the nommu kernel shares memory when it can anyway for private
read-only maps, but semantically the map should be private. this can
make a difference when debugging breakpoints are to be used, in which
case the kernel may need to ensure that the mapping is not shared.

the new behavior matches how the kernel FDPIC loader maps the main
program and/or program interpreter (dynamic linker) binary.

also fix visibility of the glue function used.

this both allows removal of some of the main remaining uses of the
SHARED macro and clears one obstacle to static-linked dlopen support,
which may be added at some point in the future.

specialized single-TLS-module versions of __copy_tls and __reset_tls
are removed and replaced with code adapted from their dynamic-linked
versions, capable of operating on a whole chain of TLS modules, and
use of the dynamic linker's DSO chain (which contains large struct dso
objects) by these functions is replaced with a new chain of struct
tls_module objects containing only the information needed for
implementing TLS. this may also yield some performance benefit
initializing TLS for a new thread when a large number of modules
without TLS have been loaded, since since there is no need to walk
structures for modules without TLS.

use weak definitions that the dynamic linker can override instead of
preprocessor conditionals on SHARED so that the same libc start and
exit code can be used for both static and dynamic linking.

this was only a tiny optimization, and static-linked binaries should
not be calling __tls_get_addr anyway since the linker is supposed to
perform relaxation, resulting in use of the local-exec TLS model.

this is the first and simplest stage of removal of the SHARED macro,
which will eventually allow libc.a and libc.so to be produced from the
same object files.

the original motivation for these #ifdefs which are now being removed
was to allow building a static-only libc using a compiler that does
not support visibility. however, SHARED was the wrong condition to
test for this anyway; various assembly-language sources refer to
hidden symbols and declare them with the .hidden directive, making it
wrong to define the referenced symbols as non-hidden. if there is a
need in the future to build libc using compilers that lack visibility,
support could be moved to the build system or perhaps the __PIC__
macro could be checked instead of SHARED.

when adding the fdpic subarchs, the need for these sub files was
overlooked. thus setjmp and longjmp performed illegal instructions.

on linux/nommu, non-writable private mappings of files may actually
use memory shared with other processes or the fs cache. the old nommu
loader code (used when mmap with MAP_FIXED fails) simply wrote over
top of the original file mapping, possibly clobbering this shared
memory. no such breakage was observed in practice, but it should have
been possible.

the new code starts by mapping anonymous writable memory on archs that
might support nommu, then maps load segments over top of it, falling
back to read if MAP_FIXED fails. we use an anonymous map rather than a
writable file map to avoid reading more data from disk than needed.
since pages cannot be loaded lazily on fault, in case of large
data/bss, mapping the full file may read a lot of data that will
subsequently be thrown away when processing additional LOAD segments.
as a result, we cannot skip the first LOAD segment when operating in
this mode.

these changes affect only non-FDPIC nommu support.

it was wrongly returning a null pointer instead of an empty string.

these files are all accepted as legacy arm syntax when producing arm
code, but legacy syntax cannot be used for producing thumb2 with
access to the full ISA. even after switching to UAL, some asm source
files contain instructions which are not valid in thumb mode, so these
will need to be addressed separately.

the idea of the three-instruction sequence being removed was to be
able to return to thumb code when used on armv4t+ from a thumb caller,
but also to be able to run on armv4 without the bx instruction
available (in which case the low bit of lr would always be 0).
however, without compiler support for generating such a sequence from
C code, which does not exist and which there is unlikely to be
interest in implementing, there is little point in having it in the
asm, and it would likely be easier to add pre-armv4t support via
enhanced linker handling of R_ARM_V4BX than at the compiler level.

removing this code simplifies adding support for building libc in
thumb2-only form (for cortex-m).

the code to save/restore vfp registers needs to build even when the
configured target does not have fpu; this is because code using vfp
fpu (but with the standard soft-float EABI) may call a libc built for
a soft-float only, and the EABI considers these registers call-saved
when they exist. thus, extra directives are used to force the
assembler to allow vfp instructions and to avoid marking the resulting
object files as requiring vfp.

moving away from using hard-coded opcode words is necessary in order
to eventually support producing thumb2-only output for cortex-m.

