remove these, they are no longer necessary or appropriate

branches: 1.2.2;
Add missing stdlib.h include

File which provides lint stubs for the functions defined by
machine-dependent assembly code, whose source files live in its directory.
(Not necessarily complete yet.)

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" part
(code is identical)

some stuff went to compat

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: 1.18.6;
* 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.

we need a silly trampoline to make _resumecontext work

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

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 swapcontext() lint stub.

Merge the nathanw_sa branch.

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).

branches: 1.7.6;
C99: Define a NAN macro in <math.h> which evaulates to a constant expression of
a single-precision quiet NaN; only to be defined on platforms that do support
this value.

Separate isnan() into a separate module, and make isinf()/isnan() use ieee.h.

Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.

Signal changes for Alpha.

add entries for the float rounding and exception functions

don't forget inifinity.c

need these for new gen/Makefile

POSIX sez:
The longjmp() function shall not cause setjmp() to return 0; if val
is 0, setjmp() shall return 1.

Fixes the _longjmp_zero and longjmp_zero test cases in the t_setjmp
test on alpha.

branches: 1.7.36;
Remove clearing of ucontext_t, the only random register values now are
argument and scratch registers (which will be reloaded after return anyway).

Save the GP register value and restore it on longjmp.

While formally GP is not a callee-saved register, for static linking
the link time optimization described in section 3.2.3 of the "Calling
Standard for Alpha Systems" allows the caller to rely on GP being the
value needed within setjmp() - and not restore it after return if setjmp
and the call site share a GP value.

This fixes longjmp/setjmp for statically linked programs, e.g. /rescue/csh
or the static tcsh variant from pkgsrc.

branches: 1.5.4; 1.5.8; 1.5.24; 1.5.28; 1.5.30;
Remove clause 3 and 4 from TNF licenses

branches: 1.4.18;
Add missing <compat/include/setjmp.h>

Put the jmpbuf's FPCR in the ucontext, not the other way around.

Fix <tab> nits.

Change longjmp to use setcontext instead of sigreturn14. Change setjmp
to use symbolic offsets via assym.h

Ensure that the stack is always 16-byte aligned by rounding sizes as needed.

All changes from rin@ except swapcontext.S, which is from me (added
symbolic constants to make the code clearer).

- Consistently use _REG_* defines from assym.h to mean register numbers,
not offsets (i.e. *8 to get the offset).
- Define and use SIZEOF_SIGINFO and UC_GREGS constants, rather than
hard-code magic numbers.

NFC -- same object code is generated.

Use a constant in the text segment to avoid the .lit8 out of range relocation
error.

Save the GP register value and restore it on longjmp.

While formally GP is not a callee-saved register, for static linking
the link time optimization described in section 3.2.3 of the "Calling
Standard for Alpha Systems" allows the caller to rely on GP being the
value needed within setjmp() - and not restore it after return if setjmp
and the call site share a GP value.

This fixes longjmp/setjmp for statically linked programs, e.g. /rescue/csh
or the static tcsh variant from pkgsrc.

branches: 1.5.2; 1.5.4; 1.5.8; 1.5.10;
Use excb before and after the mt_fpcr/mf_fpcr instructions as specified
in the alpha ARM 4.7.8.1 (Accessing the FPCR):

Because Alpha floating-point hardware can overlap the execution of
a number of floating-point instructions, accessing the FPCR must
be synchronized with other floating-point instructions. An EXCB
instruction must be issued both prior to and after accessing the
FPCR to ensure that the FPCR access is synchronized with the
execution of previous and subsequent floating-point instructions;
otherwise synchronization is not ensured.

Use unions to convert between double and uint64_t to avoid casting.

branches: 1.4.54;
Change longjmp to use setcontext instead of sigreturn14. Change setjmp
to use symbolic offsets via assym.h

Siginfo libc changes for alpha. Approved by thorpej@netbsd.org.

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.1.8;
Signal changes for Alpha.

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.1.8;
Signal changes for Alpha.

lib/libc/arch/alpha: Say STACK_ALIGNBYTES, not 0xf, where relevant.

No functional change intended.

PR kern/59327: user stack pointer is not aligned properly

lib/libc/arch/alpha/gen/_lwp.c, makecontext.c: Sort includes.

No functional change intended.

