the resolution to Austin Group issue #411 defined new semantics for
the posix_spawn dup2 file action in the (previously useless) case
where src and dest fd are equal. future issues will require the dup2
file action to remove the close-on-exec flag. without this change,
passing fds to a child with posix_spawn while avoiding fd-leak races
in a multithreaded parent required a complex dance with temporary fds.

based on patch by Petr Skocik. changes were made to preserve the
80-column formatting of the function and to remove code that became
unreachable as a result of the new functionality.

commit 06fbefd1 (first included in
release 1.1.17) introduced this regression.

patch by Adrian Bunk. it fixes the regression in all cases, but
spuriously prevents use of the clz instruction on very old compiler
versions that don't define __ARM_ARCH. this may be fixed in a more
general way at some point in the future. it also omits thumb1 logic
since building as thumb1 code is currently not supported.

commit 8c4be3e2 was written to
preclude the GLOB_PERIOD extension from matching these directory
entries, but also precluded literal matches.

adjust the check that excludes . and .. to check whether the
GLOB_PERIOD flag is in effect, so that it cannot alter behavior in
cases governed by the standard, and also don't exclude . or .. in any
case where normal glob behavior (fnmatch's FNM_PERIOD flag) would have
included one or both of them (patterns such as ".*").

it's still not clear whether this is the preferred behavior for
GLOB_PERIOD, but at least it's clear that it can no longer break
applications which are not relying on quirks of a nonstandard feature.

execvpe stack-allocates a buffer used to hold the full path
(combination of a PATH entry and the program name)
while searching through $PATH, so at least
NAME_MAX+PATH_MAX is needed.

The stack size can be made conditionally smaller
(the current 1024 appears appropriate)
should this larger size be burdensome in those situations.

MAXADDRS was chosen not to need enforcement, but the logic used to
compute it assumes the answers received match the RR types of the
queries. specifically, it assumes that only one replu contains A
record answers. if the replies to both the A and the AAAA query have
their answer sections filled with A records, MAXADDRS can be exceeded
and clobber the stack of the calling function.

this bug was found and reported by Felix Wilhelm.

the rightmost '/' character is not necessarily the delimiter before
the basename; it could be a spurious trailing character on the
directory name.

this change does not introduce any normalization of pathnames or
stripping of trailing slashes, contrary to at least glibc and perhaps
other implementations; it jusst prevents their presence from breaking
things. whether further changes should be made is an open question
that may depend on conformance and/or application compatibility
considerations.

based loosely on patch by Joakim Sindholt.

calling __unlock on t->exitlock is not valid because __unlock reads
the waiters count after making the atomic store that could allow
pthread_exit to continue and unmap the thread's stack and the object t
points to. for now, inline the __unlock logic with an unconditional
futex wake operation so that the waiters count is not needed.

once __lock/__unlock have been made safe for self-synchronized
destruction, we could switch back to using them.

the freebsd fma code failed to raise underflow exception in some
cases in nearest rounding mode (affects fmal too) e.g.

  fma(-0x1p-1000, 0x1.000001p-74, 0x1p-1022)

and the inexact exception may be raised spuriously since the fenv
is not saved/restored around the exact multiplication algorithm
(affects x86 fma too).

another issue is that the underflow behaviour when the rounded result
is the minimal normal number is target dependent, ieee754 allows two
ways to raise underflow for inexact results: raise if the result before
rounding is in the subnormal range (e.g. aarch64, arm, powerpc) or if
the result after rounding with infinite exponent range is in the
subnormal range (e.g. x86, mips, sh).

to avoid all these issues the algorithm was rewritten with mostly int
arithmetics and float arithmetics is only used to get correct rounding
and raise exceptions according to the behaviour of the target without
any fenv.h dependency. it also unifies x86 and non-x86 fma.

fmaf is not affected, fmal need to be fixed too.

this algorithm depends on a_clz_64 and it required a few spurious
instructions to make sure underflow exception is raised in a particular
corner case. (normally FORCE_EVAL(tiny*tiny) would be used for this,
but on i386 gcc is broken if the expression is constant
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57245
and there is no easy portable fix for the macro.)

this is for consistency with the way it's done in in the dynamic
linker, avoiding a deprecated C feature (non-prototype function
types), and improving code generation. GCC unnecessarily uses the
variadic calling convention (e.g. clearing rax on x86_64) when making
a call where the argument types are not known for compatibility with
wrong code which calls variadic functions this way. (C on the other
hand is clear that such calls have undefined behavior.)

