string
\param args Arguments for format string in va_list format
\return nothing
*/
#define ast_string_field_ptr_build_va(x, ptr, fmt, args) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__ast_string_field_ptr_build_va(&(x)->__field_mgr, &(x)->__field_mgr_pool, (ast_string_field *) ptr, fmt, args); \
__res__ = 0; \
} \
__res__; \
})
/*!
\brief Set a field to a complex (built) value
\param x Pointer to a structure containing fields
\param field Name of the field to set
\param fmt printf-style format string
\param args Arguments for format string in va_list format
\return nothing
*/
#define ast_string_field_build_va(x, field, fmt, args) \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
({ \
int __res__ = -1; \
if (((void *)(x)) != NULL) { \
__ast_string_field_ptr_build_va(&(x)->__field_mgr, &(x)->__field_mgr_pool, (ast_string_field *) &(x)->field, fmt, args); \
__res__ = 0; \
} \
__res__; \
})
/*!
\brief Compare the string fields in two instances of the same structure
\since 12
\param instance1 The first instance of the structure to be compared
\param instance2 The second instance of the structure to be compared
\retval zero if all string fields are equal (does not compare non-string field data)
\retval non-zero if the values of the string fields differ
*/
#define ast_string_fields_cmp(instance1, instance2) \
({ \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
int __res__ = -1; \
if (((void *)(instance1)) != NULL && ((void *)(instance2)) != NULL) { \
__res__ = __ast_string_fields_cmp(&(instance1)->__field_mgr.header->string_fields, \
&(instance2)->__field_mgr.header->string_fields); \
} \
__res__; \
})
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
int __ast_string_fields_cmp(struct ast_string_field_vector *left, struct ast_string_field_vector *right);
/*!
\brief Copy all string fields from one instance to another of the same structure
\since 12
\param copy The instance of the structure to be copied into
\param orig The instance of the structure to be copied from
\retval zero on success
\retval non-zero on error
*/
#define ast_string_fields_copy(copy, orig) \
({ \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
int __res__ = -1; \
if (((void *)(copy)) != NULL && ((void *)(orig)) != NULL) { \
__res__ = __ast_string_fields_copy(((copy)->__field_mgr_pool), \
(struct ast_string_field_mgr *)&((copy)->__field_mgr), \
(struct ast_string_field_mgr *)&((orig)->__field_mgr)); \
} \
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
__res__; \
})
stringfields: Refactor to allow fields to be added to the end of structures String fields are great, except that you can't add new ones without breaking ABI compatibility because it shifts down everything else in the structure. The only alternative is to add your own char * field to the end of the structure and manage the memory yourself which isn't ideal, especially since you then can't use the OPT_STRINGFIELD_T type. Background: The reason string fields had to be declared inside the AST_DECLARE_STRING_FIELDS block was to facilitate iteration over all declared fields for initialization, compare and copy. Since AST_DECLARE_STRING_FIELDS declared the pool, then the fields, then the manager, you could use the offsets of the pool and manager and iterate over the sequential addresses in between to access the fields. The actual pool, field allocation and field set operations don't actually care where the field is. It's just iteration over the fields that was the problem. Solution: Extended String Fields An extended string field is one that is declared outside the AST_DECLARE_STRING_FIELDS block but still (anywhere) inside the parent structure. Other than using AST_STRING_FIELD_EXTENDED instead of AST_STRING_FIELD, it looks the same as other string fields. It's storage comes from the pool and it participates in string field compare and copy operations peformed on the parent structure. It's also a valid target for the OPT_STRINGFIELD_T aco option type. Implementation: To keep track of the extended fields and make sure that ABI isn't broken, the existing embedded_pool pointer in the manager structure was repurposed to be a pointer to a separate header structure that contains the embedded_pool pointer plus a vector of fields. The length of the manager structure didn't change and the embedded_pool pointer isn't used in the macros, only the stringfields C code. A side benefit of this is that changing the header structure in the future won't break ABI. ast_string_fields_init initializes the normal string fields and appends them to the vector, and subsequent calls to ast_string_field_init_extended initialize and append the extended fields. Cleanup, ast_string_fields_cmp, and ast_string_fields_copy can now work on the vector instead of sequentially traversing the addresses between the pool and manager. The total size of a structure using string fields didn't change, whether using extended fields or not, nor have the offsets of any structure members, either inside the original block or outside. Adding an extended field to the end of a structure is the same as adding a char *. Details: The stringfield C code was pulled out from utils.c and into stringfields.c. It just made sense. Additional work was done in ast_string_field_init and ast_calloc_with_stringfields to handle the allocation of the new header structure and the vector, and the associated cleanup. In the process some additional NULL pointer checking was added. A lot of work was done in stringfields.h since the logic for compare and copy is there. Documentation was added as well as somne additional NULL checking. The ability to call ast_calloc_with_stringfields with a number of structures greater than 1 never really worked. Well, the calloc worked but there was no way to access the additional structures or clean them up. It was agreed that there was no use case for requesting more than 1 structure so an ast_assert was added to prevent it and the iteration code removed. Testing: The stringfield unit tests were updated to test both normal and extended fields. Tests for ast_string_field_ptr_set_by_fields and ast_calloc_with_stringfields were also added. As an ABI test, 13 was compiled from git and the res_pjsip_* modules, except res_pjsip itself, saved off. The patch was then added and a full compile and install was performed. Then the older res_pjsip_* moduled were copied over the installed versions so res_pjsip was new and the rest were old. No issues. contact->aor, which is a char * at the end of contact, was then changed to an extended string field and a recompile and reinstall was performed, again leaving stock versions of the the res_pjsip_* modules. Again, no issues with the res_pjsip_* modules using the old stringfield implementation and with contact->aor as a char *, and res_pjsip itself using the new stringfield implementation and contact->aor being an extended string field. Finally, several existing string fields were converted to extended string fields to test OPT_STRINGFIELD_T. Again, no issues. Change-Id: I235db338c5b178f5a13b7946afbaa5d4a0f91d61
2016-03-25 22:22:34 -06:00
int __ast_string_fields_copy(struct ast_string_field_pool *copy_pool,
struct ast_string_field_mgr *copy_mgr, struct ast_string_field_mgr *orig_mgr);
#endif /* _ASTERISK_STRINGFIELDS_H */