]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
17 years agoNew supported JMicron, ALI and ServerWorks chipsets.
Christian Brueffer [Thu, 4 Jan 2007 09:12:03 +0000 (09:12 +0000)]
New supported JMicron, ALI and ServerWorks chipsets.

MFC after: 3 days

17 years agoULE 2.0:
Jeff Roberson [Thu, 4 Jan 2007 08:56:25 +0000 (08:56 +0000)]
ULE 2.0:
 - Remove the double queue mechanism for timeshare threads.  It was slow
   due to excess cache lines in play, caused suboptimal scheduling behavior
   with niced and other non-interactive processes, complicated priority
   lending, etc.
 - Use a circular queue with a floating starting index for timeshare threads.
   Enforces fairness by moving the insertion point closer to threads with
   worse priorities over time.
 - Give interactive timeshare threads real-time user-space priorities and
   place them on the realtime/ithd queue.
 - Select non-interactive timeshare thread priorities based on their cpu
   utilization over the last 10 seconds combined with the nice value.  This
   gives us more sane priorities and behavior in a loaded system as
   compared to the old method of using the interactivity score.  The
   interactive score quickly hit a ceiling if threads were non-interactive
   and penalized new hog threads.
 - Use one slice size for all threads.  The slice is not currently
   dynamically set to adjust scheduling behavior of different threads.
 - Add some new sysctls for scheduling parameters.

Bug fixes/Clean up:
 - Fix zeroing of td_sched after initialization in sched_fork_thread() caused
   by recent ksegrp removal.
 - Fix KSE interactivity issues related to frequent forking and exiting of
   kse threads.  We simply disable the penalty for thread creation and exit
   for kse threads.
 - Cleanup the cpu estimator by using tickincr here as well.  Keep ticks and
   ltick/ftick in the same frequency.  Previously ticks were stathz and
   others were hz.
 - Lots of new and updated comments.
 - Many many others.

Tested on: up x86/amd64, 8way amd64.

17 years ago - Add three new functions to support circular run queues.
Jeff Roberson [Thu, 4 Jan 2007 08:39:58 +0000 (08:39 +0000)]
 - Add three new functions to support circular run queues.
 - runq_add_pri allows the caller to position the thread at any rqindex
   regardless of priority.
 - runq_choose_from() chooses the lowest priority thread starting from a given
   index.  The index is updated with the rqindex of the chosen thread.  This
   routine is used to pick the lowest priority relative to a given index.
 - runq_remove_idx() updates the index if the run queue that held the removed
   thread is now empty.

17 years ago - Add SRQ_BORROWING to the list of switch reasons. ULE is the only consumer
Jeff Roberson [Thu, 4 Jan 2007 08:36:40 +0000 (08:36 +0000)]
 - Add SRQ_BORROWING to the list of switch reasons.  ULE is the only consumer
   at this time.  It is used to optimize the run queue placement of threads
   who have newly elevated priorities.

17 years agoAdd new PATA only JMicron chip.
Søren Schmidt [Thu, 4 Jan 2007 07:56:12 +0000 (07:56 +0000)]
Add new PATA only JMicron chip.

17 years agoAdd sockipprotoname() function. Decode the third parameter (protocol)
Craig Rodrigues [Thu, 4 Jan 2007 04:46:59 +0000 (04:46 +0000)]
Add sockipprotoname() function.  Decode the third parameter (protocol)
of a socket() call with sockipprotoname() if the first parameter (domain)
is PF_INET or PF_INET6.

Old parsing behavior before this change:
ping6    CALL  socket(PF_INET6,SOCK_RAW,0x3a)

New behavior after this change:
ping6    CALL  socket(PF_INET6,SOCK_RAW,IPPROTO_ICMPV6)

17 years agoGenerate sockdomainname() function with auto_if_type() instead
Craig Rodrigues [Thu, 4 Jan 2007 04:28:17 +0000 (04:28 +0000)]
Generate sockdomainname() function with auto_if_type() instead
of auto_or_type.

