Unify signbitl implementations.
(Unclear why they were different, it was only the comments that
differed)

Add a comment describing why the generic version works for both 80-bit
and 128-bit double.

branches: 1.18.16;
Don't include sigtramp or foocontext routines in rumprun mode. They won't
work (since a rump kernel does not provide NetBSD _lwp syscalls nor
signals), but they cause unwanted symbols to be pulled in when the
resulting libc is linked.

problem reported by Robert Gifford on rumpkernel-users

Remove "objects built from C sources" comments. Everyone can see
they're built from C sources because the source files end in .c (???)

add fpgetprec/fpsetprec

Move AMD64's bswap64 implementation from libc to src/common and share it
with the kernel.

Rename ${SRCS.{alpha,arm,...}.gen} to ${LSRCS.{alpha,arm,...}.gen} because
they ended up in ${LSRCS}, not ${SRCS}. ${SRCS.*} namespace will be used
for more useful things. No functional changes intended.

build {frexp,ldexp,modf} in the "compat" subtree
(Drop the modf assembler version. We have the same code in libm.)
(Drop the ldexp inline-assembler version. The same code is in libm as
scalbn; the ldexp there is just a wrapper providing error handling.)

Push back the descriptions of NaN formats, and descriptions of the
distinction between signalling NaNs and quiet NaNs back into the
machine-dependent headers; treat the implementation of __nanf in the
same spirit.

IEEE 754 leaves the distinction between signalling NaNs and quiet NANs
to the implementation, and unlike our headers used to suggest they're
not identical in the interpretation of the fraction's MSb; in due
course, make those of hppa, mips, sh3, and sh5 reflect reality.

branches: 1.11.6;
* Turn isinf(3) and isnan(3) into C99-style macros.
* Make it possible for ports to override these (i.e., VAX).
* Remove isnanl(), which was internal to libc only.

Add C99 fpclassify(), isfinite(), isnormal(), and signbit() macros.

Fix tpyo. (ieee854 -> ieee754)

Add __infinityf and __infinityl, float respectively long double analogs
of __infinity.

* Since there is no single portable IEEE 754 format for a long double,
keep a common implementation of isinfl() and isnanl() to be used by
platforms where `long double' == `double'; move others into
machine-dependent code.
* In due course, consider __VFP_FP__ on arm.

Rework how dependency generation is performed:

* DPSRCS contains extra dependencies, but is _NOT_ added to CLEANFILES.
This is a change of behaviour. If a Makefile wants the clean semantics
it must specifically append to CLEANFILES.
Resolves PR toolchain/5204.

* To recap: .d (depend) files are generated for all files in SRCS and DPSRCS
that have a suffix of: .c .m .s .S .C .cc .cpp .cxx

* If YHEADER is set, automatically add the .y->.h to DPSRCS & CLEANFILES

* Ensure that ${OBJS} ${POBJS} ${LOBJS} ${SOBJS} *.d depend upon ${DPSRCS}

* Deprecate the (short lived) DEPENDSRCS


Update the various Makefiles to these new semantics; generally either
adding to CLEANFILES (because DPSRCS doesn't do that anymore), or replacing
specific .o dependencies with DPSRCS entries.

Tested with "make -j 8 distribution" and "make distribution".

Build _isinfl() and _isnanl().

Rename ieee754_function.c to function_ieee754.c, following the convention
used in exec_format.c, loadfile_format.c, subsys_machdep.c etc.
Per discussion with Christos.

Add ucontext glue for x86_64.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

libc/x86_64: Omit needless __PIC__ conditionals.

PR lib/59391: unnecessary __PIC__ conditionals clutter .S files

branches: 1.4.2;
amd64 longjmp: Restore stack first, then signal mask.

Otherwise, a pending signal may be delivered on the wrong stack when
we restore the signal mask.

