While we don't use the NC_BROADCAST value of nc_flag anywhere in the
RPC code, it is parseable by getnetconfigent(3) from /etc/netconfig.
o Clean up some "see below"'s that were cut and pasted from netconfig.h.
operation, the caller is blocked util target threads are really
suspended, also avoid suspending a thread when it is holding a
critical lock.
Fix a bug in _thr_ref_delete which tests a never set flag.
commit broke the 2**24 cases where |x| > DBL_MAX/2. There are exponent
range problems not just for denormals (underflow) but for large values
(overflow). Doubles have more than enough exponent range to avoid the
problems, but I forgot to convert enough terms to double, so there was
an x+x term which was sometimes evaluated in float precision.
Unfortunately, this is a pessimization with some combinations of systems
and compilers (it makes no difference on Athlon XP's, but on Athlon64's
it gives a 5% pessimization with gcc-3.4 but not with gcc-3.3).
Exlain the problem better in comments.
algorithm for the second step significantly to also get a perfectly
rounded result in round-to-nearest mode. The resulting optimization
is about 25% on Athlon64's and 30% on Athlon XP's (about 25 cycles
out of 100 on the former).
Using extra precision, we don't need to do anything special to avoid
large rounding errors in the third step (Newton's method), so we can
regroup terms to avoid a division, increase clarity, and increase
opportunities for parallelism. Rearrangement for parallelism loses
the increase in clarity. We end up with the same number of operations
but with a division reduced to a multiplication.
Using specifically double precision, there is enough extra precision
for the third step to give enough precision for perfect rounding to
float precision provided the previous steps are accurate to 16 bits.
(They were accurate to 12 bits, which was almost minimal for imperfect
rounding in the old version but would be more than enough for imperfect
rounding in this version (9 bits would be enough now).) I couldn't
find any significant time optimizations from optimizing the previous
steps, so I decided to optimize for accuracy instead. The second step
needed a division although a previous commit optimized it to use a
polynomial approximation for its main detail, and this division dominated
the time for the second step. Use the same Newton's method for the
second step as for the third step since this is insignificantly slower
than the division plus the polynomial (now that Newton's method only
needs 1 division), significantly more accurate, and simpler. Single
precision would be precise enough for the second step, but doesn't
have enough exponent range to handle denormals without the special
grouping of terms (as in previous versions) that requires another
division, so we use double precision for both the second and third
steps.
functions in the child after a fork() from a threaded process,
use __sys_setprocmask() rather than setprocmask() to keep our
signal handling sane. Without this fix, signals are essentially
ignored in said child and things such as protection violations
result in an endless busy loop.
Reviewed by: deischen
similar the the Solaris implementation. Repackage the krb5 GSS mechanism
as a plugin library for the new implementation. This also includes a
comprehensive set of manpages for the GSS-API functions with text mostly
taken from the RFC.
Reviewed by: Love Hörnquist Åstrand <lha@it.su.se>, ru (build system), des (openssh parts)
between a 32-bit integer and a radix-64 ASCII string. The l64a_r() function
is a NetBSD addition.
PR: 51209 (based on submission, but very different)
Reviewed by: bde, ru
distributed non-large args, this saves about 14 of 134 cycles for
Athlon64s and about 5 of 199 cycles for AthlonXPs.
Moved the check for x == 0 inside the check for subnormals. With
gcc-3.4 on uniformly distributed non-large args, this saves another
5 cycles on Athlon64s and loses 1 cycle on AthlonXPs.
Use INSERT_WORDS() and not SET_HIGH_WORD() when converting the first
approximation from bits to double. With gcc-3.4 on uniformly distributed
non-large args, this saves another 4 cycles on both Athlon64s and and
AthlonXPs.
Accessing doubles as 2 words may be an optimization on old CPUs, but on
current CPUs it tends to cause extra operations and pipeline stalls,
especially for writes, even when only 1 of the words needs to be accessed.
Removed an unused variable.
function approximation for the second step. The polynomial has degree
2 for cbrtf() and 4 for cbrt(). These degrees are minimal for the final
accuracy to be essentially the same as before (slightly smaller).
