37.2 Dynamic Linker Introspection

The GNU C Library provides various facilities for querying information from the dynamic linker.

A link map is associated with the main executable and each shared object. Some fields of the link map are accessible to applications and exposed through the struct link_map. Applications must not modify the link map directly.

Pointers to link maps can be obtained from the _r_debug variable, from the RTLD_DI_LINKMAP request for dlinfo, and from the _dl_find_object function. See below for details.

l_addr

This field contains the load address of the object. This is the offset that needs to be applied to unrelocated addresses in the object image (as stored on disk) to form an address that can be used at run time for accessing data or running code. For position-dependent executables, the load address is typically zero, and no adjustment is required. For position-independent objects, the l_addr field usually contains the address of the object’s ELF header in the process image. However, this correspondence is not guaranteed because the ELF header might not be mapped at all, and the ELF file as stored on disk might use zero as the lowest virtual address. Due to the second variable, values of the l_addr field do not necessarily uniquely identify a shared object.

On Linux, to obtain the lowest loaded address of the main program, use getauxval to obtain the AT_PHDR and AT_PHNUM values for the current process. Alternatively, call ‘dlinfo (_r_debug.r_map, &phdr)’ to obtain the number of program headers, and the address of the program header array will be stored in phdr (of type const ElfW(Phdr) *, as explained below). These values allow processing the array of program headers and the address information in the PT_LOAD entries among them. This works even when the program was started with an explicit loader invocation.

l_name

For a shared object, this field contains the file name that the the GNU C Library dynamic loader used when opening the object. This can be a relative path (relative to the current directory at process start, or if the object was loaded later, via dlopen or dlmopen). Symbolic links are not necessarily resolved.

For the main executable, l_name is ‘""’ (the empty string). (The main executable is not loaded by the GNU C Library, so its file name is not available.) On Linux, the main executable is available as /proc/self/exe (unless an explicit loader invocation was used to start the program). The file name /proc/self/exe continues to resolve to the same file even if it is moved within or deleted from the file system. Its current location can be read using readlink. See Symbolic Links. (Although /proc/self/exe is not actually a symbol link, it is only presented as one.) Note that /proc may not be mounted, in which case /proc/self/exe is not available.

If an explicit loader invocation is used (such as ‘ld.so /usr/bin/emacs’), the /proc/self/exe approach does not work because the file name refers to the dynamic linker ld.so, and not the /usr/bin/emacs program. An approximation to the executable path is still available in the info.dli_fname member after calling ‘dladdr (_r_debug.r_map->l_ld, &info)’. Note that this could be a relative path, and it is supplied by the process that created the current process, not the kernel, so it could be inaccurate.

l_ld

This is a pointer to the ELF dynamic segment, an array of tag/value pairs that provide various pieces of information that the dynamic linking process uses. On most architectures, addresses in the dynamic segment are relocated at run time, but on some architectures and in some run-time configurations, it is necessary to add the l_addr field value to obtain a proper address.

l_prev
l_next

These fields are used to maintain a double-linked linked list of all link maps within one dlmopen namespace. Note that there is currently no thread-safe way to iterate over this list. The callback-based dl_iterate_phdr interface can be used instead.

Portability note: It is not possible to create a struct link_map object and pass a pointer to a function that expects a struct link_map * argument. Only link map pointers initially supplied by the GNU C Library are permitted as arguments. In current versions of the GNU C Library, handles returned by dlopen and dlmopen are pointers to link maps. However, this is not a portable assumption, and may even change in future versions of the GNU C Library. To obtain the link map associated with a handle, see dlinfo and RTLD_DI_LINKMAP below. If a function accepts both dlopen/dlmopen handles and struct link_map pointers in its void * argument, that is documented explicitly.

37.2.1 Querying information for loaded objects

The dlinfo function provides access to internal information associated with dlopen/dlmopen handles and link maps.

Function: int dlinfo (void *handle, int request, void *arg)

| MT-Safe | AS-Unsafe corrupt | AC-Unsafe corrupt | See POSIX Safety Concepts.

This function returns information about handle in the memory location arg, based on request. The handle argument must be a pointer returned by dlopen or dlmopen; it must not have been closed by dlclose. Alternatively, handle can be a struct link_map * value for a link map of an object that has not been closed.

On success, dlinfo returns 0 for most request types; exceptions are noted below. If there is an error, the function returns -1, and dlerror can be used to obtain a corresponding error message.

The following operations are defined for use with request:

RTLD_DI_LINKMAP

The corresponding struct link_map pointer for handle is written to *arg. The arg argument must be the address of an object of type struct link_map *.

RTLD_DI_LMID

The namespace identifier of handle is written to *arg. The arg argument must be the address of an object of type Lmid_t.

