https://sourceware.org/bugzilla/show_bug.cgi?id=21782 dropped an ld
diagnostic for R_X86_64_PC32 referencing an undefined weak symbol in
-pie links. Arguably keeping the diagnostic like other ports is more
correct, since statically resolving movl foo(%rip), %eax to the
link-time zero address produces a corrupted output.
It turns out that --enable-static-pie builds do not depend on the ld
behavior. GCC generates GOT indirection for weak declarations for
-fPIE/-fPIC, so what ld does with the PC-relative relocation doesn't
really matter.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Refactor the sincos implementation - rather than rely on odd partial inlining
of preprocessed portions from sin and cos, explicitly write out the cases.
This makes sincos much easier to maintain and provides an additional 16-20%
speedup between 0 and 2^27. The overall speedup of sincos is 48% over this range.
Between 0 and PI it is 66% faster.
* sysdeps/ieee754/dbl-64/s_sin.c (__sin): Cleanup ifdefs.
(__cos): Likewise.
* sysdeps/ieee754/dbl-64/s_sin.c (__sincos): Refactor using the same
logic as sin and cos.
Refactor duplicated code into do_sin. Since all calls to do_sin use copysign to
set the sign of the result, move it inside do_sin. Small inputs use a separate
polynomial, so move this into do_sin as well (the check is based on the more
conservative case when doing large range reduction, but could be relaxed).
* sysdeps/ieee754/dbl-64/s_sin.c (do_sin): Use TAYLOR_SIN for small
inputs. Return correct sign.
(do_sincos): Remove small input check before do_sin, let do_sin set
the sign.
(__sin): Likewise.
(__cos): Likewise.
For huge inputs use the improved do_sincos function as well. Now no cases use
the correction factor returned by do_sin, do_cos and TAYLOR_SIN, so remove it.
* sysdeps/ieee754/dbl-64/s_sin.c (TAYLOR_SIN): Remove cor parameter.
(do_cos): Remove corp parameter and calculations.
(do_sin): Likewise.
(do_sincos): Remove cor variable.
(__sin): Use do_sincos for huge inputs.
(__cos): Likewise.
* sysdeps/ieee754/dbl-64/s_sincos.c (__sincos): Likewise.
(reduce_and_compute_sincos): Remove unused function.
This patch improves the accuracy of the range reduction. When the input is
large (2^27) and very close to a multiple of PI/2, using 110 bits of PI is not
enough. Improve range reduction accuracy to 136 bits. As a result the special
checks for results close to zero can be removed. The ULP of the polynomials is
at worst 0.55ULP, so there is no reason for the slow functions, and they can be
removed.
* sysdeps/ieee754/dbl-64/s_sin.c (reduce_sincos_1): Rename to
reduce_sincos, improve accuracy to 136 bits.
(do_sincos_1): Rename to do_sincos, remove fallbacks to slow functions.
(__sin): Use improved reduction and simplified do_sincos calculation.
(__cos): Likewise.
* sysdeps/ieee754/dbl-64/s_sincos.c (__sincos): Likewise.
This patch removes the large range reduction code and defers to the huge range
reduction code. The first level range reducer supports inputs up to 2^27,
which is way too large given that inputs for sin/cos are typically small
(< 10), and optimizing for a smaller range would give a significant speedup.
Input values above 2^27 are practically never used, so there is no reason for
supporting range reduction between 2^27 and 2^48. Removing it significantly
simplifies code and enables further speedups. There is about a 2.3x slowdown
in this range due to __branred being extremely slow (a better algorithm could
easily more than double performance).
* sysdeps/ieee754/dbl-64/s_sin.c (reduce_sincos_2): Remove function.
(do_sincos_2): Likewise.
(__sin): Remove middle range reduction case.
(__cos): Likewise.
* sysdeps/ieee754/dbl-64/s_sincos.c (__sincos): Remove middle range
reduction case.
This series of patches removes the slow patchs from sin, cos and sincos.
Besides greatly simplifying the implementation, the new version is also much
faster for inputs up to PI (41% faster) and for large inputs needing range
reduction (27% faster).
