10216 Commits

Author SHA1 Message Date
davidxu
97a1cdabb8 Eliminate unused code. 2006-01-16 05:33:48 +00:00
jasone
4179e6a903 Add a separate simple internal base allocator and remove base_arena, so that
there is never any need to recursively call the main allocation functions.

Remove recursive spinlock support, since it is no longer needed.

Allow chunks to be as small as the page size.

Correctly propagate OOM errors from arena_new().
2006-01-16 05:13:49 +00:00
rwatson
2ad9f38f26 Replace strncpy() with strlcpy() when parsing login time limit strings
from /etc/login.conf, or an unterminated string buffer could result.
Probably, login_times.c should reject excessively long time strings as
unparseable, rather than truncating, which might render an invalid
string valid.

Found with:	Coverity Prevent (tm)
Reviewed by:	csjp
MFC after:	3 days
2006-01-16 00:52:20 +00:00
rwatson
0254c18ba2 Remove unnecessary and undesirable 'static' from function-local keg
list, which could cause problems for multi-threaded applications
using libmemstat to monitor UMA in more than one thread
simultaneously.

MFC after:	3 days
2006-01-16 00:37:20 +00:00
marcel
7a98306144 Define NO_TLS on ia64. The dynamic TLS implementation on ia64 is
broken for non-threaded shared processes in that __tls_get_addr()
assumes the thread pointer is always initialized. This is not the
case. When arenas_map is referenced in choose_arena() and it is
defined as a thread-local variable, it will result in a SIGSEGV.

PR: ia64/91846 (describes the TLS/ia64 bug).
2006-01-16 00:32:46 +00:00
rwatson
b4fd2a3c90 Fix typo in comment.
MFC after:	3 days
2006-01-16 00:28:11 +00:00
csjp
4c05286a61 Validate that the supplied file is not empty before trying mmap(2) it
and access the pages associated with it.

Submitted by:	Wojciech A. Koszek
PR:		bin/91411
MFC after:	1 week
2006-01-15 20:30:13 +00:00
davidxu
aedf24a684 Enable mutex inheritance code in mutex_fork, I forgot to turn on it.
while here, add some comments about process shared mutex.
2006-01-14 11:33:43 +00:00
jasone
46a2bcd398 Replace malloc(), calloc(), posix_memalign(), realloc(), and free() with
a scalable concurrent allocator implementation.

Reviewed by:	current@
Approved by:	phk, markm (mentor)
2006-01-13 18:38:56 +00:00
jasone
e30d92806f Increase the number of spinlocks, since libc's malloc implementation is
about to significantly increase the number of spinlocks used.

Approved by:	markm (mentor)
2006-01-13 06:14:04 +00:00
glebius
a8f06e9570 In the splnet(9) times netgraph(4) was synchronous and if a message
had been replied, the reply was always delivered to the originator
synchronously.

With introduction of netgraph item callbacks and a wait channel with
mutex in ng_socket(4), we have fixed the problem with ngctl(8) returning
earlier than the command has been proceeded by target node. But still
ngctl(8) can return prior to the reply has arrived to its node.

To fix this:
 - Introduce a new flag for netgraph(4) messages - NGM_HASREPLY.
   This flag is or'ed with message like NGM_READONLY.
 - In netgraph userland library if we have sent a message with
   NGM_HASREPLY flag, then select(2) until reply comes.
 - Mark appropriate generic commands with NGM_HASREPLY flag,
   gathering them into one enum {}. Bump generic cookie.
2006-01-12 19:14:40 +00:00
jasone
658b743414 Fix a bitwise logic error in posix_memalign().
Reported by:	glebius
2006-01-12 18:09:25 +00:00
jasone
958adbcf18 Use posix_memalign() in valloc() rather than making assumptions about
the alignment of malloc()ed memory.

Approved by:	markm (mentor)
2006-01-12 09:29:38 +00:00
jasone
3668a2e494 In preparation for a new malloc implementation:
* Add posix_memalign().

  * Move calloc() from calloc.c to malloc.c.  Add a calloc() implementation in
    rtld-elf in order to make the loader happy (even though calloc() isn't
    used in rtld-elf).

  * Add _malloc_prefork() and _malloc_postfork(), and use them instead of
    directly manipulating __malloc_lock.