conditional execution of these instructions based on hwcap flags was
already implemented. when building for arm (non-thumb) output, the
only currently-supported configuration, this commit does not change
the code emitted.

mrc/mcr p10 coprocessor mnemonics are deprecated by some
toolchains.

contrary to commit 9367fe92, all
relevant gas versions actually do support .syntax unified.

this function is used only as a weak definition for malloc, for static
linking in programs which do not call realloc or free. since it had
external linkage and was thereby exported in libc.so's dynamic symbol
table, --gc-sections was unable to drop it. this was merely an
oversight; there's no reason for it to be external, so make it static.

this allowing the linker to drop certain weak definitions that are
only used as dummies for static linking. they could be eliminated for
shared library builds using the preprocessor instead, but we are
trying to transition to using the same object files for shared and
static libc, so a link-time solution is preferable.

based on patch by Denys Vlasenko. sorting sections and common data
symbols by alignment acts as an approximation for optimal packing,
which the linker does not actually support.

based on patch by Denys Vlasenko. the original intent for using these
options was to enable linking optimizations. these are immediately
available for static linking applications to libc.a, and will also be
used for linking libc.so in a subsequent commit.

in addition to the original motives, this change works around a whole
class of toolchain bugs where the compiler generates relative address
expressions using a weak symbol and the assembler "optimizes out" the
relocation which should result by using the weak definition. (see gas
pr 18561 and gcc pr 66609, 68178, etc. for examples.) by having
different functions and data objects in their own sections, all
relative address expressions are cross-section and thus cannot be
resolved to constants until link time. this allows us to retain
support for affected compiler/assembler versions without invasive
and fragile source-level workarounds.

this conditional path was never tested because there are no compilers
that conform to annex g (none with _Imaginary_I).

this assumption is borderline-unsafe to begin with, and fails badly
with -ffunction-sections since the linker can move the callee
arbitrarily far away when it lies in a different section.

this way, overriding these variables on the make command line (or just
re-passing the originally-passed values when invoking make) won't
suppress use of the flags added by configure.

since mremap with the MREMAP_FIXED flag is an operation that unmaps
existing mappings, it needs to use the vm lock mechanism to ensure
that any in-progress synchronization operations using vm identities
from before the call have finished.

also, the variadic argument was erroneously being read even if the
MREMAP_FIXED flag was not passed. in practice this didn't break
anything, but it's UB and in theory LTO could turn it into a hard
error.

It's quite feasible for this to happen via MREMAP_MAYMOVE.

previously, only archs that needed to do stack cleanup defined a
__cp_cancel label for acting on cancellation in their syscall asm, and
a default definition was provided by a weak alias to __cancel, the C
function. this resulted in wrong codegen for arm on gcc versions
affected by pr 68178 and possibly similar issues (like pr 66609) on
other archs, and also created an inconsistency where the __cp_begin
and __cp_end labels were treated as const data but __cp_cancel was
treated as a function. this in turn caused incorrect code generation
on archs where function pointers point to function descriptors rather
than code (for now, only sh/fdpic).

using the actual mcontext_t definition rather than an overlaid pointer
array both improves correctness/readability and eliminates some ugly
hacks for archs with 64-bit registers bit 32-bit program counter.

also fix UB due to comparison of pointers not in a common array
object.

when a library being loaded has bss (i.e. data segment with
p_memsz>p_filesz), this region needs to be zeroed with a combination
of memset and/or mmap. the regular ELF loader always did this but the
FDPIC code path omitted it, leading to objects in bss having
uninitialized/junk contents.

getnameinfo() compares the size of the given struct sockaddr with
sizeof(struct sockaddr_in) and sizeof(struct sockaddr_in6) depending on
the net family. When you add a sockaddr of size sizeof(struct
sockaddr_storage) this function will fail because the size of the
sockaddr is too big. Change the check that it only fails if the size is
too small, but make it work when it is too big for example when someone
calls this function with a struct sockaddr_storage and its size.
This fixes a problem with IoTivity 1.0.0 and musl.

glibc and bionic are only failing if it is smaller, net/freebsd
implemented the != check.
Signed-off-by: NHauke Mehrtens <hauke@hauke-m.de>

previously, transient failures like fd exhaustion or other
resource-related errors were treated the same as non-existence of
these files, leading to fallbacks or false-negative results. in
particular:

- failure to open hosts resulted in fallback to dns, possibly yielding
  EAI_NONAME for a hostname that should be defined locally, or an
  unwanted result from dns that the hosts file was intended to
  replace.