The old parsing code would incorrectly decode a socket() call in the
ping6 program as:
  CALL  socket(PF_PUP|PF_ECMA|PF_APPLETALK|PF_COIP|PF_SIP,SOCK_DGRAM,0)

The new parsing code decodes the same socket() call as:
  CALL  socket(PF_INET6,SOCK_DGRAM,0)

17 years agoAdd auto_if_type() function, which is similar to auto_switch_type().
Craig Rodrigues [Thu, 4 Jan 2007 04:18:03 +0000 (04:18 +0000)]
Add auto_if_type() function, which is similar to auto_switch_type().
However, auto_if_type() uses if/else statements in C instead
of a single switch statement, when mapping an integer value to
a #define.  For certain cases where multiple #define constants
alias to a single integer value, auto_if_type() makes things easier
to parse than auto_switch_type().

17 years agoRewrite and simplify archive_read_format_tar_skip. Compression-layer skip
Colin Percival [Wed, 3 Jan 2007 21:47:35 +0000 (21:47 +0000)]
Rewrite and simplify archive_read_format_tar_skip.  Compression-layer skip
functions are required to skip the requested distance, so we can avoid
lots of bookkeeping which would otherwise be necessary.

Reviewed by: kientzle
MFC after: 1 week

17 years agoRemove the test that ensures that when the string "nan(...)" is converted
David Schultz [Wed, 3 Jan 2007 21:28:26 +0000 (21:28 +0000)]
Remove the test that ensures that when the string "nan(...)" is converted
to floating-point, the result is a quiet NaN.  The current implementation
may return a signaling NaN, and the vendor has no plans for changing this,
for reasons explained in the comment I added.

17 years agoQuiet a compiler warning where (off_t + size_t) has a different sign
Colin Percival [Wed, 3 Jan 2007 20:01:44 +0000 (20:01 +0000)]
Quiet a compiler warning where (off_t + size_t) has a different sign
from (off_t).

Submitted by: delphij

17 years agoInsert zero-padding between sparse blocks in archive_read_data(). This
Colin Percival [Wed, 3 Jan 2007 13:16:59 +0000 (13:16 +0000)]
Insert zero-padding between sparse blocks in archive_read_data().  This
fixes "tar -c @foo.tar" where "foo.tar" contains sparse entries.

MFC after: 1 week

17 years agoWrap ipfw nat support in a new kernel config option named
Paolo Pisati [Wed, 3 Jan 2007 11:12:54 +0000 (11:12 +0000)]
Wrap ipfw nat support in a new kernel config option named
"IPFIREWALL_NAT": this way nat is turned off by default and
POLA is preserved.

Reviewed by: rwatson

17 years agoprotect against multiple inclusion (this is useful when you
Luigi Rizzo [Wed, 3 Jan 2007 10:50:03 +0000 (10:50 +0000)]
protect against multiple inclusion (this is useful when you
start working with third party usb modules, where sometimes it
is not easy to set the inclusion order so that there are no multiple
inclusions, yet you want to compile with high WARNS levels).

I am not sure if there is a standard for having a leading and/or trailing _
in the macro name, the usb code seems to use both.

There are still several unprotected headers here so it might be useful
to do the same thing on other files as well as the need arises.

MFC After: 3 days

17 years agoFix cut-and-paste bugs in the regression tests.
David Schultz [Wed, 3 Jan 2007 05:38:08 +0000 (05:38 +0000)]
Fix cut-and-paste bugs in the regression tests.

17 years agoThe distinction between quiet and signaling NaN formats is
David Schultz [Wed, 3 Jan 2007 05:00:03 +0000 (05:00 +0000)]
The distinction between quiet and signaling NaN formats is
machine-dependent; these files tell the latest version of gdtoa
what to do.

17 years agoMerge local diffs.
David Schultz [Wed, 3 Jan 2007 04:59:33 +0000 (04:59 +0000)]
Merge local diffs.

17 years agoThis commit was generated by cvs2svn to compensate for changes in r165743,
David Schultz [Wed, 3 Jan 2007 04:58:54 +0000 (04:58 +0000)]
This commit was generated by cvs2svn to compensate for changes in r165743,
which included commits to RCS files with non-trunk default branches.