this is a subtle issue with how the assembler/linker work. for the adr
pseudo-instruction used to find __hwcap, the assembler in thumb mode
generates a 16-bit thumb add instruction which can only represent
word-aligned addresses, despite not knowing the alignment of the
label. if the setjmp function is assigned a non-multiple-of-4 address
at link time, the load then loads from the wrong address (the last
instruction rather than the data containing the offset) and ends up
reading nonsense instead of the value of __hwcap. this in turn causes
the checks for floating-point/vector register sets (e.g. IWMMX) to
evaluate incorrectly, crashing when setjmp/longjmp try to save/restore
those registers.

fix based on bug report by Felix Hädicke.

patch by Jörg Krause.

MAP_{NORESERVE,LOCKED} have different values on powerpc than in generic.

under some conditions, the mmap syscall wrongly fails with EPERM
instead of ENOMEM when memory is exhausted; this is probably the
result of the kernel trying to fit the allocation somewhere that
crosses into the kernel range or below mmap_min_addr. in any case it's
a conformance bug, so work around it. for now, only handle the case of
anonymous mappings with no requested address; in other cases EPERM may
be a legitimate error.

this indirectly fixes the possibility of malloc failing with the wrong
errno value.

GLOB_PERIOD is a gnu extension, and GNU glob does not seem to honor it
except in the last path component. it's not clear whether this a bug
or intentional, but it seems reasonable that it should exclude the
special entries . and .. when walking.

changes based on report and analysis by Julien Ramseier.

some applications use getservbyport to find port numbers that are not
assigned to a service; if getservbyport always succeeds with a numeric
string as the result, they fail to find any available ports.

POSIX doesn't seem to mandate the behavior one way or another. it
specifies an abstract service database, which an implementation could
define to include numeric port strings, but it makes more sense to
align behavior with traditional implementations.

based on patch by A. Wilcox. the original patch only changed
getservbyport[_r]. to maintain a consistent view of the "service
database", I have also modified getservbyname[_r] to exclude numeric
port strings.

if the parent thread was able to set the new thread's priority before
it reached the check for 'startlock', the new thread failed to restore
its signal mask and thus ran with all signals blocked.

concept for patch by Sergei, who reported the issue; unnecessary
changes were removed and comments added since the whole 'startlock'
thing is non-idiomatic and confusing. eventually it should be replaced
with use of idiomatic synchronization primitives.

most of the found naming differences don't matter to musl, because
internally it unifies the syscall names that vary across targets,
but for external code the names should match the kernel uapi.

aarch64:
	__NR_fstatat is called __NR_newfstatat in linux.
	__NR_or1k_atomic got mistakenly copied from or1k.
arm:
	__NR_arm_sync_file_range is an alias for __NR_sync_file_range2
	__NR_fadvise64_64 is called __NR_arm_fadvise64_64 in linux,
	the old non-arm name is kept too, it should not cause issues.
	(powerpc has similar nonstandard fadvise and it uses the
	normal name.)
i386:
	__NR_madvise1 was removed from linux in commit
	303395ac3bf3e2cb488435537d416bc840438fcb 2011-11-11
microblaze:
	__NR_fadvise, __NR_fstatat, __NR_pread, __NR_pwrite
	had different name in linux.
mips:
	__NR_fadvise, __NR_fstatat, __NR_pread, __NR_pwrite, __NR_select
	had different name in linux.
mipsn32:
	__NR_fstatat is called __NR_newfstatat in linux.
or1k:
	__NR__llseek is called __NR_llseek in linux.
	the old name is kept too because that's the name musl uses
	internally.
powerpc:
	__NR_{get,set}res{gid,uid}32 was never present in powerpc linux.
	__NR_timerfd was briefly defined in linux but then got renamed.

this is mandated by C and POSIX standards and is in accordance with
glibc behavior.

Reported by Leah Neukirchen.

This aligns clearenv with the Linux man page by setting 'environ'
rather than '*environ' to NULL, and stops it from leaking entries
allocated by the libc.

Rewrite environment access functions to slim down code, fix bugs and
avoid invoking undefined behavior.

