Commit Graph

115168 Commits

Author SHA1 Message Date
Robert Watson
5d1ae027f0 Modify UMA to use critical sections to protect per-CPU caches, rather than
mutexes, which offers lower overhead on both UP and SMP.  When allocating
from or freeing to the per-cpu cache, without INVARIANTS enabled, we now
no longer perform any mutex operations, which offers a 1%-3% performance
improvement in a variety of micro-benchmarks.  We rely on critical
sections to prevent (a) preemption resulting in reentrant access to UMA on
a single CPU, and (b) migration of the thread during access.  In the event
we need to go back to the zone for a new bucket, we release the critical
section to acquire the global zone mutex, and must re-acquire the critical
section and re-evaluate which cache we are accessing in case migration has
occured, or circumstances have changed in the current cache.

Per-CPU cache statistics are now gathered lock-free by the sysctl, which
can result in small races in statistics reporting for caches.

Reviewed by:	bmilekic, jeff (somewhat)
Tested by:	rwatson, kris, gnn, scottl, mike at sentex dot net, others
2005-04-29 18:56:36 +00:00
David E. O'Brien
97c1144eb2 Enable dc(4) and rl(4) in the PAE kernel.
Both have been busdma'ed for use and tested in the Sparc64 kenrel.
2005-04-29 18:12:22 +00:00
Hajimu UMEMOTO
86726f325b NETDB_INTERNAL is not fit, here. return NO_RECOVERY for h_errno. 2005-04-29 17:36:54 +00:00
Hartmut Brandt
be7a5506e8 Rename the directive_hash.[ch] files to hash_tables.[ch] now
that there are more than one hash table in them. There is no
history to preserve here, so go without a repo-copy.

Asked for by:	Max Okumoto <okumoto@ucsd.edu>
2005-04-29 15:15:28 +00:00
Brooks Davis
e32931227c Remove the Netgear WG311 from the list. The v1 card is supported, but
the v2 card is a TI.  Since we're not attempting to keep this list
complete, removing this is best.

Reported by:	Brian Candler <B dot Candler at pobox dot com>
		Evan Dower <evantd at hotmail dot com>
MFC After:	1 day
2005-04-29 15:07:44 +00:00
Hartmut Brandt
dea3d8703e Toggle on warnings. This resolves the problem with building old releases
(getting zillions of warnings). Building an old release uses that release's
sys.mk which does not switch on these warnings, so make will be silent.
They can be switch on on the command line with the -x option to make.
This has been tested by building RELENG_5_4 on CURRENT.
2005-04-29 14:44:10 +00:00
Hajimu UMEMOTO
49e0283df8 our get{addr,name}info() is considered thread-safe. 2005-04-29 14:37:51 +00:00
Hartmut Brandt
3dece030a5 Implement a pseudo-target .WARN that allows toggeling the warning flags
for the current make. This does not override flags specified on the
command line and these settings are not passed to sub-makes.
2005-04-29 14:37:44 +00:00
Hajimu UMEMOTO
0d389e7212 we cannot use inet_ntoa(3), here. so, use inet_ntop(3), instead. 2005-04-29 12:01:36 +00:00
Søren Schmidt
e47099593a Now that probing is working in the new fashion, we need to go back to
having ata_getparm issue an ata_request and not fool around with the HW
on its own.
Needed for new HW support.
2005-04-29 11:30:03 +00:00
Maxim Konovalov
fa6c8d89c9 o Typo: s/teminal/terminal/.
Submitted by:	Michal Varga
2005-04-29 11:10:27 +00:00
Hartmut Brandt
5710486746 This commit was generated by cvs2svn to compensate for changes in r145673,
which included commits to RCS files with non-trunk default branches.
2005-04-29 11:02:18 +00:00
Hartmut Brandt
a9bfedb789 Vendor patch: fix a bug that was introduced when moving from libisc to
libbegemot: the rpoll_start_timer function needs the timeout value in
milli-seconds, not the absolute time when the timer should tick.
2005-04-29 11:02:18 +00:00
Robert Drehmel
7f31354152 Add flag to choose whether to use getgrouplist(3) or getgroups(2)
to the id_print() function.