While here:

- Tidy the code a little bit.
- Sprinkle comments to explain what's going on.
- Use `xorl %eXX,%eXX' instead of `xorq %rXX,%rXX'.
=> Same effect, one byte shorter, breaks dep chain on more uarches.
- Use forward branches for statically predicted not-taken.
=> val==0 is unlikely in longjmp

PR lib/57946

branches: 1.3.26; 1.3.34;
Put missing END() markers to set ELF symbol size.

branches: 1.2.2;
Pass PICFLAGS down to cc-as-as and use __PIC__ to decide if it is small
vs big PIC mode. Retire -DPIC and -DBIGPIC.

branches: 1.1.58; 1.1.64;
Initial commit of x86-64 specific parts of libc.

libc/x86_64: Omit needless __PIC__ conditionals.

PR lib/59391: unnecessary __PIC__ conditionals clutter .S files

branches: 1.4.2;
amd64 longjmp: Restore stack first, then signal mask.

Otherwise, a pending signal may be delivered on the wrong stack when
we restore the signal mask.

While here:

- Tidy the code a little bit.
- Sprinkle comments to explain what's going on.
- Use `xorl %eXX,%eXX' instead of `xorq %rXX,%rXX'.
=> Same effect, one byte shorter, breaks dep chain on more uarches.
- Use forward branches for statically predicted not-taken.
=> val==0 is unlikely in longjmp

PR lib/57946

branches: 1.3.26; 1.3.34;
Put missing END() markers to set ELF symbol size.

branches: 1.2.2;
Pass PICFLAGS down to cc-as-as and use __PIC__ to decide if it is small
vs big PIC mode. Retire -DPIC and -DBIGPIC.

branches: 1.1.58; 1.1.64;
Initial commit of x86-64 specific parts of libc.

Trailing whitespace

Allow storing and receiving the LWP private pointer via ucontext_t
on all platforms except VAX and IA64. Add fast access via register for
AMD64, i386 and SH3 ports. Use this fast access in libpthread to replace
the stack based pthread_self(). Implement skeleton support for Alpha,
HPPA, PowerPC, SPARC and SPARC64, but leave it disabled.

Ports that support this feature provide __HAVE____LWP_GETPRIVATE_FAST in
machine/types.h and a corresponding __lwp_getprivate_fast in
machine/mcontext.h.

This material is based upon work partially supported by
The NetBSD Foundation under a contract with Joerg Sonnenberger.

branches: 1.6.2;
Fix stack alignment for the thread stack. The AMD64 stack has 16 Bytes
alignment, but this is active before the return address is pushed onto
the stack. Fixes random crashes in va_start when the SSE registers are
saved.

Reduce diff to makecontext.

branches: 1.4.8;
Remove clause 3 and 4 from TNF licenses

branches: 1.3.18;
Add missing __RCSID()

Make sure we reference getcontext() by its internal name.

Add ucontext glue for x86_64.

Put missing END() markers to set ELF symbol size.

branches: 1.1.64; 1.1.76;
Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.1.64; 1.1.76;
Initial commit of x86-64 specific parts of libc.

Move AMD64's bswap64 implementation from libc to src/common and share it
with the kernel.

Initial commit of x86-64 specific parts of libc.

Use reach-over sources from common/lib/libc

Initial commit of x86-64 specific parts of libc.

Use reach-over sources from common/lib/libc

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.3.62; 1.3.74;
simplify: it doesn't need an i387 to kill a sign bit

fix copyback from fp stack - we need 64 bits

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.6.8; 1.6.18;
PR/44293: Paul Goyette: Fix the mess of the rounding code.

even simpler.

add a simpler version.

Instead of using a PICGOT relocation for map, use just a pc-relative one
since it is guaranteed to be close enough. Fixes issue with binutils-2.21.1a
(probably a bug).

mov -> lea to access _map table.

Initial commit of x86-64 specific parts of libc.

branches: 1.4.8;
Remove clause 3 and 4 from TNF licenses

branches: 1.3.10;
ext_int bit is no more.

branches: 1.2.16;
Make this compile.