17 years agoImport of the latest gdtoa sources, which include fixes for minor
David Schultz [Wed, 3 Jan 2007 04:58:54 +0000 (04:58 +0000)]
Import of the latest gdtoa sources, which include fixes for minor
problems relating to NaNs and rounding.

17 years agoFix rounding of 0xf for hex fp formats.
David Schultz [Wed, 3 Jan 2007 04:57:58 +0000 (04:57 +0000)]
Fix rounding of 0xf for hex fp formats.

PR: 90333

17 years ago - Fix schedgraph output with KSE threads. Call thread_switchout() after
Jeff Roberson [Wed, 3 Jan 2007 02:38:41 +0000 (02:38 +0000)]
 - Fix schedgraph output with KSE threads.  Call thread_switchout() after
   calling CTR() so we don't confuse a new kse thread with a real preemption.

17 years agofix typo: "adaptor" -> "adapter"
Ralf S. Engelschall [Tue, 2 Jan 2007 21:00:45 +0000 (21:00 +0000)]
fix typo: "adaptor" -> "adapter"

17 years agoThe nfsm_srvpathsiz() macro in nfsrv_symlink() in nfs_serv.c should
Hiroki Sato [Tue, 2 Jan 2007 20:42:08 +0000 (20:42 +0000)]
The nfsm_srvpathsiz() macro in nfsrv_symlink() in nfs_serv.c should
check length of the pathname in the range 0<=n<=NFS_MAXPATHLEN,
not 0<n<=NFS_MAXPATHLEN.  This fixes a minor interoperability problem
that the FreeBSD NFS server did not allow a symlink pointing the empty
pathname.

MFC after: 1 week

17 years agoRemove a bunch of dependencies in the IP header being the first thing in the
Julian Elischer [Tue, 2 Jan 2007 19:57:31 +0000 (19:57 +0000)]
Remove a bunch of dependencies in the IP header being the first thing in the
mbuf. First moves toward being able to cope better with having layer 2 (or
other encapsulation data) before the IP header in the packet being examined.
More commits to come to round out this functionality. This commit should
have no practical effect but clears the way for what is coming.
Revirewed by: luigi, yar
MFC After: 2 weeks

17 years agoUse the vnode interlock to close a race where pfs_vncache_alloc() could
John Baldwin [Tue, 2 Jan 2007 17:27:52 +0000 (17:27 +0000)]
Use the vnode interlock to close a race where pfs_vncache_alloc() could
attempt to vn_lock() a destroyed vnode resulting in a hang.

MFC after: 1 week
Submitted by: ups
Reviewed by: des

17 years agoClean up the struct archive used for reading an archive in the handling
Colin Percival [Tue, 2 Jan 2007 15:03:38 +0000 (15:03 +0000)]
Clean up the struct archive used for reading an archive in the handling
of @archive commands.  This bug should be harmless as long as you don't
use an excessive number of @archive commands.

MFC after: 1 week

17 years agoHandle errors which occur during archive_write_data and archive_write_close
Colin Percival [Tue, 2 Jan 2007 12:24:14 +0000 (12:24 +0000)]
Handle errors which occur during archive_write_data and archive_write_close
by printing an error message and exiting with a non-zero status code.

MFC after: 1 week

17 years agoJail_ip and jail_interface local variables were renamed to _ip and _interface
Florent Thoumie [Tue, 2 Jan 2007 11:07:13 +0000 (11:07 +0000)]
Jail_ip and jail_interface local variables were renamed to _ip and _interface
in a previous commit to avoid namespace collisions, unfortunately I missed two
of them. This leads to the ip alias being incorrectly removed in some cases
when using the stop command.

Reported by: Philipp Wuensche <cryx-freebsd@h3q.com>

17 years agoFix typo: effected -> affected
Ralf S. Engelschall [Tue, 2 Jan 2007 09:12:37 +0000 (09:12 +0000)]
Fix typo: effected -> affected

Submitted by: Gordon Stratton <tsr2600 (at) gmail (dot) com>

17 years agoExpand a comment for clarity and fix a missed substitution.
Scott Long [Tue, 2 Jan 2007 04:51:26 +0000 (04:51 +0000)]
Expand a comment for clarity and fix a missed substitution.