Use getgrouplist(3) for the case when an user was specified,
and getgroups(2) when no user was given.
That reverts to the expected behaviour and makes it easy to
implement an option later to force using getgrouplist(3).
2005-04-29 10:11:18 +00:00
Robert Drehmel
5b8e7b6b29 Fix spelling error. 2005-04-29 08:37:52 +00:00
Pawel Jakub Dawidek
760fde0677 Remove trailing spaces. 2005-04-29 07:57:50 +00:00
Warner Losh
4e30440d7c Add a detach for pci bridge and pci bus drivers. This allows one to
theoretically unload pci bridges or pci drivers.  It will also allow
detach to work if one needed to detach a subtree.

This is inspired by looking at the p4 commits from bms to his 5.4
tree, but I didn't look at the final results.
2005-04-29 06:22:41 +00:00
Ruslan Ermilov
f9067637f7 Fix the following warnings on amd64:
/usr/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter/netinet/ip_frag.c: In function `fr_ipid_newfrag':
/usr/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter/netinet/ip_frag.c:397: warning: cast to pointer from integer of different size
/usr/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter/netinet/ip_frag.c: In function `fr_ipid_knownfrag':
/usr/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter/netinet/ip_frag.c:582: warning: cast from pointer to integer of different size
2005-04-29 05:57:17 +00:00
Scott Long
a304e92b2b Only create the rdpti alias if the asr device creation succeeds. 2005-04-29 04:47:11 +00:00
Scott Long
a8e6f0b6fa For whatever reason, we don't allow USB on PAE. Since it's a dependency
for EHCI, exclude that driver also.
2005-04-29 02:59:19 +00:00
Scott Long
1c7e93581d Don't bother pretending that CAM will send CAM_DATA_PHYS pointers. It's
a concept that is fundamentally broken with PAE.
2005-04-29 02:58:23 +00:00
Scott Long
9fa98e70b0 Update the file.* entries for the new home of hwpmc 2005-04-29 02:40:16 +00:00
Marcel Moolenaar
d98d9b126b In pcib_alloc_resource() check if the resource allocation request is
for the VGA I/O or memory ranges, when it's not within the default
ranges decoded by the bridge. When allocation for VGA addresses is
attempted, check that the bridge has the VGA Enable bit set before
allowing it.
As such, newbusified VGA drivers can allocate their resources when
the VGA adapter is behind a PCI-to-PCI bridge.

Reviewed by: imp@, jhb@
2005-04-29 02:15:40 +00:00
Marcel Moolenaar
9929ff6b5b Add pci_is_vga_ioport_range() and pci_is_vga_memory_range() as inline
functions. These functions centralize the details of which I/O port
and memory ranges belong to VGA.

Reviewed by: imp@, jhb@
2005-04-29 02:03:11 +00:00
Marcel Moolenaar
d1ee178775 Add defines for the Bridge Control Register bits.
Obtained from: jhb@
2005-04-29 01:58:27 +00:00
Darren Reed
a63b0146dc IPFIlter problems that prevented building should all now be resolved so
remove this temporary measure.
2005-04-28 23:29:00 +00:00
Jesus R. Camou
cebee7e48e Update the manual page for ppp(8).
PR:		docs/78605
Submitted by:	John E. Hein <jhein@timing.com>
Approved by:	trhodes (mentor)
MFC after:	1 day
2005-04-28 22:31:37 +00:00
Søren Schmidt
ca8c70e9e3 Provide a default setmode method.
This shaves off multiple copies of the same setmode stub.
2005-04-28 22:15:44 +00:00
Søren Schmidt
8dad6b7be5 Rearrange the way the reset code is called.
Prepare for different looking controllers.
2005-04-28 22:08:08 +00:00
Darren Reed
e01de6cda3 Don't use quad_t on FreeBSD (deprecated) so use "long long" instead.
Someday this should be converted to uint64_t and printstate.c changed to
use those horrid PRiud64 things.
2005-04-28 21:36:30 +00:00
Stefan Farfeleder
66116c07a7 Revert the last change, the conversion from long double to double can raise
unwanted underflow exceptions.

