mips32r6 and mips64r6 are actually new isas at both the asm source and
opcode levels (pre-r6 code cannot run on r6) and thus need to be
treated as a new subarch. the following changes are made, some of
which yield code generation improvements for non-r6 targets too:

- add subarch logic in configure script and reloc.h files for dynamic
  linker name.

- suppress use of .set mips2 asm directives (used to allow mips2
  atomic instructions on baseline mips1 builds; the kernel has to
  emulate them on mips1) except when actually needed. they cause wrong
  instruction encodings on r6, and pessimize inlining on at least some
  compilers.

- only hard-code sync instruction encoding on mips1.

- use "ZC" constraint instead of "m" constraint for llsc memory
  operands on r6, where the ll/sc instructions no longer accept full
  16-bit offsets.

- only hard-code rdhwr instruction encoding with .word on targets
  (pre-r2) where it may need trap-and-emulate by the kernel.
  otherwise, just use the instruction mnemonic, and allow an arbitrary
  destination register to be used.

the two/three/four byte memmem specializations are not prepared to
handle haystacks shorter than the needle; they unconditionally read at
least up to the needle length and subtract from the haystack length.
if the haystack is shorter, the remaining haystack length underflows
and produces an unbounded search which will eventually either crash or
find a spurious match.

the top-level memmem function attempted to avoid this case already by
checking for haystack shorter than needle, but it failed to re-check
after using memchr to remove the maximal prefix not containing the
first byte of the needle.

commits e24984ef and
16b55298 inadvertently disabled the
a_spin implementations for i386, x86_64, and x32 by defining a macro
named a_pause instead of a_spin. this should not have caused any
functional regression, but it inhibited cpu relaxation while spinning
for locks.

bug reported by George Kulakowski.

the comparison f->wpos > f->buf has undefined behavior when f->wpos is
a null pointer, despite the intuition (and actual compiler behavior,
for all known compilers) being that NULL > ptr is false for all valid
pointers ptr.

the purpose of the comparison is to determine if the write buffer is
non-empty, and the idiom used elsewhere for that is comparison against
f->wbase, which is either a null pointer when not writing, or equal to
f->buf when writing. in the former case, both f->wpos and f->wbase are
null; in the latter they are both non-null and point into the same
array.

allows the os to free the marked pages lazily on memory pressure.
expected to increase malloc performance.
new in linux commit 854e9ed09dedf0c19ac8640e91bcc74bc3f9e5c9

new flag for exclusive wakeup mode when an event source fd is attached
to multiple epoll fds but they should not all receive the events.
new in linux commit df0108c5da561c66c333bb46bfe3c1fc65905898

new socket options for setting classic or extended BPF program
for sockets in a SO_REUSEPORT group.  added in linux commit
538950a1b7527a0a52ccd9337e3fcd304f027f13

new in linux commit 715f504b118998c41a2079a17e16bf5a8a114885
same as IP_HDRINCL but for SOL_IPV6 sockets.

it was introduced for offloading copying between regular files
in linux commit 29732938a6289a15e907da234d6692a2ead71855

(microblaze and sh does not yet have the syscall number.)

add userfaultfd, membarrier and mlock2 system call numbers.

currently five targets use the same mman.h constants and the rest
share most constants too, so move them to sys/mman.h before the
bits/mman.h include where the differences can be corrected by
redefinition of the macros.

this fixes two minor bugs: POSIX_MADV_DONTNEED was wrong on most
targets (it should be the same as MADV_DONTNEED), and sh defined
the x86-only MAP_32BIT mmap flag.

the idiom fprintf(f, "%.*s", n, "") was wrongly used in vfwprintf as a
means of producing n spaces; instead it produces no output. the
correct form is fprintf(f, "%*s", n, ""), using width instead of
precision, since for %s the later is a maximum rather than a minimum.

these changes should not affect generated code, but they reflect that
the underlying objects operated on by a_cas_p are supposed to have
type volatile void *, not volatile long. in theory a compiler could
treat the effective type mismatch in the "m" memory operands as
undefined behavior.

apparently clang does not accept matching-register input and output
constraints that differ in size (32-bit vs 64-bit).

based on patch by Jaydeep Patil.

the SPE ABI may be compatible with soft-float, but actually making it
work requires some additional work, so for now it's best to make sure
broken builds don't happen.

Some PowerPC CPUs (e.g. Freescale MPC85xx) have a completely different
instruction set for floating point operations (SPE).
Executing regular PowerPC floating point instructions results in
"Illegal instruction" errors.

