Commit Graph

15332 Commits

Author SHA1 Message Date
Joseph Myers
2ee094ff75 Don't disable CMPLXL macro for __NO_LONG_DOUBLE_MATH (bug 15488). 2013-05-18 12:12:38 +00:00
Joseph Myers
3779b5b64a Make libm-test look up ulps by name at runtime. 2013-05-18 12:10:59 +00:00
Joseph Myers
bb38759d6d Fix remainder exceptions and directed-rounding results (bugs 15480, 15485). 2013-05-17 19:04:08 +00:00
Joseph Myers
a00bdcf0e0 Simplify libm-test extra-output initialization. 2013-05-17 19:02:19 +00:00
Joseph Myers
de407f79a2 Don't handle ulps for integer tests in libm-test.inc. 2013-05-17 19:01:27 +00:00
Joseph Myers
8269107fe6 Test more cases of "inexact" exceptions in libm-test.inc. 2013-05-17 14:45:50 +00:00
Siddhesh Poyarekar
48a18de1e1 Prevent optimizing out of benchmark function call
Resolves: #15424

The compiler would optimize the benchmark function call out of the
loop and call it only once, resulting in blazingly fast times for some
benchmarks (notably atan, sin and cos).  Mark the inputs as volatile
so that the code is forced to read again from the input for each
iteration.
2013-05-17 19:10:33 +05:30
Adhemerval Zanella
13d3b41a36 PowerPC: fix hypot/hypotf check for -INF 2013-05-17 08:15:59 -05:00
Maciej W. Rozycki
2848b10585 MIPS: soft-fp NaN representation corrections
[BZ #15442] This adds support for the inverse interpretation of the
quiet bit of IEEE 754 floating-point NaN data that some processors
use.  This includes in particular MIPS architecture processors; the
payload used for the canonical qNaN encoding is updated accordingly
so as not to interfere with the quiet bit.
2013-05-16 23:33:55 +01:00
Joseph Myers
c58b274f01 Test for errno setting in more pole error cases. 2013-05-16 21:57:25 +00:00
Joseph Myers
0ab349044b Remove ENOSYS tests in libm-test.inc. 2013-05-16 19:09:54 +00:00
Joseph Myers
323e5cb792 Remove libm-test support for TEST_* inside functions. 2013-05-16 19:09:03 +00:00
Ryan S. Arnold
e054f49430 Add #include <stdint.h> for uint[32|64]_t usage (except installed headers). 2013-05-16 11:32:54 -05:00
Joseph Myers
8c75f67421 Convert TEST_extra tests from code to data. 2013-05-16 15:53:40 +00:00
Richard Henderson
e6e49e59c1 De-stringify constants in math/atest-exp2.c. 2013-05-16 06:34:01 -07:00
Joseph Myers
105a07dfc0 Support testing "inexact" exceptions in libm-test.inc. 2013-05-16 13:19:38 +00:00
Joseph Myers
e9eee333b9 Test for errno setting in more overflow error cases. 2013-05-16 13:18:48 +00:00
Joseph Myers
1c38ff73fd Convert TEST_cc_c tests from code to data. 2013-05-16 13:17:55 +00:00
Joseph Myers
15c7c18dc6 Convert TEST_f_L tests from code to data. 2013-05-16 13:16:56 +00:00
Peter Collingbourne
f137ff1383 Remove const attribute on get_log2
This function is not const, as it can modify log2_m and log2_m_inited.
2013-05-15 14:27:53 -07:00
Joseph Myers
3608cb241e Convert TEST_f_l tests from code to data. 2013-05-15 19:48:17 +00:00
Peter Collingbourne
1deff3dca1 Use movq for 64-bit operations
The EXTRACT_WORDS64 and INSERT_WORDS64 macros use movd for a 64-bit
operation.  Somehow gcc manages to turn this into movq, but LLVM won't.

2013-05-15  Peter Collingbourne  <pcc@google.com>

	* sysdeps/x86_64/fpu/math_private.h (MOVQ): New macro.
	(EXTRACT_WORDS64) Use where appropriate.
	(INSERT_WORDS64) Likewise.
2013-05-15 20:33:45 +02:00
Peter Collingbourne
791f3ba0db Use x constraints for operands to vfmaddss and vfmaddsd
While these instructions accept memory operands, only one operand
may be a memory operand.  Giving two operands xm constraints gives
the compiler the option of using memory for both operands, which
would result in invalid assembly code.  Using x for all operands is
more appropriate, as most x86_64 calling conventions will pass the
arguments in registers anyway.

2013-05-15  Peter Collingbourne  <pcc@google.com>

	* sysdeps/x86_64/fpu/multiarch/s_fma.c (__fma_fma4): Replace xm
	constraints with x constraints.
	* sysdeps/x86_64/fpu/multiarch/s_fmaf.c (__fmaf_fma4): Likewise.
2013-05-15 20:31:53 +02:00
Peter Collingbourne
39b1f6172a Move _obstack_compat out of common
it is impossible to create an alias of a common symbol (as
compat_symbol does), because common symbols do not have a section or
an offset until linked.  GNU as tolerates aliases of common symbols by
simply creating another common symbol, but other assemblers (notably
LLVM's integrated assembler) are less tolerant.