C99 7.12.3.1: add IEEE library portions of fpclassify. Don't build these
yet as the VAX implementation is still subject to discussion.

Put missing END() markers to set ELF symbol size.

branches: 1.3.62; 1.3.74;
Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.1.10; 1.1.20;
add fpgetprec/fpsetprec

Put missing END() markers to set ELF symbol size.

branches: 1.4.8; 1.4.18;
PR/44293: Paul Goyette: Fix the mess of the rounding code.

Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.3.62; 1.3.74;
Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.5.6; 1.5.16;
let one bit more through to SSE, to make FP_X_IMP work

branches: 1.4.2;
Fix fpsetmask(3) to be able to clear bits previously set.

Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.1.10; 1.1.20;
add fpgetprec/fpsetprec

Put missing END() markers to set ELF symbol size.

branches: 1.4.8; 1.4.18;
PR/44293: Paul Goyette: Fix the mess of the rounding code.

Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Put missing END() markers to set ELF symbol size.

branches: 1.5.62; 1.5.74;
we also have to reset bits in the i387 status
as required

-clear status bits in mxcsr as required
-return the old accumulated status bits

Attempt to juggle the 2 seperarate status/mask bit sets for plain FP
and XMM to provide a consistent interface.

Add internal names for fp{get,set}{mask,round,sticky}(), and
externally-visible weak aliases for the internal names.

Initial commit of x86-64 specific parts of libc.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

make sure the infinity returned by HUGE_VALL has the "implicit"
bit set, otherwise it is invalid
This code is unlikely to be hit because gcc (and clang) use
their builtins for these special values.

branches: 1.3.2;
make infinityl match what gdtoa thinks.

Add missing __RCSID()

Add __infinityf and __infinityl, float respectively long double analogs
of __infinity.

branches: 1.3.8;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.10;
ext_int bit is no more.

branches: 1.1.16;
C99 7.12.3.2: add library portions of isfinite, but don't build these just
yet.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

ext_int bit is no more.

branches: 1.5.16;
* Turn isinf(3) and isnan(3) into C99-style macros.
* Make it possible for ports to override these (i.e., VAX).
* Remove isnanl(), which was internal to libc only.

Err, rename some members added in previous to make them reflect their
semantics better.

Retire local union { floating-type; struct ieee_prec; } in favor of those
available from <machine/ieee.h>

Sync with i386:
Update to account for separate handling of the explicit integer bit;
simplifies a little.

* Since there is no single portable IEEE 754 format for a long double,
keep a common implementation of isinfl() and isnanl() to be used by
platforms where `long double' == `double'; move others into
machine-dependent code.
* In due course, consider __VFP_FP__ on arm.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

Revert previous and make frach be 0xc0000000 for NAN's generated by strtold()
as suggested by tron.

Make NAN tests consistent with other platforms and FreeBSD.

branches: 1.6.34;
ext_int bit is no more.

branches: 1.5.16;
* Turn isinf(3) and isnan(3) into C99-style macros.
* Make it possible for ports to override these (i.e., VAX).
* Remove isnanl(), which was internal to libc only.

Err, rename some members added in previous to make them reflect their
semantics better.

Retire local union { floating-type; struct ieee_prec; } in favor of those
available from <machine/ieee.h>

Sync with i386:
Update to account for separate handling of the explicit integer bit;
simplifies a little.

* Since there is no single portable IEEE 754 format for a long double,
keep a common implementation of isinfl() and isnanl() to be used by
platforms where `long double' == `double'; move others into
machine-dependent code.
* In due course, consider __VFP_FP__ on arm.

replaced by code in libc/compat/gen

Fix -Wshadow warnings.

Initial commit of x86-64 specific parts of libc.