Make it possible to run these devices in soft-float mode.

This is the minimal fix for __putenv leaving a pointer to freed heap
storage in __env_map array, which could later on lead to errors such
as double-free.

patch by Mahesh Bodapati and Jaydeep Patil of Imagination
Technologies.

this change is made in preparation for adding the mips64 port, which
needs a 64-bit (and mips64-specific) form of the R_INFO macro, but
it's a better abstraction anyway.

based on part of the mips64 port patch by Mahesh Bodapati and Jaydeep
Patil of Imagination Technologies.

expf(-NAN) was treated as expf(-large) which unconditionally
returns +0, so special case +-NAN.
reported by Petr Hosek.

This brings the call to an actually usable speed.
Quick unscientific benchmark: 14ns : 102ns :: vDSO : syscall

This is a GNU extension, but a fairly minor one, for a system call that
otherwise has no libc wrapper.

This is a workaround to treat * as literal * at the start of a BRE.

Ideally ^ would be treated as an anchor at the start of any BRE
subexpression and similarly $ would be an anchor at the end of any
subexpression.  This is not required by the standard and hard to do
with the current code, but it's the existing practice.  If it is
changed, * should be treated as literal after such anchor as well.

commit 7eaa76fc made * invalid at
the start of a BRE subexpression, but it should be accepted as
literal * there according to the standard.

This patch does not fix subexpressions starting with ^*.

name_from_hosts failed to account for the possibility of an address
family error from name_from_numeric, wrongly counting such a return as
success and using the uninitialized address data as part of the
results passed up to the caller.

non-matching address family entries cannot simply be ignored or
results would be inconsistent with respect to whether AF_UNSPEC or a
specific address family is queried. instead, record that a
non-matching entry was seen, and fail the lookup with EAI_NONAME of no
matching-family entries are found.

at present this is done only for consistency, since this file defines
its own a_cas_p rather than using the new generic one from atomic.h
added in commit 225f6a6b. these
definitions may however be useful if we ever need to add other
pointer-sized atomic operations.

No current ports do this, but it will be useful for porting to 64-bit ll/sc
architectures, such as mips64 and powerpc64.

this follows the principle of having the source tree layout define
build semantics. it also makes it possible for crt/$(ARCH) to define
additional installable files, which may be needed for midipix and
other future targets.

the nt32 and nt64 archs will be provided by the midipix project for
building musl on top of its posix-like syscall layer for windows. at
present the needed arch files are in a separate repository, but having
the tuple matching in the upstream configure script should make it
possible to overlay the arch files without needing any further
patching.

commit e4355bd6 moved the math asm
from external source files to inline asm, but unfortunately, all
current releases of clang use the wrong inline asm constraint codes
for float and double ("w" and "P" instead of "t" and "w",
respectively). this patch adds detection for the bug in configure,
and, for now, just disables the affected asm on broken clang versions.

in order to take advantage of the fpu in -mfloat-abi=softfp mode, the
__VFP_FP__ (presence of vfp fpu) was checked instead of checking for
__ARM_PCS_VFP (hardfloat EABI variant). however, the latter macro is
the one that's actually specified by the ABI documents rather than
being compiler-specific, and should also be checked in case __VFP_FP__
is not defined on some compilers or some configurations.

these additions were made based on scanning commit authors since the
last update, at the time of the 1.1.7 release, and adding everyone
with either substantial code contributions or a pattern of ongoing
simple patch submission.

the dynamic linker was found to hang when used as the PT_INTERP, but
not when invoked as a command. the mechanism of this failure was not
determined, but the cause is clear:

commit 5552ce52 removed the SHARED
macro, but arch/sh/crt_arch.h is still using it to choose the right
form of the crt/ldso entry point code. moving the forced definition
from rcrt1.c to dlstart.c restores the old behavior. eventually the
logic should be changed to fully remove the SHARED macro or at least
rename it to something more reasonable.

commit 80fbaac4 broke all soft-float
archs, where gcc defines __GCC_IEC_559==0 because rounding modes and
exception flags are not supported. for now, just check for
__FAST_MATH__ as an indication of broken float. this won't detect all
possible misconfigurations but it probably catches the most common
one.

commit 2f853dd6 moved the error
handling for $(ARCH) not being set such that it applied to all
targets, including clean and distclean. previously these targets
worked even in an unconfigured tree. to restore the old behavior, make
most of the makefile body conditional on $(ARCH) being set/non-empty
and produce the error via a fake "all" target in the conditional
branch for the case where $(ARCH) is empty.