2013-05-15  Peter Collingbourne  <pcc@google.com>

	* malloc/obstack.c (_obstack_compat): Add initializer.
-
2013-05-15 20:29:20 +02:00
Edjunior Machado
12fba011bd Update s390/bits/siginfo.h
2013-05-15  Edjunior Machado  <emachado@linux.vnet.ibm.com>

	* sysdeps/unix/sysv/linux/s390/bits/siginfo.h (siginfo_t):
	Remove si_trapno and add si_addr_lsb to _sifields.sigfault.
	(si_trapno): Remove macro.
	(si_addr_lsb): Define new macro.
	(BUS_MCEERR_AR, BUS_MCEERR_AO): Define new values.
2013-05-15 20:15:19 +02:00
Joseph Myers
b861c6c4c9 Consistently use TEST_f_L in tests of llrint and llround. 2013-05-15 16:58:41 +00:00
Joseph Myers
7abeee129e Convert TEST_f_b tests from code to data. 2013-05-15 12:04:07 +00:00
Joseph Myers
cbe8c4d333 Test for errno setting in more domain error cases. 2013-05-15 10:43:19 +00:00
Joseph Myers
5575c0e5cd Disable libm-test errno testing for TEST_INLINE. 2013-05-15 10:42:18 +00:00
Joseph Myers
d369f53178 Convert TEST_ff_i tests from code to data. 2013-05-15 10:41:39 +00:00
David S. Miller
781fd74820 Update Sparc ULPs.
* sysdeps/sparc/fpu/libm-test-ulps: Update.
2013-05-14 15:44:29 -07:00
Joseph Myers
5e90827081 Fix whitespace in RUN_TEST_LOOP_2_f. 2013-05-14 21:17:19 +00:00
Joseph Myers
f71172e551 Correct types of fields in libm-test.inc structures. 2013-05-14 21:00:56 +00:00
Joseph Myers
8cfa635ab8 Convert TEST_ffI_f1 tests from code to data. 2013-05-14 20:36:03 +00:00
Joseph Myers
2bcb36b265 Add test for setjmp / longjmp and floating-point state. 2013-05-14 19:49:09 +00:00
Joseph Myers
acbd839a07 Convert TEST_f_i tests from code to data. 2013-05-14 19:13:43 +00:00
Joseph Myers
e414d745ee Fix function name in scalbln_test call to END. 2013-05-14 19:12:54 +00:00
Joseph Myers
d9c2a0fd2c Convert TEST_fI_f1 tests from code to data. 2013-05-14 19:12:10 +00:00
Joseph Myers
9dc9095d56 Convert TEST_fF_f1 tests from code to data. 2013-05-14 14:49:26 +00:00
Joseph Myers
87aa21df6f Convert TEST_f_f1 tests from code to data. 2013-05-14 12:01:19 +00:00
Carlos O'Donell
141af660d8 Add comments to vDSO hwcap loading process.
Loading of the vDSO pseudo-hwcap from the type 2 GNU note is
a rather arcane and poorly documented process. Given that I had
a chance to review this code today I thought I would add all
of the things I had to lookup to verify the validity of the
process.

With a single .note.GNU the vDSO can register up to 64 flags,
though in practice you are limited to 64 - _DL_FIRST_EXTRA
bits which on x86 is 12 bits.

The only use of this that I know of is in the Xen support
in Linux where they use the 1st bit to indicate "nosegneg".
I see "We use bit 1 to avoid bugs in some versions of glibc
when bit 0 is used; the choice is otherwise arbitrary.", but
no reference to a glibc bug anywhere. The code as-is should
support bit zero, so we still have that free for future use.

The kernel, glibc, and ld.so.cache must coordinate to ensure
that bit values don't go too high and are used consistently.

---

2013-05-13  Carlos O'Donell  <carlos@redhat.com>

	* elf/dl-hwcaps.c (_dl_important_hwcaps): Comment vDSO hwcap loading.
	* elf/ldconfig.c (is_hwcap_platform): Comment each hwcap check.
	(main): Comment "tls" pseudo-hwcap.
2013-05-14 00:06:35 -04:00
Joseph Myers
68fc074ca0 Convert TEST_fl_f tests from code to data. 2013-05-13 22:38:15 +00:00
Joseph Myers
243216e126 Convert TEST_fi_f tests from code to data. 2013-05-13 19:45:36 +00:00
Joseph Myers
6a1992e268 Convert TEST_c_f tests from code to data. 2013-05-13 18:58:17 +00:00
Joseph Myers
0e400df55d Convert TEST_if_f tests from code to data. 2013-05-13 18:57:35 +00:00
Joseph Myers
6d33265c72 Consistently use TEST_fl_f in tests of scalbln. 2013-05-13 18:56:50 +00:00
Adhemerval Zanella
f98ece5fe2 Define decimal constant M_1_DIV_El in libm-test.inc
This patch replaces the 1.0 / M_El by the decimal constant M_1_DIV_El.
2013-05-13 12:08:42 -05:00
Siddhesh Poyarekar
43fe811b73 Use HP_TIMING for benchmarks if available
HP_TIMING uses native timestamping instructions if available, thus
greatly reducing the overhead of recording start and end times for
function calls.  For architectures that don't have HP_TIMING
available, we fall back to the clock_gettime bits.  One may also
override this by invoking the benchmark as follows:

  make USE_CLOCK_GETTIME=1 bench

and get the benchmark results using clock_gettime.  One has to do
`make bench-clean` to ensure that the benchmark programs are rebuilt.
2013-05-13 13:44:32 +05:30
Siddhesh Poyarekar
0f7d347bd0 Make _LIB_VERSION a weak symbol
That way it can live alongside _LIB_VERSION in libieee.a for
statically compiled programs.

Resolves #14582.
2013-05-13 11:46:36 +05:30
Joseph Myers
4f184d30cb Convert TEST_fff_f tests from code to data. 2013-05-12 21:07:58 +00:00