ULP is ~0.55 with no errors found after testing 1.6 billion inputs across most
of the range with mpsin and mpcos. The number of incorrectly rounded results
(ie. ULP >0.5) is at most ~2750 per million inputs between 0.125 and 0.5,
the average is ~850 per million between 0 and PI.
Tested on AArch64 and x86_64 with no regressions.
The first patch removes the slow paths for the cases where the input is small
and doesn't require range reduction. Update ULP tables for sin, cos and sincos
on AArch64 and x86_64.
* sysdeps/aarch64/libm-test-ulps: Update ULP for sin, cos, sincos.
* sysdeps/ieee754/dbl-64/s_sin.c (__sin): Remove slow paths for small
inputs.
(__cos): Likewise.
* sysdeps/x86_64/fpu/libm-test-ulps: Update ULP for sin, cos, sincos.
Unlike GCC, llvm always uses an integrated assembler, which attempts to
recognized all `asm` statements written in the C code. glibc uses some
syntactically invalid asm statements to emit constants into assembly that
are later extracted with a sed or AWK script.
This change fixes two such invalid `asm` statements by wrapping the
output in a `.ascii` directive.. This does not break the sed/AWK (the same
special sequence is output) but it makes the statement syntactically valid.
See cf8e3f8757 for a previous fix for the same issue.
Bug 22639 reports localtime failing to handle time offset transitions
correctly in 2039 and later on platforms with 64-bit time_t.
The problem is the use of SECSPERDAY (constant 86400) in calculations
such as
t = ((year - 1970) * 365
+ /* Compute the number of leapdays between 1970 and YEAR
(exclusive). There is a leapday every 4th year ... */
+ ((year - 1) / 4 - 1970 / 4)
/* ... except every 100th year ... */
- ((year - 1) / 100 - 1970 / 100)
/* ... but still every 400th year. */
+ ((year - 1) / 400 - 1970 / 400)) * SECSPERDAY;
where t is of type time_t and year is of type int. Before my commit
92bd70fb85 (an update from tzcode,
included in 2.26 and later releases), SECSPERDAY was obtained from a
file imported from tzcode, where the value included a cast to
int_fast32_t. On 64-bit platforms, glibc defines int_fast32_t to be
long int, so 64-bit, but my patch resulted in it changing to int.
(The bug would probably have existed even before my patch for x32,
which has 64-bit time_t but 32-bit int_fast32_t, but I haven't
verified that.)
This patch fixes the problem by including a cast to time_t in the
definition of SECSPERDAY. (64-bit time support for 32-bit systems
should move such code that isn't a public interface to using the
internal 64-bit version of time_t throughout.)
Tested for x86_64 and x86.
[BZ #22639]
* time/tzset.c (SECSPERDAY): Cast to time_t.
* time/tst-y2039.c: New file.
* time/Makefile (tests): Add tst-y2039.
These comments should make it easier to see the (small) diff introduced
in cf8e3f8757. Without these comments, the diff may get list on a future
upstream merge.
The gen-as-const and gen-py-const scripts are used to generate integer constant
definitions from a list of constant C-expressions. This is achieved by
generating a C program with inline `asm` statements, that depend on
these constant expressions. During compilation, the constant expressions
are evaluated, and included in the inline asm. The build process
generates only the assembly, and then used `sed` to extract the values
from the assembly text.
This is clever. It allows the build process to extract the value of C
statements built under the target architecture. The implementation is a
bit fragile, but it is not immediately obvious to me how it could be
improved.
This change slightly modifies `gen-as-const` and `gen-py-const` to emit
valid assembly directives instead of invalid directives that were
previously emitted. Since the values are extracted via string parsing,
this has no effect on the values extracted. This is needed because the
LLVM assembler validates all statements before emitting them, whereas it
appears GCC will literally emit any `asm` directives without validation
or recognition.
This completes the deprecation and removal of this inclusion, which
was begun in the 2.25 release.
* posix/sys/types.h: Don't include sys/sysmacros.h.
* misc/sys/sysmacros.h: Remove the conditional deprecation
warnings for the macros defined by this header.
+ Use DOT_MACHINE macro instead of ".machine" instruction.
+ Use __isinf and __isinff instead of builtin versions.
+ In s_logb, s_logbf and s_logbl functions, used float versions to
calculate "ret = x & 0x7f800000;" expression.
