Commit Graph

5671 Commits

Author SHA1 Message Date
bde
c553ad248f Add an alternative view of the bits in an 80-bit long double (64+16
instead of 32+32+15+1) on all arches that have such long doubles (amd64,
ia64 and i386).  Large objects should be be accessed in large units,
and the 32+32+15+1[+padding] decomposition asks for almost the opposite
of that, sometimes resulting in very slow accesses depending on how
well the compiler ignores what we ask for and converts to the best
units for the given machine.  E.g., on Athlons, there is a 10-20 cycle
penalty for accessing the middle 32-bit word immediately after an
80-bit store.

Whether actually using the alternative view is better is very machine-
dependent.  A 32+32+16 view is probably best with old 32-bit systems
and gcc through 4.2.1.  The compiler should mostly avoid the view and
generate best accesses, but gcc-4.2.1 is far from doing that.  I think
64+16 is best for now.  Similarly for doubles -- they should be using
64+0 especially on 64-bit machines, but fdlibm uses 32+32 extensively
for them.  Fortunately, in 64-bit mode for doubles, gcc already ignores
the 32+32-bit view and generates best accesses in many cases.
2008-01-17 16:39:07 +00:00
remko
af0f4dc1e0 Fix some style nits.
Prodded by:	brueffer
MFC After:	3 days
2008-01-16 19:36:21 +00:00
jhb
184b0a421c Remove some now-unused macros.
MFC after:	1 week
2008-01-15 18:55:52 +00:00
jhb
c02890da0b Put back the openpty(3) and ptsname(3) fixes but don't disable ptsname(3)
on pts(4) devices this time.  This fixes the issues while leaving pts(4)
enabled on HEAD.
2008-01-15 15:36:23 +00:00
cperciva
2f49f42d98 Back out last commit, since it accidentally broke pts.
The security fix will be re-committed soon, hopefully without breaking
anything.
2008-01-15 13:59:13 +00:00
das
c41cd4bf35 In getttyent(3), if /etc/ttys doesn't end in a newline, don't
freak out and keep trying to expand the buffer until realloc()
fails.

PR:	114398
2008-01-15 06:50:50 +00:00
cperciva
533f13b8b2 Fix issues which allow snooping on ptys. [08:01]
Fix an off-by-one error in inet_network(3). [08:02]

Security: FreeBSD-SA-08:01.pty
Security: FreeBSD-SA-08:02.libc
2008-01-14 22:56:05 +00:00
das
00c36da743 Changing 'r' to a size_t in the previous commit turned quicksort
into slowsort for some sequences because different parts of the
code used 'r' to store two different things, one of which was
signed. Clean things up by splitting 'r' into two variables, and
use a more meaningful name.
2008-01-14 09:21:34 +00:00
das
1daf1db8d4 Use size_t to avoid overflow when sorting arrays larger than 2 GB.
PR:		111085
MFC after:	2 weeks
2008-01-13 02:11:10 +00:00
jhb
34de26d108 Add a feature_present(3) function which checks to see if a named kernel
feature is present by checking the kern.features sysctl MIB.

MFC after:	1 week
2008-01-10 22:11:21 +00:00
jhb
8cd9437636 Add a new file descriptor type for IPC shared memory objects and use it to
implement shm_open(2) and shm_unlink(2) in the kernel:
- Each shared memory file descriptor is associated with a swap-backed vm
  object which provides the backing store.  Each descriptor starts off with
  a size of zero, but the size can be altered via ftruncate(2).  The shared
  memory file descriptors also support fstat(2).  read(2), write(2),
  ioctl(2), select(2), poll(2), and kevent(2) are not supported on shared
  memory file descriptors.
- shm_open(2) and shm_unlink(2) are now implemented as system calls that
  manage shared memory file descriptors.  The virtual namespace that maps
  pathnames to shared memory file descriptors is implemented as a hash
  table where the hash key is generated via the 32-bit Fowler/Noll/Vo hash
  of the pathname.
- As an extension, the constant 'SHM_ANON' may be specified in place of the
  path argument to shm_open(2).  In this case, an unnamed shared memory
  file descriptor will be created similar to the IPC_PRIVATE key for
  shmget(2).  Note that the shared memory object can still be shared among
  processes by sharing the file descriptor via fork(2) or sendmsg(2), but
  it is unnamed.  This effectively serves to implement the getmemfd() idea
  bandied about the lists several times over the years.
- The backing store for shared memory file descriptors are garbage
  collected when they are not referenced by any open file descriptors or
  the shm_open(2) virtual namespace.