RTLD_DI_ORIGIN

The value of the $ORIGIN dynamic string token for handle is written to the character array starting at arg as a null-terminated string.

This request type should not be used because it is prone to buffer overflows.

RTLD_DI_SERINFO
RTLD_DI_SERINFOSIZE

These requests can be used to obtain search path information for handle. For both requests, arg must point to a Dl_serinfo object. The RTLD_DI_SERINFOSIZE request must be made first; it updates the dls_size and dls_cnt members of the Dl_serinfo object. The caller should then allocate memory to store at least dls_size bytes and pass that buffer to a RTLD_DI_SERINFO request. This second request fills the dls_serpath array. The number of array elements was returned in the dls_cnt member in the initial RTLD_DI_SERINFOSIZE request. The caller is responsible for freeing the allocated buffer.

This interface is prone to buffer overflows in multi-threaded processes because the required size can change between the RTLD_DI_SERINFOSIZE and RTLD_DI_SERINFO requests.

RTLD_DI_TLS_DATA

This request writes the address of the TLS block (in the current thread) for the shared object identified by handle to *arg. The argument arg must be the address of an object of type void *. A null pointer is written if the object does not have any associated TLS block.

RTLD_DI_TLS_MODID

This request writes the TLS module ID for the shared object handle to *arg. The argument arg must be the address of an object of type size_t. The module ID is zero if the object does not have an associated TLS block.

RTLD_DI_PHDR

This request writes the address of the program header array to *arg. The argument arg must be the address of an object of type const ElfW(Phdr) * (that is, const Elf32_Phdr * or const Elf64_Phdr *, as appropriate for the current architecture). For this request, the value returned by dlinfo is the number of program headers in the program header array.

The dlinfo function is a GNU extension.

The remainder of this section documents the _dl_find_object function and supporting types and constants.

Data Type: struct dl_find_object

This structure contains information about a main program or loaded object. The _dl_find_object function uses it to return result data to the caller.

unsigned long long int dlfo_flags

Currently unused and always 0.

void *dlfo_map_start

The start address of the inspected mapping. This information comes from the program header, so it follows its convention, and the address is not necessarily page-aligned.

void *dlfo_map_end

The end address of the mapping.

struct link_map *dlfo_link_map

This member contains a pointer to the link map of the object.

void *dlfo_eh_frame

This member contains a pointer to the exception handling data of the object. See DLFO_EH_SEGMENT_TYPE below.

This structure is a GNU extension.

Macro: int DLFO_STRUCT_HAS_EH_DBASE

On most targets, this macro is defined as 0. If it is defined to 1, struct dl_find_object contains an additional member dlfo_eh_dbase of type void *. It is the base address for DW_EH_PE_datarel DWARF encodings to this location.

This macro is a GNU extension.

Macro: int DLFO_STRUCT_HAS_EH_COUNT

On most targets, this macro is defined as 0. If it is defined to 1, struct dl_find_object contains an additional member dlfo_eh_count of type int. It is the number of exception handling entries in the EH frame segment identified by the dlfo_eh_frame member.

This macro is a GNU extension.

Macro: int DLFO_EH_SEGMENT_TYPE

On targets using DWARF-based exception unwinding, this macro expands to PT_GNU_EH_FRAME. This indicates that dlfo_eh_frame in struct dl_find_object points to the PT_GNU_EH_FRAME segment of the object. On targets that use other unwinding formats, the macro expands to the program header type for the unwinding data.

This macro is a GNU extension.

Function: int _dl_find_object (void *address, struct dl_find_object *result)

| MT-Safe | AS-Safe | AC-Safe | See POSIX Safety Concepts.

On success, this function returns 0 and writes about the object surrounding the address to *result. On failure, -1 is returned.

The address can be a code address or data address. On architectures using function descriptors, no attempt is made to decode the function descriptor. Depending on how these descriptors are implemented, _dl_find_object may return the object that defines the function descriptor (and not the object that contains the code implementing the function), or fail to find any object at all.

On success address is greater than or equal to result->dlfo_map_start and less than result->dlfo_map_end, that is, the supplied code address is located within the reported mapping.

This function returns a pointer to the unwinding information for the object that contains the program code address in result->dlfo_eh_frame. If the platform uses DWARF unwinding information, this is the in-memory address of the PT_GNU_EH_FRAME segment. See DLFO_EH_SEGMENT_TYPE above. In case address resides in an object that lacks unwinding information, the function still returns 0, but sets result->dlfo_eh_frame to a null pointer.

_dl_find_object itself is thread-safe. However, if the application invokes dlclose for the object that contains address concurrently with _dl_find_object or after the call returns, accessing the unwinding data for that object or the link map (through result->dlfo_link_map) is not safe. Therefore, the application needs to ensure by other means (e.g., by convention) that address remains a valid code address while the unwinding information is processed.

This function is a GNU extension.