Preparation for tidying for:

PR kern/59327: user stack pointer is not aligned properly

branches: 1.8.4;
Ensure that the stack is always 16-byte aligned by rounding sizes as needed.

All changes from rin@ except swapcontext.S, which is from me (added
symbolic constants to make the code clearer).

Fix the build, _UC_UNIQUE has been renamed _UC_TLSBASE

branches: 1.6.4; 1.6.10;
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: 1.5.8; 1.5.20;
Remove clause 3 and 4 from TNF licenses

branches: 1.4.18;
Add missing __RCSID()

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

Merge the nathanw_sa branch.

branches: 1.1.2;
file _lwp.c was initially added on branch nathanw_sa.

branches: 1.3.8;
add matching END(). found via -Wa,--fatal-warnings

branches: 1.2.30;
blow this up by sticking a copyright notice to it

we need a silly trampoline to make _resumecontext work

these went to compat

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.12;
catch up with <machine/asm.h> changes, SYS.h changes, and the removal
of DEFS.h

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

Use reach-over sources from common/lib/libc

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.8;
Use insbl to reduce by 1 instruction.

Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.

Use reach-over sources from common/lib/libc

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.3.8;
Whitespace fix.

Use insbl to reduce by 1 instruction.

Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.

Don't intentionally enable integer overflow traps (ITOH Yasufumi)
Closes port-alpha/10065

branches: 1.8.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.7.12;
catch up with <machine/asm.h> changes, SYS.h changes, and the removal
of DEFS.h

remove unused local label

deal properly (i hope!) with registers' upper 32 bits, when doing
32-bit division and remainder. Sometimes, when optimizing, they could
have been different than previously expected (and could have caused
hokey results).

kill some now-bogus XXX's

save & restore the argument registers, too. from Trevor Blackwell

strange corner cases:
(1) zero top 32 bits for 32-bit unsigned ops; 32-bit value can
be accidentally(?) sign-extended by the compiler/assembler.
(2) fix return value for mod's that cause traps

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

On ports without __HAVE_LONG_DOUBLE, make fabsl alias fabs.

For some reason, fabs lives in libc, not in libm, and our tests now
detect when fabs or fabsl is missing from libm. For those ports that
sometimes have long double and sometimes don't, make it conditional.

Still missing: fabs _and_ fabsl on ia64. Need help from an itanium
wizard! Other portmasters: Please take a look and see if I missed
any ports that might have long double where this alias will not work.

branches: 1.3.102; 1.3.104;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.12;
catch up with <machine/asm.h> changes, SYS.h changes, and the removal
of DEFS.h

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

defintion -> definition

branches: 1.9.8; 1.9.18;
Use excb before and after the mt_fpcr/mf_fpcr instructions as specified
in the alpha ARM 4.7.8.1 (Accessing the FPCR):

Because Alpha floating-point hardware can overlap the execution of
a number of floating-point instructions, accessing the FPCR must
be synchronized with other floating-point instructions. An EXCB
instruction must be issued both prior to and after accessing the
FPCR to ensure that the FPCR access is synchronized with the
execution of previous and subsequent floating-point instructions;
otherwise synchronization is not ensured.

Use unions to convert between double and uint64_t to avoid casting.

branches: 1.8.42;
__asm__ -> __asm

Add missing __RCSID()

add a cast appease gcc3.3

netbsd.org->NetBSD.org

sweep of my licenses (userland files w/o only my copyright) for
consistency. (no functional changes)

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.10;
Pull in __flt_rounds() prototype.

get & set FP rounding direction. these actually seem to work.

Add missing __RCSID()

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

minor cleanup

branches: 1.2.10;
* New MI man page for fp{set,get}{round,sticky,mask}(3).
* Implement fpgetsticky() for alpha.
* Direct fpsetsticky() and fp{get,set}mask() into alpha kernel via sysarch(2).
* Define new sysarch(2) stub for above and install and distribute sysarch.h
for alpha. (The fpcr IS user mode r/w, but for reasons beyond the scope
of a commit message kernel calls are needed.) And much kernel Magick is
required before these do anything, but this way programs compiled under
1.4 will DTRT on future snapshots and releases.

branches: 1.1.12;
functions to set FP exception masks & sticky exceptions. they abort, for now.

