@node Dynamic Linker
@c @node Dynamic Linker, Internal Probes, Threads, Top
@c %MENU% Loading programs and shared objects.
@chapter Dynamic Linker
@cindex dynamic linker
@cindex dynamic loader

The @dfn{dynamic linker} is responsible for loading dynamically linked
programs and their dependencies (in the form of shared objects).  The
dynamic linker in @theglibc{} also supports loading shared objects (such
as plugins) later at run time.

Dynamic linkers are sometimes called @dfn{dynamic loaders}.

@menu
* Dynamic Linker Invocation::   Explicit invocation of the dynamic linker.
* Dynamic Linker Introspection::    Interfaces for querying mapping information.
* Dynamic Linker Hardening::    Avoiding unexpected issues with dynamic linking.
@end menu

@node Dynamic Linker Invocation
@section Dynamic Linker Invocation

@cindex program interpreter
When a dynamically linked program starts, the operating system
automatically loads the dynamic linker along with the program.
@Theglibc{} also supports invoking the dynamic linker explicitly to
launch a program.  This command uses the implied dynamic linker
(also sometimes called the @dfn{program interpreter}):

@smallexample
sh -c 'echo "Hello, world!"'
@end smallexample

This command specifies the dynamic linker explicitly:

@smallexample
ld.so /bin/sh -c 'echo "Hello, world!"'
@end smallexample

Note that @command{ld.so} does not search the @env{PATH} environment
variable, so the full file name of the executable needs to be specified.

The @command{ld.so} program supports various options.  Options start
@samp{--} and need to come before the program that is being launched.
Some of the supported options are listed below.

@table @code
@item --list-diagnostics
Print system diagnostic information in a machine-readable format.
@xref{Dynamic Linker Diagnostics}.
@end table

@menu
* Dynamic Linker Diagnostics::   Obtaining system diagnostic information.
@end menu

@node Dynamic Linker Diagnostics
@subsection Dynamic Linker Diagnostics
@cindex diagnostics (dynamic linker)

The @samp{ld.so --list-diagnostics} produces machine-readable
diagnostics output.  This output contains system data that affects the
behavior of @theglibc{}, and potentially application behavior as well.

The exact set of diagnostic items can change between releases of
@theglibc{}.  The output format itself is not expected to change
radically.

The following table shows some example lines that can be written by the
diagnostics command.

@table @code
@item dl_pagesize=0x1000
The system page size is 4096 bytes.

@item env[0x14]="LANG=en_US.UTF-8"
This item indicates that the 21st environment variable at process
startup contains a setting for @code{LANG}.

@item env_filtered[0x22]="DISPLAY"
The 35th environment variable is @code{DISPLAY}.  Its value is not
included in the output for privacy reasons because it is not recognized
as harmless by the diagnostics code.

@item path.prefix="/usr"
This means that @theglibc{} was configured with @code{--prefix=/usr}.

@item path.system_dirs[0x0]="/lib64/"
@itemx path.system_dirs[0x1]="/usr/lib64/"
The built-in dynamic linker search path contains two directories,
@code{/lib64} and @code{/usr/lib64}.
@end table

@menu
* Dynamic Linker Diagnostics Format::  Format of ld.so output.
* Dynamic Linker Diagnostics Values::  Data contain in ld.so output.
@end menu

@node Dynamic Linker Diagnostics Format
@subsubsection Dynamic Linker Diagnostics Format

As seen above, diagnostic lines assign values (integers or strings) to a
sequence of labeled subscripts, separated by @samp{.}.  Some subscripts
have integer indices associated with them.  The subscript indices are
not necessarily contiguous or small, so an associative array should be
used to store them.  Currently, all integers fit into the 64-bit
unsigned integer range.  Every access path to a value has a fixed type
(string or integer) independent of subscript index values.  Likewise,
whether a subscript is indexed does not depend on previous indices (but
may depend on previous subscript labels).

A syntax description in ABNF (RFC 5234) follows.  Note that
@code{%x30-39} denotes the range of decimal digits.  Diagnostic output
lines are expected to match the @code{line} production.

@c ABNF-START
@smallexample
HEXDIG = %x30-39 / %x61-6f ; lowercase a-f only
ALPHA = %x41-5a / %x61-7a / %x7f ; letters and underscore
ALPHA-NUMERIC = ALPHA / %x30-39 / "_"
DQUOTE = %x22 ; "

; Numbers are always hexadecimal and use a 0x prefix.
hex-value-prefix = %x30 %x78
hex-value = hex-value-prefix 1*HEXDIG

; Strings use octal escape sequences and \\, \".
string-char = %x20-21 / %x23-5c / %x5d-7e ; printable but not "\
string-quoted-octal = %x30-33 2*2%x30-37
string-quoted = "\" ("\" / DQUOTE / string-quoted-octal)
string-value = DQUOTE *(string-char / string-quoted) DQUOTE

value = hex-value / string-value

label = ALPHA *ALPHA-NUMERIC
index = "[" hex-value "]"
subscript = label [index]

line = subscript *("." subscript) "=" value
@end smallexample

@node Dynamic Linker Diagnostics Values
@subsubsection Dynamic Linker Diagnostics Values

As mentioned above, the set of diagnostics may change between
@theglibc{} releases.  Nevertheless, the following table documents a few
common diagnostic items.  All numbers are in hexadecimal, with a
@samp{0x} prefix.