Adjust the rounding between steps 2 and 3 to match. Unfortunately,
for cbrt(), this breaks the claimed accuracy slightly although incorrect
rounding doesn't. Claim less accuracy since its not worth pessimizing
the polynomial or relying on exhaustive testing to get insignificantly
more accuracy.
This saves about 30 cycles on Athlons (mainly by avoiding 2 divisions)
so it gives an overall optimization in the 10-25% range (a larger
percentage for float precision, especially in 32-bit mode, since other
overheads are more dominant for double precision, surprisingly more
in 32-bit mode).
- in preparing for the third approximation, actually make t larger in
magnitude than cbrt(x). After chopping, t must be incremented by 2
ulps to make it larger, not 1 ulp since chopping can reduce it by
almost 1 ulp and it might already be up to half a different-sized-ulp
smaller than cbrt(x). I have not found any cases where this is
essential, but the think-time error bound depends on it. The relative
smallness of the different-sized-ulp limited the bug. If there are
cases where this is essential, then the final error bound would be
5/6+epsilon instead of of 4/6+epsilon ulps (still < 1).
- in preparing for the third approximation, round more carefully (but
still sloppily to avoid branches) so that the claimed error bound of
0.667 ulps is satisfied in all cases tested for cbrt() and remains
satisfied in all cases for cbrtf(). There isn't enough spare precision
for very sloppy rounding to work:
- in cbrt(), even with the inadequate increment, the actual error was
0.6685 in some cases, and correcting the increment increased this
a little. The fix uses sloppy rounding to 25 bits instead of very
sloppy rounding to 21 bits, and starts using uint64_t instead of 2
words for bit manipulation so that rounding more bits is not much
costly.
- in cbrtf(), the 0.667 bound was already satisfied even with the
inadequate increment, but change the code to almost match cbrt()
anyway. There is not enough spare precision in the Newton
approximation to double the inadequate increment without exceeding
the 0.667 bound, and no spare precision to avoid this problem as
in cbrt(). The fix is to round using an increment of 2 smaller-ulps
before chopping so that an increment of 1 ulp is enough. In cbrt(),
we essentially do the same, but move the chop point so that the
increment of 1 is not needed.
Fixed comments to match code:
- in cbrt(), the second approximation is good to 25 bits, not quite 26 bits.
- in cbrt(), don't claim that the second approximation may be implemented
in single precision. Single precision cannot handle the full exponent
range without minor but pessimal changes to renormalize, and although
single precision is enough, 25 bit precision is now claimed and used.
Added comments about some of the magic for the error bound 4/6+epsilon.
I still don't understand why it is 4/6+ and not 6/6+ ulps.
Indent comments at the right of code more consistently.
to be compatible with symbol versioning support as implemented by
GNU libc and documented by http://people.redhat.com/~drepper/symbol-versioning
and LSB 3.0.
Implement dlvsym() function to allow lookups for a specific version of
a given symbol.
means:
o Remove Elf64_Quarter,
o Redefine Elf64_Half to be 16-bit,
o Redefine Elf64_Word to be 32-bit,
o Add Elf64_Xword and Elf64_Sxword for 64-bit entities,
o Use Elf_Size in MI code to abstract the difference between
Elf32_Word and Elf64_Word.
o Add Elf_Ssize as the signed counterpart of Elf_Size.
MFC after: 2 weeks
on probationary terms: it may go away again if it transpires it is
a bad idea.
This extensible printf version will only be used if either
environment variable USE_XPRINTF is defined
or
one of the extension functions are called.
or
the global variable __use_xprintf is set greater than zero.
In all other cases our traditional printf implementation will
be used.
The extensible version is slower than the default printf, mostly
because less opportunity for combining I/O operation exists when
faced with extensions. The default printf on the other hand
is a bad case of spaghetti code.
The extension API has a GLIBC compatible part and a FreeBSD version
of same. The FreeBSD version exists because the GLIBC version may
run afoul of our FILE * locking in multithreaded programs and it
even further eliminate the opportunities for combining I/O operations.
Include three demo extensions which can be enabled if desired: time
(%T), hexdump (%H) and strvis (%V).