Add some casts to get rid of "bitwise op on signed value is non-portable"
warning from lint.

branches: 1.10.2;
Use excb before and after the mt_fpcr/mf_fpcr instructions as specified
in the alpha ARM 4.7.8.1 (Accessing the FPCR):

Because Alpha floating-point hardware can overlap the execution of
a number of floating-point instructions, accessing the FPCR must
be synchronized with other floating-point instructions. An EXCB
instruction must be issued both prior to and after accessing the
FPCR to ensure that the FPCR access is synchronized with the
execution of previous and subsequent floating-point instructions;
otherwise synchronization is not ensured.

Use unions to convert between double and uint64_t to avoid casting.

branches: 1.9.42;
__asm__ -> __asm

Add missing __RCSID()

add a cast appease gcc3.3

netbsd.org->NetBSD.org

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

remove the trapb ops bracketing mf_fpcr, in this one case they aren't needed

branches: 1.3.4;
sweep of my licenses (userland files w/o only my copyright) for
consistency. (no functional changes)

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.1.18;
get & set FP rounding direction. these actually seem to work.

branches: 1.9.6;
Backout previous: the license sweep touched these files in error, so
restore the old license.

branches: 1.8.2;
Remove clause 3 and 4 from TNF licenses

Add missing __RCSID()

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

Complete rewrite. Call the new sysarch(ALPHA_FPGETSTICKY, ...).

branches: 1.4.4;
sweep of my licenses (userland files w/o only my copyright) for
consistency. (no functional changes)

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.6;
* New MI man page for fp{set,get}{round,sticky,mask}(3).
* Implement fpgetsticky() for alpha.
* Direct fpsetsticky() and fp{get,set}mask() into alpha kernel via sysarch(2).
* Define new sysarch(2) stub for above and install and distribute sysarch.h
for alpha. (The fpcr IS user mode r/w, but for reasons beyond the scope
of a commit message kernel calls are needed.) And much kernel Magick is
required before these do anything, but this way programs compiled under
1.4 will DTRT on future snapshots and releases.

branches: 1.1.12;
functions to set FP exception masks & sticky exceptions. they abort, for now.

fix old style decls

branches: 1.4.44; 1.4.50;
Add missing __RCSID()

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

branches: 1.2.10;
* New MI man page for fp{set,get}{round,sticky,mask}(3).
* Implement fpgetsticky() for alpha.
* Direct fpsetsticky() and fp{get,set}mask() into alpha kernel via sysarch(2).
* Define new sysarch(2) stub for above and install and distribute sysarch.h
for alpha. (The fpcr IS user mode r/w, but for reasons beyond the scope
of a commit message kernel calls are needed.) And much kernel Magick is
required before these do anything, but this way programs compiled under
1.4 will DTRT on future snapshots and releases.

branches: 1.1.12;
functions to set FP exception masks & sticky exceptions. they abort, for now.

Add some casts to get rid of "bitwise op on signed value is non-portable"
warning from lint.

branches: 1.11.2;
Use excb before and after the mt_fpcr/mf_fpcr instructions as specified
in the alpha ARM 4.7.8.1 (Accessing the FPCR):

Because Alpha floating-point hardware can overlap the execution of
a number of floating-point instructions, accessing the FPCR must
be synchronized with other floating-point instructions. An EXCB
instruction must be issued both prior to and after accessing the
FPCR to ensure that the FPCR access is synchronized with the
execution of previous and subsequent floating-point instructions;
otherwise synchronization is not ensured.

Use unions to convert between double and uint64_t to avoid casting.

branches: 1.10.42;
__asm__ -> __asm

Add missing __RCSID()

add a cast appease gcc3.3

netbsd.org->NetBSD.org

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

tweak barrier ops some more

Delete the trapb ops bracketing mf_fpcr, they aren't needed.
Change the trapb ops bracketing mt_fpcr to excb ops.

branches: 1.3.4;
sweep of my licenses (userland files w/o only my copyright) for
consistency. (no functional changes)

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.1.18;
get & set FP rounding direction. these actually seem to work.

fix old style decls

branches: 1.4.44; 1.4.50;
Add missing __RCSID()

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

branches: 1.2.10;
* New MI man page for fp{set,get}{round,sticky,mask}(3).
* Implement fpgetsticky() for alpha.
* Direct fpsetsticky() and fp{get,set}mask() into alpha kernel via sysarch(2).
* Define new sysarch(2) stub for above and install and distribute sysarch.h
for alpha. (The fpcr IS user mode r/w, but for reasons beyond the scope
of a commit message kernel calls are needed.) And much kernel Magick is
required before these do anything, but this way programs compiled under
1.4 will DTRT on future snapshots and releases.

branches: 1.1.12;
functions to set FP exception masks & sticky exceptions. they abort, for now.