@table @code
@item dl_dst_lib=@var{string}
The @code{$LIB} dynamic string token expands to @var{string}.

@cindex HWCAP (diagnostics)
@item dl_hwcap=@var{integer}
@itemx dl_hwcap2=@var{integer}
The HWCAP and HWCAP2 values, as returned for @code{getauxval}, and as
used in other places depending on the architecture.

@cindex page size (diagnostics)
@item dl_pagesize=@var{integer}
The system page size is @var{integer} bytes.

@item dl_platform=@var{string}
The @code{$PLATFORM} dynamic string token expands to @var{string}.

@item dso.libc=@var{string}
This is the soname of the shared @code{libc} object that is part of
@theglibc{}.  On most architectures, this is @code{libc.so.6}.

@item env[@var{index}]=@var{string}
@itemx env_filtered[@var{index}]=@var{string}
An environment variable from the process environment.  The integer
@var{index} is the array index in the environment array.  Variables
under @code{env} include the variable value after the @samp{=} (assuming
that it was present), variables under @code{env_filtered} do not.

@item path.prefix=@var{string}
This indicates that @theglibc{} was configured using
@samp{--prefix=@var{string}}.

@item path.sysconfdir=@var{string}
@Theglibc{} was configured (perhaps implicitly) with
@samp{--sysconfdir=@var{string}} (typically @code{/etc}).

@item path.system_dirs[@var{index}]=@var{string}
These items list the elements of the built-in array that describes the
default library search path.  The value @var{string} is a directory file
name with a trailing @samp{/}.

@item path.rtld=@var{string}
This string indicates the application binary interface (ABI) file name
of the run-time dynamic linker.

@item version.release="stable"
@itemx version.release="development"
The value @code{"stable"} indicates that this build of @theglibc{} is
from a release branch.  Releases labeled as @code{"development"} are
unreleased development versions.

@cindex version (diagnostics)
@item version.version="@var{major}.@var{minor}"
@itemx version.version="@var{major}.@var{minor}.9000"
@Theglibc{} version.  Development releases end in @samp{.9000}.

@cindex auxiliary vector (diagnostics)
@item auxv[@var{index}].a_type=@var{type}
@itemx auxv[@var{index}].a_val=@var{integer}
@itemx auxv[@var{index}].a_val_string=@var{string}
An entry in the auxiliary vector (specific to Linux).  The values
@var{type} (an integer) and @var{integer} correspond to the members of
@code{struct auxv}.  If the value is a string, @code{a_val_string} is
used instead of @code{a_val}, so that values have consistent types.

The @code{AT_HWCAP} and @code{AT_HWCAP2} values in this output do not
reflect adjustment by @theglibc{}.

@item uname.sysname=@var{string}
@itemx uname.nodename=@var{string}
@itemx uname.release=@var{string}
@itemx uname.version=@var{string}
@itemx uname.machine=@var{string}
@itemx uname.domain=@var{string}
These Linux-specific items show the values of @code{struct utsname}, as
reported by the @code{uname} function.  @xref{Platform Type}.

@item aarch64.cpu_features.@dots{}
These items are specific to the AArch64 architectures.  They report data
@theglibc{} uses to activate conditionally supported features such as
BTI and MTE, and to select alternative function implementations.

@item aarch64.processor[@var{index}].@dots{}
These are additional items for the AArch64 architecture and are
described below.

@item aarch64.processor[@var{index}].requested=@var{kernel-cpu}
The kernel is told to run the subsequent probing on the CPU numbered
@var{kernel-cpu}.  The values @var{kernel-cpu} and @var{index} can be
distinct if there are gaps in the process CPU affinity mask.  This line
is not included if CPU affinity mask information is not available.

@item aarch64.processor[@var{index}].observed=@var{kernel-cpu}
This line reports the kernel CPU number @var{kernel-cpu} on which the
probing code initially ran.  If the CPU number cannot be obtained,
this line is not printed.

@item aarch64.processor[@var{index}].observed_node=@var{node}
This reports the observed NUMA node number, as reported by the
@code{getcpu} system call.  If this information cannot be obtained, this
line is not printed.

@item aarch64.processor[@var{index}].midr_el1=@var{value}
The value of the @code{midr_el1} system register on the processor
@var{index}.  This line is only printed if the kernel indicates that
this system register is supported.

@item aarch64.processor[@var{index}].dczid_el0=@var{value}
The value of the @code{dczid_el0} system register on the processor
@var{index}.

@cindex CPUID (diagnostics)
@item x86.cpu_features.@dots{}
These items are specific to the i386 and x86-64 architectures.  They
reflect supported CPU features and information on cache geometry, mostly
collected using the CPUID instruction.