%T can format time_t (%T), struct timeval (%lT) and struct timespec (%llT)
in one of two human readable duration formats:
"%.3llT" -> "20349.245"
"%#.3llT" -> "5h39m9.245"
%H will hexdump a sequence of bytes and takes a pointer and a length
argument. The width specifies number of bytes per line.
"%4H" -> "65 72 20 65"
"%+4H" -> "0000 65 72 20 65"
"%#4H" -> "65 72 20 65 |er e|"
"%+#4H" -> "0000 65 72 20 65 |er e|"
%V will dump a string in strvis format.
"%V" -> "Hello\tWor\377ld" (C-style)
"%0V" -> "Hello\011Wor\377ld" (octal)
"%+V" -> "Hello%09Wor%FFld" (http-style)
Tests, comments, bugreports etc are most welcome.
allocate a memory block. sscanf calls __svfscanf which in turn calls
fread, fread triggers mutex initialization but the mutex is not
destroyed in sscanf, this leads to memory leak. To avoid the memory
leak and performance issue, we create a none MT-safe version of fread:
__fread, and instead let __svfscanf call __fread.
PR: threads/90392
Patch submitted by: dhartmei
MFC after: 7 days
the second step of approximating cbrt(x). It turns out to be neither
very magic not nor very good. It is just the (2,2) Pade approximation
to 1/cbrt(r) at r = 1, arranged in a strange way to use fewer operations
at a cost of replacing 4 multiplications by 1 division, which is an
especially bad tradeoff on machines where some of the multiplications
can be done in parallel. A Remez rational approximation would give
at least 2 more bits of accuracy, but the (2,2) Pade approximation
already gives 6 more bits than needed. (Changed the comment which
essentially says that it gives 3 more bits.)
Lower order Pade approximations are not quite accurate enough for
double precision but are plenty for float precision. A lower order
Remez rational approximation might be enough for double precision too.
However, rational approximations inherently require an extra division,
and polynomial approximations work well for 1/cbrt(r) at r = 1, so I
plan to switch to using the latter. There are some technical
complications that tend to cost a division in another way.
This gives an optimization of between 9 and 22% on Athlons (largest
for cbrt() on amd64 -- from 205 to 159 cycles).
We extracted the sign bit and worked with |x|, and restored the sign
bit as the last step. We avoided branches to a fault by using accesses
to FP values as bits to clear and restore the sign bit. Avoiding
branches is usually good, but the bit access macros are not so good
(especially for setting FP values), and here they always caused pipeline
stalls on Athlons. Even using branches would be faster except on args
that give perfect branch misprediction, since only mispredicted branches
cause stalls, but it possible to avoid touching the sign bit in FP
values at all (except to preserve it in conversions from bits to FP
not related to the sign bit). Do this. The results are identical
except in 2 of the 3 unsupported rounding modes, since all the
approximations use odd rational functions so they work right on strictly
negative values, and the special case of -0 doesn't use an approximation.
For some denormalized long double values, a bug in __hldtoa() (called
from *printf()'s %A format) results in a base 16 digit being rounded
up from 0xf to 0x10.
When this digit is subsequently converted to string format, an index
of 10 reaches past the end of the uppper-case hex/char array, picking
up whatever the code segment happen to contain at that address.
This mostly seem to be some character from the upper half of the
byte range.
When using the %a format instead of %A, the first character past
the end of the lowercase hex/char table happens to be index 0 in
the uppercase hex/char table hextable and therefore the string
representation features a '0', which is supposedly correct.
This leads me to belive that the proper fix _may_ be as simple as
masking all but the lower four bits off after incrementing a hex-digit
in libc/gdtoa/_hdtoa.c:roundup(). I worry however that the upper
bit in 0x10 indicates a carry not carried.
Until das@ or bde@ finds time to visit this issue, extend the
hexdigit arrays with a 17th index containing '?' so that we get a
invalid but consistent and printable output in both %a and %A formats
whenever this bug strikes.
This unmasks the bug in the %a format therefore solving the real
issue may both become easier and more urgent.
Possibly related to: PR 85080
With help by: bde@
<cbrt(x) in bits> ~= <x in bits>/3 + BIAS.
Keep the large comments only in the double version as usual.
Fixed some style bugs (mainly grammar and spelling errors in comments).
