commit b91cdbe2, in fixing another
issue, changed the logic for how alt-form octal adds the leading zero
to adjust the precision rather than using a prefix character. this
wrongly suppressed the zero flag by mimicing an explicit precision
given by the format string. switch back to using a prefix character.

based on bug report and patch by Dmitry V. Levin, but simplified.

this reverts commit 2c1f8fd5. without
the _Noreturn attribute, the compiler cannot use asserts to perform
reachability/range analysis. this leads to missed optimizations and
spurious warnings.

the original backtrace problem that prompted the removal of _Noreturn
was not clearly documented at the time, but it seems to happen only
when libc was built without -g, which also breaks many other
backtracing cases.

This makes the result consistent with sysconf(_SC_OPEN_MAX).

O_FSYNC was never defined and is legacy/wrong, nothing seems to use it.

the gnu config.sub script recognizes several mipsisa64* cpu types
that musl supports as mips64 targets.

there was a copy paste error that could cause large ulp errors
in atan2l, atanl, asinl and acosl on aarch64, mips64 and mipsn32.

(the implementation is from freebsd fdlibm, but the tail end
of the polynomial was wrong. 128 bit long double functions
are not yet tested so this went undetected.)

linux containers use separate mount namespace so the /proc
symlink might not point to the right device if the fd was
opened in the parent namespace, in this case return ENOENT.

userfaultfd, membarrier and mlock2 syscalls got wired up in linux
commit fbce3befd60d40639bf3c6b60f7477b2f988f92d

only matters if swapcontext is used in a signal handler running on an
altstack, new in linux commit 2a74213838104a41588d86fd5e8d344972891ace

for GPRS tunneling protocol, new in linux commit
459aa660eb1d8ce67080da1983bb81d716aa5a69

macros for qualcom ip router protocol, new in linux commit
bdabad3e363d825ddf9679dd431cca0b2c30f881

despite sh not generally using register-pair alignment for 64-bit
syscall arguments, there are arch-specific versions of the syscall
entry points for pread and pwrite which include a dummy argument for
alignment before the 64-bit offset argument.

revert commit 8c316e9e. it was wrong
and does not match how the kernel API works.

the FIXME comment here was overlooked at the time locale support was
added.

this code was already under #if 0, but could be confusing if a reader
didn't notice that, and it's almost surely full of bugs and/or
inconsistencies with the current code that uses the gethostbyname2_r
backend.

Since commit 2f853dd6, all generated
headers are generated under the obj directory, which is already ignored.

It's identical to the generic version, after evaluating the endian
preprocessor checks in the generic version.

modern compilers (for gcc, versions 4.8 and later) automatically
pre-include <stdc-predef.h> to obtain the values of certain predefined
macros specified by ISO C but which reflect properties of the library
implementation, not just the compiler. provide values indicating that
wchar_t is Unicode-encoded and that Annex F (IEEE floating point) is
supported unless the compiler indicates otherwise.

based on patch by Masanori Ogino.

these changes still do not yield a fully-conforming abort, but they
fix two known issues:

- per POSIX, termination via SIGKILL is not "abnormal", but both ISO C
  and POSIX require abort to yield abnormal termination.

- raising SIGKILL fails to do anything to pid 1 in some containers.

now, the trapping instruction produced by a_crash() is expected to
produce abnormal termination, without the risk of invoking a signal
handler since SIGILL and SIGSEGV are blocked, and _Exit, which
contains an infinite loop analogous to the one being removed from
abort itself, is used as a last resort.

this implementation still fails to produce an exit status as if the
process terminated via SIGABRT in cases where SIGABRT is blocked or
ignored, but fixing that is not easy; the obvious pseudo-solutions all
have subtle race conditions where a concurrent fork or exec can expose
incorrect signal state.

commit befa5866 performed this change
for struct definitions that did not also involve typedef, but omitted
the latter.

last time elf.h was thoroughly updated sh was not yet supported
so these processor specific e_flags were missing.

new regset in linux v3.18 for ptrace.

see
https://sourceware.org/ml/libc-alpha/2016-01/msg00822.html

marks different RLD_MAP for debugging PIE binaries.

following
http://www.sco.com/developers/gabi/latest/ch4.sheader.html

see
https://sourceware.org/ml/libc-alpha/2015-03/msg00580.html

following
https://sourceware.org/ml/libc-alpha/2014-05/msg00332.html

add ilp32 related relocs and alternative names for a few macros following
https://sourceware.org/ml/libc-alpha/2014-11/msg00455.html

see
https://sourceware.org/ml/libc-alpha/2013-11/msg00315.html
https://sourceware.org/ml/libc-alpha/2013-11/msg00314.html

following the corresponding binutils and glibc changes
https://sourceware.org/ml/binutils/2013-10/msg00372.html

it was changed to EM_OR1K in 200d1547
as that was meant to be the official name, but glibc and the latest
gabi spec still uses the EM_OPENRISC name:
http://www.sco.com/developers/gabi/latest/ch4.eheader.html
binutils defines both macros so we should do the same for backward
compatibility.

placing the opening brace on the same line as the struct keyword/tag
is the style I prefer and seems to be the prevailing practice in more
recent additions.

these changes were generated by the command:

find include/ arch/*/bits -name '*.h' \
-exec sed -i '/^struct [^;{]*$/{N;s/\n/ /;}' {} +

and subsequently checked by hand to ensure that the regex did not pick
up any false positives.

with this change, all three files are identical.

it seems it was a typo.

these were incorrectly using the generic definitions.

same changes to the defined macros as in powerpc and generic bits.