@item x86.processor[@var{index}].@dots{}
These are additional items for the i386 and x86-64 architectures, as
described below.  They mostly contain raw data from the CPUID
instruction.  The probes are performed for each active CPU for the
@code{ld.so} process, and data for different probed CPUs receives a
uniqe @var{index} value.  Some CPUID data is expected to differ from CPU
core to CPU core.  In some cases, CPUs are not correctly initialized and
indicate the presence of different feature sets.

@item x86.processor[@var{index}].requested=@var{kernel-cpu}
The kernel is told to run the subsequent probing on the CPU numbered
@var{kernel-cpu}.  The values @var{kernel-cpu} and @var{index} can be
distinct if there are gaps in the process CPU affinity mask.  This line
is not included if CPU affinity mask information is not available.

@item x86.processor[@var{index}].observed=@var{kernel-cpu}
This line reports the kernel CPU number @var{kernel-cpu} on which the
probing code initially ran.  If the CPU number cannot be obtained,
this line is not printed.

@item x86.processor[@var{index}].observed_node=@var{node}
This reports the observed NUMA node number, as reported by the
@code{getcpu} system call.  If this information cannot be obtained, this
line is not printed.

@item x86.processor[@var{index}].cpuid_leaves=@var{count}
This line indicates that @var{count} distinct CPUID leaves were
encountered.  (This reflects internal @code{ld.so} storage space, it
does not directly correspond to @code{CPUID} enumeration ranges.)

@item x86.processor[@var{index}].ecx_limit=@var{value}
The CPUID data extraction code uses a brute-force approach to enumerate
subleaves (see the @samp{.subleaf_eax} lines below).  The last
@code{%rcx} value used in a CPUID query on this probed CPU was
@var{value}.

@item x86.processor[@var{index}].cpuid.eax[@var{query_eax}].eax=@var{eax}
@itemx x86.processor[@var{index}].cpuid.eax[@var{query_eax}].ebx=@var{ebx}
@itemx x86.processor[@var{index}].cpuid.eax[@var{query_eax}].ecx=@var{ecx}
@itemx x86.processor[@var{index}].cpuid.eax[@var{query_eax}].edx=@var{edx}
These lines report the register contents after executing the CPUID
instruction with @samp{%rax == @var{query_eax}} and @samp{%rcx == 0} (a
@dfn{leaf}).  For the first probed CPU (with a zero @var{index}), only
leaves with non-zero register contents are reported.  For subsequent
CPUs, only leaves whose register contents differs from the previously
probed CPUs (with @var{index} one less) are reported.

Basic and extended leaves are reported using the same syntax.  This
means there is a large jump in @var{query_eax} for the first reported
extended leaf.

@item x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].eax=@var{eax}
@itemx x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].ebx=@var{ebx}
@itemx x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].ecx=@var{ecx}
@itemx x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].edx=@var{edx}
This is similar to the leaves above, but for a @dfn{subleaf}.  For
subleaves, the CPUID instruction is executed with @samp{%rax ==
@var{query_eax}} and @samp{%rcx == @var{query_ecx}}, so the result
depends on both register values.  The same rules about filtering zero
and identical results apply.

@item x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].until_ecx=@var{ecx_limit}
Some CPUID results are the same regardless the @var{query_ecx} value.
If this situation is detected, a line with the @samp{.until_ecx}
selector ins included, and this indicates that the CPUID register
contents is the same for @code{%rcx} values between @var{query_ecx}
and @var{ecx_limit} (inclusive).

@item x86.processor[@var{index}].cpuid.subleaf_eax[@var{query_eax}].ecx[@var{query_ecx}].ecx_query_mask=0xff
This line indicates that in an @samp{.until_ecx} range, the CPUID
instruction preserved the lowested 8 bits of the input @code{%rcx} in
the output @code{%rcx} registers.  Otherwise, the subleaves in the range
have identical values.  This special treatment is necessary to report
compact range information in case such copying occurs (because the
subleaves would otherwise be all different).

@item x86.processor[@var{index}].xgetbv.ecx[@var{query_ecx}]=@var{result}
This line shows the 64-bit @var{result} value in the @code{%rdx:%rax}
register pair after executing the XGETBV instruction with @code{%rcx}
set to @var{query_ecx}.  Zero values and values matching the previously
probed CPU are omitted.  Nothing is printed if the system does not
support the XGETBV instruction.
@end table

@node Dynamic Linker Introspection
@section Dynamic Linker Introspection

@Theglibc{} provides various functions for querying information from the
dynamic linker.

@deftypefun {int} dlinfo (void *@var{handle}, int @var{request}, void *@var{arg})
@safety{@mtsafe{}@asunsafe{@asucorrupt{}}@acunsafe{@acucorrupt{}}}
@standards{GNU, dlfcn.h}
This function returns information about @var{handle} in the memory
location @var{arg}, based on @var{request}.  The @var{handle} argument
must be a pointer returned by @code{dlopen} or @code{dlmopen}; it must
not have been closed by @code{dlclose}.