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).

branches: 1.4.10;
Trivial KNF.

Make the C version of frexp() consitent across platforms, and handle the
Inf/NaN case.

Just some minor cleanup.

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

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).

branches: 1.6.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.5.2;
Fix unaligned accesses more correctly.

Fix unaligned accesses.

delint

const poisoning.

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

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).

branches: 1.5.6;
Delint.

Separate isnan() into a separate module, and make isinf()/isnan() use ieee.h.

delint

Need internal names for isinf() and isnan().

branches: 1.1.2;
libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

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).

branches: 1.2.6;
Delint.

Separate isnan() into a separate module, and make isinf()/isnan() use ieee.h.

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).

Fix -Wshadow warnings.

branches: 1.9.10;
Another microoptimization. Take advantage of the addition already having set
the flags for the <=0 comparison.

Reuse the trick of loading 0.0 (rather than 1.0) and then modifying the
exponent. It's a bit faster on some machines.

Remove more unused mulexp variables.

ldexp(denormal, exp>1023) would generate the wrong result in all non-overflow
cases. Totally rework this code to fix this bug *and* make it faster.

Don't ever return -0.0.

Minor style change.

Sync with SPARC version.

Just some minor cleanup.

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

lib/libc/arch/alpha: Say STACK_ALIGNBYTES, not 0xf, where relevant.

No functional change intended.

PR kern/59327: user stack pointer is not aligned properly

lib/libc/arch/alpha/gen/_lwp.c, makecontext.c: Sort includes.

No functional change intended.

Preparation for tidying for:

PR kern/59327: user stack pointer is not aligned properly

branches: 1.7.4;
Ensure that the stack is always 16-byte aligned by rounding sizes as needed.

All changes from rin@ except swapcontext.S, which is from me (added
symbolic constants to make the code clearer).

Explicitly add back __resumecontext wrapper

Mark _resumecontext as dead. ANSIfy all declarations for it.

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

branches: 1.3.30;
we need a silly trampoline to make _resumecontext work

Merge the nathanw_sa branch.

branches: 1.1.2;
file makecontext.c was initially added on branch nathanw_sa.

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).

branches: 1.3.4;
fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.2;
Fix doubled 'the's.

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

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: 1.4.2; 1.4.32;
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).

Align __nanf[] object with (sigh) gnuisms

branches: 1.1.6;
C99: Define a NAN macro in <math.h> which evaulates to a constant expression of
a single-precision quiet NaN; only to be defined on platforms that do support
this value.

Mark _resumecontext as dead. ANSIfy all declarations for it.

branches: 1.4.8;
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.

Merge the nathanw_sa branch.

branches: 1.1.2;
file resumecontext.c was initially added on branch nathanw_sa.

these went to compat

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.3.8;
Make setjmp() work again after the changes to sigaltstack().

catch up with <machine/asm.h> changes, SYS.h changes, and the removal
of DEFS.h

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

these went to compat

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well. No functional changes.

branches: 1.2.12;
catch up with <machine/asm.h> changes, SYS.h changes, and the removal
of DEFS.h

libc support for the alpha. more functions should be written in assembler,
but that's optimization. 8-)

Ensure that the stack is always 16-byte aligned by rounding sizes as needed.

All changes from rin@ except swapcontext.S, which is from me (added
symbolic constants to make the code clearer).

- Consistently use _REG_* defines from assym.h to mean register numbers,
not offsets (i.e. *8 to get the offset).
- Define and use SIZEOF_SIGINFO and UC_GREGS constants, rather than
hard-code magic numbers.

NFC -- same object code is generated.

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

branches: 1.2.30;
Merge the nathanw_sa branch.

branches: 1.1.2;
file swapcontext.S was initially added on branch nathanw_sa.