- failure to open services resulted in EAI_SERVICE.

- failure to open resolv.conf resulted in querying localhost rather
  than the configured nameservers.

now, only permanent errors trigger the fallback behaviors above; all
other errors are reportable to the caller as EAI_SYSTEM.

the buffer enlargement logic here accounted for the terminating null
byte, but not for the possibility of hitting the delimiter in the
buffer-refill code path that uses getc_unlocked, in which case two
additional bytes (the delimiter and the null termination) are written
without another chance to enlarge the buffer.

this patch and the corresponding bug report are by Felix Janda.

the option to suppress executable stack tagging was placed in CFLAGS,
which is treated as optional and overridable by the build system. if a
user replaces CFLAGS after configure has run, it could get lost,
resulting in a libc.so that's flagged as needing executable stack,
which would cause the kernel to map the initial stack as executable.

move -Wa,--noexecstack to CFLAGS_C99FSE, the make variable used for
mandatory compiler options.

these per-target CFLAGS adjustments are mandatory additions to the
command line for building the affected targets, not part of the
user-provided CFLAGS for tuning. my intent was always that the
variable append operations would take place after user settings, but
when a variable is set on the command line, it overrides all
definitions in the makefile, including target-specific ones.

based on patch by Szabolcs Nagy.

Some armhf gcc toolchains (built with --with-float=hard but without
--with-fpu=vfp*) do not pass -mfpu=vfp to the assembler and then
binutils rejects the UAL mnemonics for VFP unless there is an .fpu vfp
directive in the asm source.

POSIX requires pthread_join to synchronize memory on success.  The
futex wait inside __timedwait_cp cannot handle this because it's not
called in all cases.  Also, in the case of a spurious wake, tid can
become zero between the wake and when the joining thread checks it.

when determining which module an address belongs to, all function
descriptor ranges must be checked first, in case the allocated memory
falls inside another module's memory range.

dladdr itself must also check addresses against function descriptors
before doing a best-match search against the symbol table. even when
doing the latter (e.g. for code addresses obtained from mcontext_t),
also check whether the best-match was a function, and if so, replace
the result with a function descriptor address. which is the nominal
"base address" of the function and which the caller needs if it
intends to subsequently call the matching function.

since commits 2907afb8 and
6fc30c24, __dls2 is no longer called
via symbol lookup, but instead uses relative addressing that needs to
be resolved at link time. on some linker versions, and/or if
-Bsymbolic-functions is not used, the linker may leave behind a
dynamic relocation, which is not suitable for bootstrapping the
dynamic linker, if the reference to __dls2 is marked hidden but the
definition is not actually hidden. correcting the definition to use
hidden visibility fixes the problem.

the static-PIE entry point rcrt1 was likewise affected and is also
fixed by this patch.

we need access to all instructions in order for runtime selection of
atomic model to work correctly. without this patch, some versions of
gcc instruct gas to reject instructions outside the target isa level.

other archs use asm for the thread pointer load, so making that asm
volatile is sufficient to inform the compiler that it has a "side
effect" (crashing or giving the wrong result if the thread pointer was
not yet initialized) that prevents reordering. however, powerpc and
or1k have dedicated general purpose registers for the thread pointer
and did not need to use any asm to access it; instead, "local register
variables with a specified register" were used. however, there is no
specification for ordering constraints on this type of usage, and
presumably use of the thread pointer could be reordered across its
initialization.

to impose an ordering, I have added empty volatile asm blocks that
produce the "local register variable with a specified register" as
an output constraint.

this builds on commits a603a75a and
0ba35d69 to ensure that a compiler
cannot conclude that it's valid to reorder the asm to a point before
the thread pointer is set up, or to treat the inline function as if it
were declared with attribute((const)).

other archs already use volatile asm for thread pointer loading.