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

The following operations are defined for use with @var{request}:

@vtable @code
@item RTLD_DI_LINKMAP
The corresponding @code{struct link_map} pointer for @var{handle} is
written to @code{*@var{arg}}.  The @var{arg} argument must be the
address of an object of type @code{struct link_map *}.

@item RTLD_DI_LMID
The namespace identifier of @var{handle} is written to
@code{*@var{arg}}.  The @var{arg} argument must be the address of an
object of type @code{Lmid_t}.

@item RTLD_DI_ORIGIN
The value of the @code{$ORIGIN} dynamic string token for @var{handle} is
written to the character array starting at @var{arg} as a
null-terminated string.

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

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

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

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

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

The @code{dlinfo} function is a GNU extension.
@end deftypefun

The remainder of this section documents the @code{_dl_find_object}
function and supporting types and constants.

@deftp {Data Type} {struct dl_find_object}
@standards{GNU, dlfcn.h}
This structure contains information about a main program or loaded
object.  The @code{_dl_find_object} function uses it to return
result data to the caller.

@table @code
@item unsigned long long int dlfo_flags
Currently unused and always 0.

@item 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.

@item void *dlfo_map_end
The end address of the mapping.

@item struct link_map *dlfo_link_map
This member contains a pointer to the link map of the object.

@item void *dlfo_eh_frame
This member contains a pointer to the exception handling data of the
object.  See @code{DLFO_EH_SEGMENT_TYPE} below.

@end table

This structure is a GNU extension.
@end deftp

@deftypevr Macro int DLFO_STRUCT_HAS_EH_DBASE
@standards{GNU, dlfcn.h}
On most targets, this macro is defined as @code{0}.  If it is defined to
@code{1}, @code{struct dl_find_object} contains an additional member
@code{dlfo_eh_dbase} of type @code{void *}.  It is the base address for
@code{DW_EH_PE_datarel} DWARF encodings to this location.

This macro is a GNU extension.
@end deftypevr

@deftypevr Macro int DLFO_STRUCT_HAS_EH_COUNT
@standards{GNU, dlfcn.h}
On most targets, this macro is defined as @code{0}.  If it is defined to
@code{1}, @code{struct dl_find_object} contains an additional member
@code{dlfo_eh_count} of type @code{int}.  It is the number of exception
handling entries in the EH frame segment identified by the
@code{dlfo_eh_frame} member.

This macro is a GNU extension.
@end deftypevr

@deftypevr Macro int DLFO_EH_SEGMENT_TYPE
@standards{GNU, dlfcn.h}
On targets using DWARF-based exception unwinding, this macro expands to
@code{PT_GNU_EH_FRAME}.  This indicates that @code{dlfo_eh_frame} in
@code{struct dl_find_object} points to the @code{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.
@end deftypevr

@deftypefun {int} _dl_find_object (void *@var{address}, struct dl_find_object *@var{result})
@standards{GNU, dlfcn.h}
@safety{@mtsafe{}@assafe{}@acsafe{}}
On success, this function returns 0 and writes about the object
surrounding the address to @code{*@var{result}}.  On failure, -1 is
returned.

The @var{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, @code{_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 @var{address} is greater than or equal to
@code{@var{result}->dlfo_map_start} and less than
@code{@var{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 @var{address} in
@code{@var{result}->dlfo_eh_frame}.  If the platform uses DWARF
unwinding information, this is the in-memory address of the
@code{PT_GNU_EH_FRAME} segment.  See @code{DLFO_EH_SEGMENT_TYPE} above.
In case @var{address} resides in an object that lacks unwinding information,
the function still returns 0, but sets @code{@var{result}->dlfo_eh_frame}
to a null pointer.

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

This function is a GNU extension.
@end deftypefun

@node Dynamic Linker Hardening
@section Avoiding Unexpected Issues With Dynamic Linking

This section details recommendations for increasing application
robustness, by avoiding potential issues related to dynamic linking.
The recommendations have two main aims: reduce the involvement of the
dynamic linker in application execution after process startup, and
restrict the application to a dynamic linker feature set whose behavior
is more easily understood.

Key aspects of limiting dynamic linker usage after startup are: no use
of the @code{dlopen} function, disabling lazy binding, and using the
static TLS model.  More easily understood dynamic linker behavior
requires avoiding name conflicts (symbols and sonames) and highly
customizable features like the audit subsystem.

Note that while these steps can be considered a form of application
hardening, they do not guard against potential harm from accidental or
deliberate loading of untrusted or malicious code.  There is only
limited overlap with traditional security hardening for applications
running on GNU systems.

@subsection Restricted Dynamic Linker Features

Avoiding certain dynamic linker features can increase predictability of
applications and reduce the risk of running into dynamic linker defects.

@itemize @bullet
@item
Do not use the functions @code{dlopen}, @code{dlmopen}, or
@code{dlclose}.  Dynamic loading and unloading of shared objects
introduces substantial complications related to symbol and thread-local
storage (TLS) management.