Pointed out by:	das
2005-04-28 19:45:55 +00:00
Hajimu UMEMOTO
850bc9af6b sync _map_v4v6_host*() with bind9's. it treats align better bit.
Obtained from:	BIND9
2005-04-28 18:52:40 +00:00
Hajimu UMEMOTO
b8ab0f4506 we don't need mutex lock to call _gethostbynis*(), anymore. 2005-04-28 18:21:11 +00:00
Hajimu UMEMOTO
aa2f4ec72a make gethostby*() thread-safe. 2005-04-28 18:03:43 +00:00
Hajimu UMEMOTO
7b671d902b _gethostbynis{addr,name}() can handle an IPv6, now. 2005-04-28 17:44:44 +00:00
Darren Reed
0c3757df94 Patches from Ruslan Ermilov to address problems compiling LINT 2005-04-28 16:33:15 +00:00
Darren Reed
cf4ee2c933 Patches from Ruslam Ermilov to remove NetBSD bits from Makefiles and cleanup
build problems with rescue.
2005-04-28 16:26:35 +00:00
Robert Drehmel
58b6761db3 We need to pass NGROUPS + 1 to getgrouplist(3) to display
NGROUPS groups.  getgrouplist(3) may put a duplicate group
id into the passed array (it sets [0] and [1] to the value
of the gid argument), but id_print() sorts them out.
2005-04-28 16:11:37 +00:00
Robert Drehmel
68b9b81e79 - Merge two functions for printing `id' output.
Showing the ids of both an user given by an argument to `id',
   and the current user, is now handled in a single function.
   Displaying the current user's ids was inaccurate because
   getgroups(2) had been used.  getgroups(2) returns the current
   kernel state of a user's groups, which may not always be
   correct if /etc/group was recently changed.
 - Fix a few style bugs.

PR:		bin/78085
2005-04-28 15:55:54 +00:00
Hartmut Brandt
48c49ace00 Introduce a flag to enable extended warnings (-x) and make them off
by default. This should fix the problem of getting lots of errors
when building with an up-to-date make and old *.mk files.
2005-04-28 15:37:25 +00:00
Hajimu UMEMOTO
a2a775011c make getnetby*() thread-safe. 2005-04-28 15:32:55 +00:00
Ruslan Ermilov
1afc807b33 Be resistant to DESTDIR being set to some unkosher value, e.g. "/".
Noticed by:	Steve Ames
2005-04-28 14:45:57 +00:00
Paul Saab
4bb10cf193 Add support for the P600 and name the E400. 2005-04-28 14:40:23 +00:00
Hartmut Brandt
1b26a92962 Add an initial regression test suite for make(1).
Submitted by:	Max Okumoto <okumotu@ucsd.edu> (partly, initial version)
2005-04-28 13:20:48 +00:00
Lukas Ertl
bc2d4d6784 Only allow RAID5 plexes to be parity checked.
PR:           kern/80427
Submitty by:  Stijn Hoop <stijn@win.tue.nl>
2005-04-28 13:09:00 +00:00
Robert Drehmel
989090c052 - Act according to the documentation (man page):
When adding users from a preformatted file, do not exit
   silently when empty lines or lines starting with a '#'
   are encountered - ignore them instead.
 - Fix a spelling error in a comment.

PR:		bin/80058
2005-04-28 12:44:02 +00:00
Robert Drehmel
92bc8d4e8e - Set negative row or column argument values to zero.
Negative values would produce undefined behaviour including
   a possible segmentation fault.
 - Explicitly initialize the global row and column variables
   to zero.

PR:		bin/80348
2005-04-28 12:37:15 +00:00
Hartmut Brandt
b1880272f1 Use a minimal perfect hash for the special sources/targets too. Add
the corresponding magic to create the hash function to the Makefile.
2005-04-28 12:05:43 +00:00
Joseph Koshy
6b8c8cd85f Return the correct register number in the 'get_msr()' MD function.
Only allow a process to use the x86 RDPMC instruction if it has
allocated and attached a PMC to itself.

Inform the MD layer of the "pseudo context switch out" that needs
to be done when the last thread of a process is exiting.
2005-04-28 08:13:19 +00:00
Hartmut Brandt
47cc56ab55 Fix the $$FreeBSD$$ that should be written verbatim to directive_hash.c
so that it doesn't get replaced with the Makefile's revision.
2005-04-28 08:04:46 +00:00