Submitted by:	dillon, peter (previous versions)
Submitted by:	rwatson (I based this on his version)
Reviewed by:	alc (suggested converting getmemfd() to shm_open())
2008-01-08 21:58:16 +00:00
jasone
6aea4f4c16 Enable both sbrk(2)- and mmap(2)-based memory acquisition methods by
default.  This has the disadvantage of rendering the datasize resource
limit irrelevant, but without this change, legitimate uses of more
memory than will fit in the data segment are thwarted by default.

Fix chunk_alloc_mmap() to work correctly if initial mapping is not
chunk-aligned and mapping extension fails.
2008-01-03 23:22:13 +00:00
jasone
573b21b457 Fix a major chunk-related memory leak in chunk_dealloc_dss_record(). [1]
Clean up DSS-related locking and protect all pertinent variables with
dss_mtx (remove dss_chunks_mtx).  This fixes race conditions that could
cause chunk leaks.

Reported by:	[1] kris
2007-12-31 06:19:48 +00:00
jasone
1f55b95c0b Fix a bug related to sbrk() calls that could cause address space leaks.
This is a long-standing bug, but until recent changes it was difficult
to trigger, and even then its impact was non-catastrophic, with the
exception of revision 1.157.

Optimize chunk_alloc_mmap() to avoid the need for unmapping pages in the
common case.  Thanks go to Kris Kennaway for a patch that inspired this
change.

Do not maintain a record of previously mmap'ed chunk address ranges.
The original intent was to avoid the extra system call overhead in
chunk_alloc_mmap(), which is no longer a concern.  This also allows some
simplifications for the tree of unused DSS chunks.

Introduce huge_mtx and dss_chunks_mtx to replace chunks_mtx.  There was
no compelling reason to use the same mutex for these disjoint purposes.

Avoid memset() for huge allocations when possible.

Maintain two trees instead of one for tracking unused DSS address
ranges.  This allows scalable allocation of multi-chunk huge objects in
the DSS.  Previously, multi-chunk huge allocation requests failed if the
DSS could not be extended.
2007-12-31 00:59:16 +00:00
jasone
138c627354 Back out premature commit of previous version. 2007-12-28 09:21:12 +00:00
jasone
c907683663 Maintain two trees instead of one (old_chunks --> old_chunks_{ad,szad}) in
order to support re-use of multi-chunk unused regions within the DSS for
huge allocations.  This generalization is important to correct function
when mmap-based allocation is disabled.

Avoid zeroing re-used memory in the DSS unless it really needs to be
zeroed.
2007-12-28 07:24:19 +00:00
jasone
93c7f7517b Release chunks_mtx for all paths through chunk_dealloc().
Reported by:	kris
2007-12-28 02:15:08 +00:00
jasone
15ff969441 Add the 'D' and 'M' run time options, and use them to control whether
memory is acquired from the system via sbrk(2) and/or mmap(2).  By default,
use sbrk(2) only, in order to support traditional use of resource limits.
Additionally, when both options are enabled, prefer the data segment to
anonymous mappings, in order to coexist better with large file mappings
in applications on 32-bit platforms.  This change has the potential to
increase memory fragmentation due to the linear nature of the data
segment, but from a performance perspective this is mitigated by the use
of madvise(2). [1]

Add the ability to interpret integer prefixes in MALLOC_OPTIONS
processing.  For example, MALLOC_OPTIONS=lllllllll can now be specified as
MALLOC_OPTIONS=9l.