@item
Without the @code{dlopen} function, @code{dlsym} and @code{dlvsym}
cannot be used with shared object handles.  Minimizing the use of both
functions is recommended.  If they have to be used, only the
@code{RTLD_DEFAULT} pseudo-handle should be used.

@item
Use the local-exec or initial-exec TLS models.  If @code{dlopen} is not
used, there are no compatibility concerns for initial-exec TLS.  This
TLS model avoids most of the complexity around TLS access.  In
particular, there are no TLS-related run-time memory allocations after
process or thread start.

If shared objects are expected to be used more generally, outside the
hardened, feature-restricted context, lack of compatibility between
@code{dlopen} and initial-exec TLS could be a concern.  In that case,
the second-best alternative is to use global-dynamic TLS with GNU2 TLS
descriptors, for targets that fully implement them, including the fast
path for access to TLS variables defined in the initially loaded set of
objects.  Like initial-exec TLS, this avoids memory allocations after
thread creation, but only if the @code{dlopen} function is not used.

@item
Do not use lazy binding.  Lazy binding may require run-time memory
allocation, is not async-signal-safe, and introduces considerable
complexity.

@item
Make dependencies on shared objects explicit.  Do not assume that
certain libraries (such as @code{libc.so.6}) are always loaded.
Specifically, if a main program or shared object references a symbol,
create an ELF @code{DT_NEEDED} dependency on that shared object, or on
another shared object that is documented (or otherwise guaranteed) to
have the required explicit dependency.  Referencing a symbol without a
matching link dependency results in underlinking, and underlinked
objects cannot always be loaded correctly: Initialization of objects may
not happen in the required order.

@item
Do not create dependency loops between shared objects (@code{libA.so.1}
depending on @code{libB.so.1} depending on @code{libC.so.1} depending on
@code{libA.so.1}).  @Theglibc{} has to initialize one of the objects in
the cycle first, and the choice of that object is arbitrary and can
change over time.  The object which is initialized first (and other
objects involved in the cycle) may not run correctly because not all of
its dependencies have been initialized.

Underlinking (see above) can hide the presence of cycles.

@item
Limit the creation of indirect function (IFUNC) resolvers.  These
resolvers run during relocation processing, when @theglibc{} is not in
a fully consistent state.  If you write your own IFUNC resolvers, do
not depend on external data or function references in those resolvers.

@item
Do not use the audit functionality (@code{LD_AUDIT}, @code{DT_AUDIT},
@code{DT_DEPAUDIT}).  Its callback and hooking capabilities introduce a
lot of complexity and subtly alter dynamic linker behavior in corner
cases even if the audit module is inactive.

@item
Do not use symbol interposition.  Without symbol interposition, the
exact order in which shared objects are searched are less relevant.

Exceptions to this rule are copy relocations (see the next item), and
vague linkage, as used by the C++ implementation (see below).

@item
One potential source of symbol interposition is a combination of static
and dynamic linking, namely linking a static archive into multiple
dynamic shared objects.  For such scenarios, the static library should
be converted into its own dynamic shared object.

A different approach to this situation uses hidden visibility for
symbols in the static library, but this can cause problems if the
library does not expect that multiple copies of its code coexist within
the same process, with no or partial sharing of state.

@item
If you use shared objects that are linked with @option{-Wl,-Bsymbolic}
(or equivalent) or use protected visibility, the code for the main
program must be built as @option{-fpic} or @option{-fPIC} to avoid
creating copy relocations (and the main program must not use copy
relocations for other reasons).  Using @option{-fpie} or @option{-fPIE}
is not an alternative to PIC code in this context.

@item
Be careful about explicit section annotations.  Make sure that the
target section matches the properties of the declared entity (e.g., no
writable objects in @code{.text}).

@item
Ensure that all assembler or object input files have the recommended
security markup, particularly for non-executable stack.

@item
Avoid using non-default linker flags and features.  In particular, do
not use the @code{DT_PREINIT_ARRAY} dynamic tag, and do not flag
objects as @code{DF_1_INITFIRST}.  Do not change the default linker
script of BFD ld.  Do not override ABI defaults, such as the dynamic
linker path (with @option{--dynamic-linker}).

@item
Some features of @theglibc{} indirectly depend on run-time code loading
and @code{dlopen}.  Use @code{iconv_open} with built-in converters only
(such as @code{UTF-8}).  Do not use NSS functionality such as
@code{getaddrinfo} or @code{getpwuid_r} unless the system is configured
for built-in NSS service modules only (see below).
@end itemize

Several considerations apply to ELF constructors and destructors.

@itemize @bullet
@item
The dynamic linker does not take constructor and destructor priorities
into account when determining their execution order.  Priorities are
only used by the link editor for ordering execution within a
completely linked object.  If a dynamic shared object needs to be
initialized before another object, this can be expressed with a
@code{DT_NEEDED} dependency on the object that needs to be initialized
earlier.