* avoid using int-typed iterators where size_t would be correct;
* use strncmp instead of memcmp consistently;
* tighten prologues by invoking __strchrnul;
* handle NULL environ.

putenv:
* handle "=value" input via unsetenv too (will return -1/EINVAL);
* rewrite and simplify __putenv; fix the leak caused by failure to
  deallocate entry added by preceding setenv when called from putenv.

setenv:
* move management of libc-allocated entries to this translation unit,
  and use no-op weak symbols in putenv/unsetenv;

unsetenv:
* rewrite; this fixes UB caused by testing a free'd pointer against
  NULL on entry to subsequent loops.

Not changed:
Failure to extend allocation tracking array (previously __env_map, now
env_alloced) is ignored rather than causing to report -1/ENOMEM to the
caller; the worst-case consequence is leaking this allocation when it
is removed or replaced in a subsequent environment access.

Initially UB in unsetenv was reported by Alexander Cherepanov.
Using a weak alias to avoid pulling in malloc via unsetenv was
suggested by Rich Felker.

the DFA table controlling accepted ranges for the f4 prefix used an
incorrect upper bound of 0xa0 where it should have been 0x90, allowing
such sequences to be accepted and decoded as non-Unicode-scalar values
0x110000 through 0x11ffff.

the value computed as an output limit that bounds the amount of input
consumed below the input limit was incorrectly being used as the
actual amount of input consumed. instead, compute the actual amount of
input consumed as a difference of pointers before and after the
conversion.

patch by Mikhail Kremnyov.

Glibc renamed the linux uapi HWCAP_* macros to HWCAP_ARM_*
so have both variants in case some code depends on it.
(The HWCAP2_ macros are not defined in glibc currently so those
only have the linux uapi variant.)

counts leading zero bits of a 64bit int, undefined on zero input.
(has nothing to do with atomics, added to atomic.h so target specific
helper functions are together.)

there is a logarithmic generic implementation and another in terms of
a 32bit a_clz_32 on targets where that's available.

aligned with linux arch/s390/include/asm/elf.h
(these macros should be exported into uapi, but they are not)

it's in the reserved namespace for the latter, where it seems it was
historically defined, and some programs expect to find it there.

patch by Pascal Cuoq.

It is possible for argv[0] to be a null pointer, but the __progname
variable is used to implement functions in src/legacy/err.c that do not
expect it to be null. It is also available to the user via the
program_invocation_name alias as a GNU extension, and the implementation
in Glibc initializes it to a pointer to empty string rather than NULL.

Since argv[0] is usually non-null and it's preferable to keep those
variables in BSS, implement the fallbacks in __init_libc, which also
allows to have an intermediate fallback to AT_EXECFN.

it is defined in linux asm/sockios.h since commit
ae40eb1ef30ab4120bd3c8b7e3da99ee53d27a23 (linux v2.6.22)
but was missing from musl by accident.

in musl the sockios macros are exposed in sys/ioctl.h together
with other ioctl requests instead of in sys/socket.h because of
namespace rules. (glibc has them in sys/socket.h under _GNU_SOURCE.)

commit c002668e inadvertently moved
the check for unflushed write buffer outside of the scope of the
existing lock.

Due to a missing ":" in an asm() statement, the "memory" clobber is
considered by gcc as an input operand and not a clobber, which causes a
build failure.

passing to pthread_join the id of a thread which is not joinable
results in undefined behavior.

in principle the check to trap does not necessarily work if
pthread_detach was called after thread creation, since no effort is
made here to synchronize access to t->detached, but the check is
well-defined and harmless for callers which did not invoke UB, and
likely to help catch erroneous code that would otherwise mysteriously
hang.

patch by William Pitcock.

The TOC pointer is constant within a single dso, but needs to be saved
and restored around cross-dso calls.  The PLT stub saves it to the
caller's stack frame, and the linker adds code to the caller to restore
it.

With a local call, as within a single dso or with static linking, this
doesn't happen and the TOC pointer is always in r2.  Therefore,
setjmp/longjmp need to save/restore the TOC pointer from/to different
locations depending on whether the call to setjmp was a local or non-local
call.

It is always safe for longjmp to restore to both r2 and the caller's stack.
If the call to setjmp was local, and only r2 matters and the stack location
will be ignored, but is required by the ABI to be reserved for the TOC
pointer.  If the call was non-local, then only the stack location matters,
and whatever is restored into r2 will be clobbered anyway when the caller
reloads r2 from the stack.

A little extra care is required for sigsetjmp, because it uses setjmp
internally.  After the second return from this setjmp call, r2 will contain
the caller's TOC pointer instead of libc's TOC pointer.  We need to save
and restore the correct libc pointer before we can tail call to
__sigsetjmp_tail.