If SRCS contains any .S files then remove the corresponding .c file from
SRCS and add to LSRCS (for lint).
Change the 'string' Makefiles so that the arch/*/string/Makefile.inc
need only specify the .S files for that architecture and not the .c
files for all the files they don't override.

branches: 1.3.8;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.30;
Get the stack alignment right: the stack arguments must start
at a 16-byte aligned address, the return address is at an
"odd" address.
approved by fvdl

Add ucontext glue for x86_64.

build {frexp,ldexp,modf} in the "compat" subtree
(Drop the modf assembler version. We have the same code in libm.)
(Drop the ldexp inline-assembler version. The same code is in libm as
scalbn; the ldexp there is just a wrapper providing error handling.)

Return the right value for modf(). Fixes PR 22503.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22280, verified by myself.

Initial commit of x86-64 specific parts of libc.

Make NAN a compile time constant (with help from the compiler). Suggested
by krister. Fixes PR 40695.
Make references to the old construct (which we can't remove for binary
compatibility reasons) emit a linker warning.

branches: 1.3.2; 1.3.32;
Push back the descriptions of NaN formats, and descriptions of the
distinction between signalling NaNs and quiet NaNs back into the
machine-dependent headers; treat the implementation of __nanf in the
same spirit.

IEEE 754 leaves the distinction between signalling NaNs and quiet NANs
to the implementation, and unlike our headers used to suggest they're
not identical in the interpretation of the fraction's MSb; in due
course, make those of hppa, mips, sh3, and sh5 reflect reality.

Clean up some rampant code duplication wrt ieee number handling:
- Add alignment-safe double and float unions.
- Use the above for the __infinity and __nan constants on all
architectures that use the standard ieee754 representation of
those constants.
- Add a single copy of various ieee754 math functions (frexp, isinf,
isnan, ldexp and modf) that had numerous duplicates among the
arch-specific directories.
- Use the above functions on all architectures where the generic C
versions where used. Architectures that had local assembly
routines are untouched (for those functions only).

Initial commit of x86-64 specific parts of libc.

libc/x86_64: Omit needless __PIC__ conditionals.

PR lib/59391: unnecessary __PIC__ conditionals clutter .S files

branches: 1.6.36;
Put missing END() markers to set ELF symbol size.

branches: 1.5.2;
Pass PICFLAGS down to cc-as-as and use __PIC__ to decide if it is small
vs big PIC mode. Retire -DPIC and -DBIGPIC.

branches: 1.4.4; 1.4.8; 1.4.28;
Remove clause 3 and 4 from TNF licenses

branches: 1.3.16;
-check for a NULL uc_link before calling setcontext(2); do a real exit(3)
instead with exit handlers and all that
-If setcontext(2) returns (which it shouldn't), do an _exit(-1) rather
than relying on a sensible return value to use as exit code. Makes
it less likely that an ucontext corruption goes unnoticed.

alignment trouble: uc_link is at offset 8

Add ucontext glue for x86_64.

Unify signbitl implementations.
(Unclear why they were different, it was only the comments that
differed)

Add a comment describing why the generic version works for both 80-bit
and 128-bit double.

branches: 1.2.8; 1.2.64;
Remove clause 3 and 4 from TNF licenses

branches: 1.1.30;
Add C99 fpclassify(), isfinite(), isnormal(), and signbit() macros.

libc/x86_64: Omit needless __PIC__ conditionals.

PR lib/59391: unnecessary __PIC__ conditionals clutter .S files

branches: 1.6.36;
Put missing END() markers to set ELF symbol size.

branches: 1.5.2;
Pass PICFLAGS down to cc-as-as and use __PIC__ to decide if it is small
vs big PIC mode. Retire -DPIC and -DBIGPIC.

branches: 1.4.36; 1.4.42;
Don't clobber callee save registers r12 & r13 in swapcontext. They
were saved in the stack, but the subsequent swapcontext would return
to the earlier stack frame and never restore them. Fixes puffs
with default compiler flags on amd64.

Thanks to Mihai Chelaru for providing access to an amd64 machine
for debugging this problem.

branches: 1.3.2; 1.3.10;
Use correct stack offsets after pushing values onto it.

branches: 1.2.2; 1.2.4;
Make sure we reference getcontext() by its internal name.

Add ucontext glue for x86_64.