@item
The recommendations to avoid cyclic dependencies and symbol
interposition make it less likely that ELF objects are accessed before
their ELF constructors have run.  However, using @code{dlsym} and
@code{dlvsym}, it is still possible to access uninitialized facilities
even with these restrictions in place.  (Of course, access to
uninitialized functionality is also possible within a single shared
object or the main executable, without resorting to explicit symbol
lookup.)  Consider using dynamic, on-demand initialization instead.  To
deal with access after de-initialization, it may be necessary to
implement special cases for that scenario, potentially with degraded
functionality.

@item
Be aware that when ELF destructors are executed, it is possible to
reference already-deconstructed shared objects.  This can happen even in
the absence of @code{dlsym} and @code{dlvsym} function calls, for
example if client code using a shared object has registered callbacks or
objects with another shared object.  The ELF destructor for the client
code is executed before the ELF destructor for the shared objects that
it uses, based on the expected dependency order.

@item
If @code{dlopen} and @code{dlmopen} are not used, @code{DT_NEEDED}
dependency information is complete, and lazy binding is disabled, the
execution order of ELF destructors is expected to be the reverse of the
ELF constructor order.  However, two separate dependency sort operations
still occur.  Even though the listed preconditions should ensure that
both sorts produce the same ordering, it is recommended not to depend on
the destructor order being the reverse of the constructor order.
@end itemize

The following items provide C++-specific guidance for preparing
applications.  If another programming language is used and it uses these
toolchain features targeted at C++ to implement some language
constructs, these restrictions and recommendations still apply in
analogous ways.

@itemize @bullet
@item
C++ inline functions, templates, and other constructs may need to be
duplicated into multiple shared objects using vague linkage, resulting
in symbol interposition.  This type of symbol interposition is
unproblematic, as long as the C++ one definition rule (ODR) is followed,
and all definitions in different translation units are equivalent
according to the language C++ rules.

@item
Be aware that under C++ language rules, it is unspecified whether
evaluating a string literal results in the same address for each
evaluation.  This also applies to anonymous objects of static storage
duration that GCC creates, for example to implement the compound
literals C++ extension.  As a result, comparing pointers to such
objects, or using them directly as hash table keys, may give unexpected
results.

By default, variables of block scope of static storage have consistent
addresses across different translation units, even if defined in
functions that use vague linkage.

@item
Special care is needed if a C++ project uses symbol visibility or
symbol version management (for example, the GCC @samp{visibility}
attribute, the GCC @option{-fvisibility} option, or a linker version
script with the linker option @option{--version-script}).  It is
necessary to ensure that the symbol management remains consistent with
how the symbols are used.  Some C++ constructs are implemented with
the help of ancillary symbols, which can make complicated to achieve
consistency.  For example, an inline function that is always inlined
into its callers has no symbol footprint for the function itself, but
if the function contains a variable of static storage duration, this
variable may result in the creation of one or more global symbols.
For correctness, such symbols must be visible and bound to the same
object in all other places where the inline function may be called.
This requirement is not met if the symbol visibility is set to hidden,
or if symbols are assigned a textually different symbol version
(effectively creating two distinct symbols).

Due to the complex interaction between ELF symbol management and C++
symbol generation, it is recommended to use C++ language features for
symbol management, in particular inline namespaces.

@item
The toolchain and dynamic linker have multiple mechanisms that bypass
the usual symbol binding procedures.  This means that the C++ one
definition rule (ODR) still holds even if certain symbol-based isolation
mechanisms are used, and object addresses are not shared across
translation units with incompatible type definitions.

This does not matter if the original (language-independent) advice
regarding symbol interposition is followed.  However, as the advice may
be difficult to implement for C++ applications, it is recommended to
avoid ODR violations across the entire process image.  Inline namespaces
can be helpful in this context because they can be used to create
distinct ELF symbols while maintaining source code compatibility at the
C++ level.

@item
Be aware that as a special case of interposed symbols, symbols with the
@code{STB_GNU_UNIQUE} binding type do not follow the usual ELF symbol
namespace isolation rules: such symbols bind across @code{RTLD_LOCAL}
boundaries.  Furthermore, symbol versioning is ignored for such symbols;
they are bound by symbol name only.  All their definitions and uses must
therefore be compatible.  Hidden visibility still prevents the creation
of @code{STB_GNU_UNIQUE} symbols and can achieve isolation of
incompatible definitions.

@item
C++ constructor priorities only affect constructor ordering within one
shared object.  Global constructor order across shared objects is
consistent with ELF dependency ordering if there are no ELF dependency
cycles.

@item
C++ exception handling and run-time type information (RTTI), as
implemented in the GNU toolchain, is not address-significant, and
therefore is not affected by the symbol binding behaviour of the dynamic
linker.  This means that types of the same fully-qualified name (in
non-anonymous namespaces) are always considered the same from an
exception-handling or RTTI perspective.  This is true even if the type
information object or vtable has hidden symbol visibility, or the
corresponding symbols are versioned under different symbol versions, or
the symbols are not bound to the same objects due to the use of
@code{RTLD_LOCAL} or @code{dlmopen}.