Reported by:	[1] rwatson
Design review:	[1] alc, peter, rwatson
2007-12-27 23:29:44 +00:00
jhb
3db380a937 Fix a typo in regards to the ENOENT error.
PR:		docs/118929
Submitted by:	mymtom of hotmail
MFC after:	3 days
2007-12-27 21:55:49 +00:00
jhb
590aeb53cc Clean up some of the pts(4) vs pty(4) stuff in grantpt(3) and friends:
- Use PTY* for all pty(4) related constants.
- Use PTMX* for all pts(4) related constants.
- Consistently use _PATH_DEV PTMX rather than "/dev/ptmx".
- Revert 1.7 and properly fix it by using the correct prefix string for
  pts(4) masters.

MFC after:	3 days
2007-12-21 21:26:08 +00:00
imp
d140ab1c44 Reduce lock contention for simple cases.
# this really should be done with pthread_once, but I've debugged this code.

Reviewed by: arch@
2007-12-19 04:30:10 +00:00
imp
7d62c26d32 Add note about other systems. 2007-12-19 03:33:13 +00:00
das
a2bd2e60b5 Move all the xprintf-related symbols to FBSDprivate_1.0.
Discussed with:	deischen, kan, phk
2007-12-18 23:49:05 +00:00
das
ac3245defa Since nan() is supposed to work the same as strtod("nan(...)", NULL),
my original implementation made both use the same code. Unfortunately,
this meant libm depended on a vendor header at compile time and previously-
unexposed vendor bits in libc at runtime.

Hence, I just wrote my own version of the relevant vendor routine. As it
turns out, mine has a factor of 8 fewer of lines of code, and is a bit more
readable anyway. The strtod() and *scanf() routines still use vendor code.

Reviewed by:	bde
2007-12-18 23:46:32 +00:00
jasone
b720912697 Use fixed point integer math instead of floating point math when
calculating run sizes.  Use of the floating point unit was a potential
pessimization to context switching for applications that do not otherwise
use floating point math. [1]

Reformat cpp macro-related comments to improve consistency.

Submitted by:	das
2007-12-18 05:27:57 +00:00
das
b5d3ffed42 Catch up with vfprintf.c,v 1.77. 2007-12-18 01:20:33 +00:00
bushman
9d2bf43707 Moved logging out of the nss_method_lookup() in order not to
flood logs with failed fallback method lookup attempts.
2007-12-17 16:12:57 +00:00
jasone
65edaae0a7 Refactor features a bit in order to make it possible to disable lazy
deallocation and dynamic load balancing via the MALLOC_LAZY_FREE and
MALLOC_BALANCE knobs.  This is a non-functional change, since these
features are still enabled when possible.

Clean up a few things that more pedantic compiler settings would cause
complaints over.
2007-12-17 01:20:04 +00:00
das
bb384eba43 Implement and document nan(), nanf(), and nanl(). This commit
adds two new directories in msun: ld80 and ld128. These are for
long double functions specific to the 80-bit long double format
used on x86-derived architectures, and the 128-bit format used on
sparc64, respectively.
2007-12-16 21:19:28 +00:00
das
d3ab161cfb Export gdtoa's __ULto{x,Q}_D2A routine in a private namespace so
libm can use it.
2007-12-16 21:15:57 +00:00
das
60328a8764 Arrange so that the NaN returned by strtod("nan", NULL) is the same as
the NaN returned by strtod("nan()", NULL).
2007-12-16 21:15:09 +00:00
deischen
85b00ba4b1 Increment the version namespace for 8.0-current. New symbols and
symbols whose ABI has changed should be added to FBSD_1.1.
2007-12-14 20:49:06 +00:00
jhb
4a8d2d68f1 Update posix_openpt(3) to handle 512 ptys. This was missed in the earlier
pty(4) changes.