Approved by:	phk, markm (mentor)
2006-01-12 07:28:21 +00:00
ru
40e5b400fd Build shared library on behalf of bsnmpd. 2006-01-12 07:26:57 +00:00
ceri
d80ea38223 I wrote getnetconfig where I meant getnetpath in the previous revision. 2006-01-11 13:57:15 +00:00
davidxu
84187b477c The thr_new sysscall was already in libc, don't generate it. 2006-01-11 06:10:05 +00:00
grog
d8f7824ea7 Add references to fhopen, fhstat, getfh, lgetfh and fhstatfs.
Pointed out by: Antony Curtis <antony@mysql.com>
2006-01-10 23:24:47 +00:00
davidxu
1218aa5627 Use macro STATIC_LIB_REQUIRE to declare a symbol should be linked into
static binary.
2006-01-10 04:53:03 +00:00
davidxu
36bb818343 Rescue pthread_set_name_np for compatible reason, remove unused code. 2006-01-09 08:07:22 +00:00
davidxu
8798f9bcec Tweak macro THR_LOCK_RELEASE a bit for non-PTHREAD_INVARIANTS case. 2006-01-09 07:32:22 +00:00
davidxu
db9594be6e Return real detached state. 2006-01-09 03:59:51 +00:00
davidxu
6f27a53b6f Fix a bug recently introduced, the _thread_active_count should be
decreased if thread can not be created.
2006-01-08 10:13:18 +00:00
davidxu
d1f7b6d605 Allow background threads to be suspended. 2006-01-08 01:49:31 +00:00
davidxu
4c4f339838 Try to reduce total time needed for suspending all threads,
first broadcast signals to all threads, then enter a wait loop.
2006-01-08 01:48:51 +00:00
davidxu
f04292bb19 Remove functions i386_get_gsbase and i386_set_gsbase, they were already
in libc.
2006-01-07 06:01:43 +00:00
ceri
4882287a5e o Document the possibility of putting 'b' in the flag field.
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.
2006-01-06 19:39:16 +00:00
davidxu
d6c88c0f27 Refine thread suspension code, now thread suspension is a blockable
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.
2006-01-05 13:51:22 +00:00
bde
76448b8654 Oops, on amd64 (and probably on all non-i386 systems), the previous
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.
2006-01-05 09:18:48 +00:00
dds
eab4be14c1 Document the recently-added EINVAL behavior.
MFC after:	1 week
2006-01-05 08:55:56 +00:00
bde
cbc5231d53 Use double precision internally to optimize cbrtf(), and change the
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.
2006-01-05 07:57:31 +00:00
davidxu
45180066f9 1. Add SIGEV_THREAD notification for mq_notify.
2. Reuse current timer code and abstract some common code to
   to support both timer and mqueue.
2006-01-04 11:48:02 +00:00
brian
9105db3d27 For the ``#ifdef NOTYET'' code that allows calling non-async-safe
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
2006-01-03 15:34:27 +00:00
davidxu
46ad9344c2 Remove in-progress wait code to sync with libpthread's behavior. 2006-01-03 13:30:23 +00:00
dfr
9eb9045215 Use the correct shared-library version number (the same as the one used
by the standalone version of heimdal GSS-API). If any compat issues
arise, I may increment the version number once more.
2006-01-01 11:01:01 +00:00
csjp
3c839b2489 Document the LOGIN_SETMAC setusercontext(3) flag. While we are here, drop
in an external reference to mac_set_proc(3).
2005-12-30 06:16:20 +00:00
dfr
d9cbcb50b5 Add a new extensible GSS-API layer which can support GSS-API plugins,
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)
2005-12-29 14:40:22 +00:00
dougb
5356ece7af Updated versions of header files generated per the instructions
in src/contrib/bind9/FREEBSD-Upgrade for the 9.2.3 import
2005-12-29 04:29:03 +00:00
grehan
ac8a981b11 gmon now supported on powerpc 2005-12-29 04:10:52 +00:00
grehan
2153c7c91d The minbrk symbol is hidden the same on powerpc as other FreeBSD platforms. 2005-12-29 04:09:38 +00:00
trhodes
8190ceb049 Add a64l(), l64a(), and l64a_r() XSI extentions. These functions convert
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
2005-12-24 22:37:59 +00:00
phk
62bc67a619 Add abort2 manual page.
Submitted by:	"Wojciech A. Koszek" <dunstan@freebsd.czest.pl>
Edited by:	phk
2005-12-23 12:27:42 +00:00
cognet
2cf710a449 Explicitely use a "signed char" instead of a "char", for those archs where
char defaults to unsigned.
2005-12-22 14:23:54 +00:00
davidxu
2d34e45c94 Let _mutex_cv_lock call internal functiona mutex_lock_common. 2005-12-21 05:14:07 +00:00
davidxu
5ac55e9c46 Hide umtx API symbols as well. 2005-12-21 03:53:29 +00:00
davidxu
8d9e23fb51 1. Retire macro SCLASS, instead simply use language keyword and
put variables in thr_init.c.
2. Hide all global symbols which won't be exported.
2005-12-21 03:14:06 +00:00
davidxu
3e60ee73cb Follow the mistake in libpthread, the first version name in libpthread
is LIBTHREAD_1_0, but really it should be LIBPTHREAD_1_0.
Fix it so libmap.conf works again (it was broken by recent versioning
code in rtld_elf).
2005-12-21 02:34:20 +00:00
bde
761a5296f9 Extract the high and low words together. With gcc-3.4 on uniformly
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.
2005-12-20 01:21:30 +00:00
davidxu
a8cc7e07c2 Clear return code to zero if joiner successfully waited joinee.
Bug reported by: jasone at connonware when using ports lang/onyx
MFC after: 3 days
2005-12-19 03:20:55 +00:00
bde
ce5f09f38d Use a minimax polynomial approximation instead of a Pade rational
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).
2005-12-19 00:22:03 +00:00