This can cause issues in applications that contain multiple incompatible
definitions of the same type.  Inline namespaces can be used to create
distinct symbols at the ELF layer, avoiding this type of issue.

@item
C++ exception handling across multiple @code{dlmopen} namespaces may
not work, particular with the unwinder in GCC versions before 12.
Current toolchain versions are able to process unwinding tables across
@code{dlmopen} boundaries.  However, note that type comparison is
name-based, not address-based (see the previous item), so exception
types may still be matched in unexpected ways.  An important special
case of exception handling, invoking destructors for variables of block
scope, is not impacted by this RTTI type-sharing.  Likewise, regular
virtual member function dispatch for objects is unaffected (but still
requires that the type definitions match in all directly involved
translation units).

Once more, inline namespaces can be used to create distinct ELF symbols
for different types.

@item
Although the C++ standard requires that destructors for global objects
run in the opposite order of their constructors, the Itanium C++ ABI
requires a different destruction order in some cases.  As a result, do
not depend on the precise destructor invocation order in applications
that use @code{dlclose}.

@item
Registering destructors for later invocation allocates memory and may
silently fail if insufficient memory is available.  As a result, the
destructor is never invoked.  This applies to all forms of destructor
registration, with the exception of thread-local variables (see the next
item).  To avoid this issue, ensure that such objects merely have
trivial destructors, avoiding the need for registration, and deallocate
resources using a different mechanism (for example, from an ELF
destructor).

@item
A similar issue exists for @code{thread_local} variables with thread
storage duration of types that have non-trivial destructors.  However,
in this case, memory allocation failure during registration leads to
process termination.  If process termination is not acceptable, use
@code{thread_local} variables with trivial destructors only.
Functions for per-thread cleanup can be registered using
@code{pthread_key_create} (globally for all threads) and activated
using @code{pthread_setspecific} (on each thread).  Note that a
@code{pthread_key_create} call may still fail (and
@code{pthread_create} keys are a limited resource in @theglibc{}), but
this failure can be handled without terminating the process.
@end itemize

@subsection Producing Matching Binaries

This subsection recommends tools and build flags for producing
applications that meet the recommendations of the previous subsection.

@itemize @bullet
@item
Use BFD ld (@command{bfd.ld}) from GNU binutils to produce binaries,
invoked through a compiler driver such as @command{gcc}.  The version
should be not too far ahead of what was current when the version of
@theglibc{} was first released.

@item
Do not use a binutils release that is older than the one used to build
@theglibc{} itself.

@item
Compile with @option{-ftls-model=initial-exec} to force the initial-exec
TLS model.

@item
Link with @option{-Wl,-z,now} to disable lazy binding.

@item
Link with @option{-Wl,-z,relro} to enable RELRO (which is the default on
most targets).

@item
Specify all direct shared objects dependencies using @option{-l} options
to avoid underlinking.  Rely on @code{.so} files (which can be linker
scripts) and searching with the @option{-l} option.  Do not specify the
file names of shared objects on the linker command line.

@item
Consider using @option{-Wl,-z,defs} to treat underlinking as an error
condition.

@item
When creating a shared object (linked with @option{-shared}), use
@option{-Wl,-soname,lib@dots{}} to set a soname that matches the final
installed name of the file.

@item
Do not use the @option{-rpath} linker option.  (As explained below, all
required shared objects should be installed into the default search
path.)

@item
Use @option{-Wl,--error-rwx-segments} and @option{-Wl,--error-execstack} to
instruct the link editor to fail the link if the resulting final object
would have read-write-execute segments or an executable stack.  Such
issues usually indicate that the input files are not marked up
correctly.

@item
Ensure that for each @code{LOAD} segment in the ELF program header, file
offsets, memory sizes, and load addresses are multiples of the largest
page size supported at run time.  Similarly, the start address and size
of the @code{GNU_RELRO} range should be multiples of the page size.

Avoid creating gaps between @code{LOAD} segments.  The difference
between the load addresses of two subsequent @code{LOAD} segments should
be the size of the first @code{LOAD} segment.  (This may require linking
with @option{-Wl,-z,noseparate-code}.)

This may not be possible to achieve with the currently available link
editors.

@item
If the multiple-of-page-size criterion for the @code{GNU_RELRO} region
cannot be achieved, ensure that the process memory image right before
the start of the region does not contain executable or writable memory.
@c https://sourceware.org/pipermail/libc-alpha/2022-May/138638.html
@end itemize

@subsection Checking Binaries

In some cases, if the previous recommendations are not followed, this
can be determined from the produced binaries.  This section contains
suggestions for verifying aspects of these binaries.

@itemize @bullet
@item
To detect underlinking, examine the dynamic symbol table, for example
using @samp{readelf -sDW}.  If the symbol is defined in a shared object
that uses symbol versioning, it must carry a symbol version, as in
@samp{pthread_kill@@GLIBC_2.34}.