MFC after:	3 days
2007-12-13 00:08:59 +00:00
wes
9c13fe545e Remove license clause 3 to agree with the now-standard BSD license.
Prompted by:	Glenn Halperin, Symbian Software
2007-12-12 18:33:06 +00:00
bushman
b02556dae5 Implementing 'fallback' nsswitch source. 'fallback' source is used
when particular function can't be found in nsswitch-module. For
example, getgrouplist(3) will use module-supplied 'getgroupmembership'
function (which can work in an optimal way for such source as LDAP) and
will fall back to the stanard iterate-through-all-groups implementation
otherwise.

PR:		ports/114655
Submitted by:	Michael Hanselmann <freebsd AT hansmi DOT ch>
Reviewed by:	brooks (mentor)
2007-12-12 10:08:03 +00:00
phantom
b52b3d4702 Remove 3rd clause of license
Per request of: glenn halperin at symbian.com
2007-12-12 07:43:23 +00:00
ache
7ec3418f8e Fix typo in the comment 2007-12-11 20:39:32 +00:00
das
ec6292aa68 Remove some test instrumentation. (The Symbol.map changes broke it anyway.) 2007-12-09 21:00:12 +00:00
das
0ae05e01f7 Fix handling of subnormals on i386/ia64/amd64.
PR:	85080
2007-12-09 19:48:57 +00:00
remko
f956375817 Make the warning a bit less 'broad' then it used to be. The access
is seems to be a problem for SUID applications, which we like to
prevent as much as possible.

PR:		docs/39530
Submitted by:	Soren Spies <sspies at apple dot com>
MFC After:	3 days
2007-12-08 22:50:35 +00:00
ume
b787f17ecc Merge BIND 9.4.2 into main chunk. 2007-12-03 15:13:44 +00:00
ume
c98e1a2037 This commit was generated by cvs2svn to compensate for changes in r174223,
which included commits to RCS files with non-trunk default branches.
2007-12-03 15:07:58 +00:00
ume
488e5a6dee Vendor import of BIND 9.4.2 2007-12-03 15:07:58 +00:00
des
e6b2882442 Since jb@ fixed the type of dd_lock in <dirent.h>, these casts are no
longer required.
2007-12-03 14:33:51 +00:00
das
3f2a2dba4d In scanf, round according to the current rounding mode. 2007-12-03 07:17:33 +00:00
jasone
f7cf554c6b Only zero large allocations when necessary (for calloc()). 2007-11-28 00:17:34 +00:00
jasone
a32c67c9ab Add _pthread_mutex_init_calloc_cb to libc's map, for which malloc defines
a stub.
2007-11-27 16:22:21 +00:00
jasone
afb1c25b9f Document the B and L MALLOC_OPTIONS. 2007-11-27 03:18:26 +00:00
jasone
4db288d2b3 Implement dynamic load balancing of thread-->arena mapping, based on lock
contention.  The intent is to dynamically adjust to load imbalances, which
can cause severe contention.

Use pthread mutexes where possible instead of libc "spinlocks" (they aren't
actually spin locks).  Conceptually, this change is meant only to support
the dynamic load balancing code by enabling the use of spin locks, but it
has the added apparent benefit of substantially improving performance due to
reduced context switches when there is moderate arena lock contention.

Proper tuning parameter configuration for this change is a finicky business,
and it is very much machine-dependent.  One seemingly promising solution
would be to run a tuning program during operating system installation that
computes appropriate settings for load balancing.  (The pthreads adaptive
spin locks should probably be similarly tuned.)
2007-11-27 03:17:30 +00:00
jasone
2dd595aefe Implement lazy deallocation of small objects. For each arena, maintain a
vector of slots for lazily freed objects.  For each deallocation, before
doing the hard work of locking the arena and deallocating, try several times
to randomly insert the object into the vector using atomic operations.

This approach is particularly effective at reducing contention for
multi-threaded applications that use the producer-consumer model, wherein
one producer thread allocates objects, then multiple consumer threads
deallocate those objects.
2007-11-27 03:13:15 +00:00