@item
Examine the dynamic segment with @samp{readelf -dW} to check that all
the required @code{NEEDED} entries are present.  (It is not necessary to
list indirect dependencies if these dependencies are guaranteed to
remain during the evolution of the explicitly listed direct
dependencies.)

@item
The @code{NEEDED} entries should not contain full path names including
slashes, only @code{sonames}.

@item
For a further consistency check, collect all shared objects referenced
via @code{NEEDED} entries in dynamic segments, transitively, starting at
the main program.  Then determine their dynamic symbol tables (using
@samp{readelf -sDW}, for example).  Ideally, every symbol should be
defined at most once, so that symbol interposition does not happen.

If there are interposed data symbols, check if the single interposing
definition is in the main program.  In this case, there must be a copy
relocation for it.  (This only applies to targets with copy relocations.)

Function symbols should only be interposed in C++ applications, to
implement vague linkage.  (See the discussion in the C++ recommendations
above.)

@item
Using the previously collected @code{NEEDED} entries, check that the
dependency graph does not contain any cycles.

@item
The dynamic segment should also mention @code{BIND_NOW} on the
@code{FLAGS} line or @code{NOW} on the @code{FLAGS_1} line (one is
enough).

@item
For shared objects (not main programs), if the program header has a
@code{PT_TLS} segment, the dynamic segment (as shown by @samp{readelf
-dW}) should contain the @code{STATIC_TLS} flag on the @code{FLAGS}
line.

If @code{STATIC_TLS} is missing in shared objects, ensure that the
appropriate relocations for GNU2 TLS descriptors are used (for example,
@code{R_AARCH64_TLSDESC} or @code{R_X86_64_TLSDESC}).

@item
There should not be a reference to the symbols @code{__tls_get_addr},
@code{__tls_get_offset}, @code{__tls_get_addr_opt} in the dynamic symbol
table (in the @samp{readelf -sDW} output).  Thread-local storage must be
accessed using the initial-exec (static) model, or using GNU2 TLS
descriptors.

@item
Likewise, the functions @code{dlopen}, @code{dlmopen}, @code{dlclose}
should not be referenced from the dynamic symbol table.

@item
For shared objects, there should be a @code{SONAME} entry that matches
the file name (the base name, i.e., the part after the slash).  The
@code{SONAME} string must not contain a slash @samp{/}.

@item
For all objects, the dynamic segment (as shown by @samp{readelf -dW})
should not contain @code{RPATH} or @code{RUNPATH} entries.

@item
Likewise, the dynamic segment should not show any @code{AUDIT},
@code{DEPAUDIT}, @code{AUXILIARY}, @code{FILTER}, or
@code{PREINIT_ARRAY} tags.

@item
If the dynamic segment contains a (deprecated) @code{HASH} tag, it
must also contain a @code{GNU_HASH} tag.

@item
The @code{INITFIRST} flag (undeer @code{FLAGS_1}) should not be used.

@item
The program header must not have @code{LOAD} segments that are writable
and executable at the same time.

@item
All produced objects should have a @code{GNU_STACK} program header that
is not marked as executable.  (However, on some newer targets, a
non-executable stack is the default, so the @code{GNU_STACK} program
header is not required.)
@end itemize

@subsection Run-time Considerations

In addition to preparing program binaries in a recommended fashion, the
run-time environment should be set up in such a way that problematic
dynamic linker features are not used.

@itemize @bullet
@item
Install shared objects using their sonames in a default search path
directory (usually @file{/usr/lib64}).  Do not use symbolic links.
@c This is currently not standard practice.

@item
The default search path must not contain objects with duplicate file
names or sonames.

@item
Do not use environment variables (@code{LD_@dots{}} variables such as
@code{LD_PRELOAD} or @code{LD_LIBRARY_PATH}, or @code{GLIBC_TUNABLES})
to change default dynamic linker behavior.

@item
Do not install shared objects in non-default locations.  (Such locations
are listed explicitly in the configuration file for @command{ldconfig},
usually @file{/etc/ld.so.conf}, or in files included from there.)

@item
In relation to the previous item, do not install any objects it
@code{glibc-hwcaps} subdirectories.

@item
Do not configure dynamically-loaded NSS service modules, to avoid
accidental internal use of the @code{dlopen} facility.  The @code{files}
and @code{dns} modules are built in and do not rely on @code{dlopen}.

@item
Do not truncate and overwrite files containing programs and shared
objects in place, while they are used.  Instead, write the new version
to a different path and use @code{rename} to replace the
already-installed version.

@item
Be aware that during a component update procedure that involves multiple
object files (shared objects and main programs), concurrently starting
processes may observe an inconsistent combination of object files (some
already updated, some still at the previous version).  For example,
this can happen during an update of @theglibc{} itself.
@end itemize

@c FIXME these are undocumented:
@c dladdr
@c dladdr1
@c dlclose
@c dlerror
@c dlmopen
@c dlopen
@c dlsym
@c dlvsym