17 years agoFix compiling.
David Xu [Tue, 2 Jan 2007 04:14:01 +0000 (04:14 +0000)]
Fix compiling.

17 years agoMake sure that all of the fields in the header are clean. It was possible for
Scott Long [Tue, 2 Jan 2007 04:12:34 +0000 (04:12 +0000)]
Make sure that all of the fields in the header are clean.  It was possible for
unsafe flags to leak from one command to another.

17 years agoRemove old man page.
Tim Kientzle [Tue, 2 Jan 2007 03:42:16 +0000 (03:42 +0000)]
Remove old man page.

Thanks to: Pawel Worach

17 years agoAdd a simple tool for checking the command queues in the MFI driver. Also add
Scott Long [Mon, 1 Jan 2007 23:50:04 +0000 (23:50 +0000)]
Add a simple tool for checking the command queues in the MFI driver.  Also add
a simple wrapper for running the MegaCLI tool.

17 years agostyle(9) nit. Prefer struct[space]name[space]{ to make grep searches more
Christian S.J. Peron [Mon, 1 Jan 2007 21:49:39 +0000 (21:49 +0000)]
style(9) nit. Prefer struct[space]name[space]{ to make grep searches more
in line with that we find in the rest of the tree.

17 years agoForgot to mention that the last revision was fix for:
Daniel Gerzo [Mon, 1 Jan 2007 21:22:32 +0000 (21:22 +0000)]
Forgot to mention that the last revision was fix for:

PR: docs/106315
MFC-after: 3 days
Approved by: trhodes (mentor)

17 years ago- Do not claim that ipfw is the only firewall package available
Daniel Gerzo [Mon, 1 Jan 2007 21:16:42 +0000 (21:16 +0000)]
- Do not claim that ipfw is the only firewall package available
  in FreeBSD.

Approved by: trhodes (mentor), keramida (mentor)

17 years agoDuring an upgrade, if /boot/kernel.prev exists then the backup of
Ceri Davies [Mon, 1 Jan 2007 17:25:51 +0000 (17:25 +0000)]
During an upgrade, if /boot/kernel.prev exists then the backup of
/boot/kernel can fail.  Try to remove /boot/kernel.prev and cope the
best we can if that fails.

PR: bin/106376
Submitted by: Nobuyuki Koganemaru
Approved by: ru
MFC after: 2 weeks

17 years agoClean up pfr_kentry_pl2 as well. This fixes a kernel panic in the vm.zone
Max Laier [Mon, 1 Jan 2007 16:51:11 +0000 (16:51 +0000)]
Clean up pfr_kentry_pl2 as well.  This fixes a kernel panic in the vm.zone
sysctl after unloading pf.

Submitted by: Earl Lapus
MFC after: 3 days

17 years agoFix amd64 build.
Xin LI [Mon, 1 Jan 2007 14:47:45 +0000 (14:47 +0000)]
Fix amd64 build.

Submitted by: Divacky Roman <xdivac02 stud fit vutbr cz>

17 years agoTeach the stub policy about some of the more recent entry points that have
Christian S.J. Peron [Mon, 1 Jan 2007 01:47:18 +0000 (01:47 +0000)]
Teach the stub policy about some of the more recent entry points that have
been introduced to the MAC framework:

mpo_associate_nfsd_label
mpo_create_mbuf_from_firewall
mpo_check_system_nfsd
mpo_check_vnode_mmap_downgrade
mpo_check_vnode_mprotect
mpo_init_syncache_label
mpo_destroy_syncache_label
mpo_init_syncache_from_inpcb
mpo_create_mbuf_from_syncache

MFC after: 2 weeks [1]

[1] The syncache related entry points will NOT be MFCed as the changes in
    the syncache subsystem are not present in RELENG_6 yet.

17 years agoRemove conditional return of 1. For the MAC_STATIC case at this point in
Christian S.J. Peron [Mon, 1 Jan 2007 01:40:29 +0000 (01:40 +0000)]
Remove conditional return of 1. For the MAC_STATIC case at this point in
the code, one being returned is invariant.

Discussed with: rwatson
MFC after: 1 week

17 years agoMFp4: Fix bit name for SPI SR register
Warner Losh [Mon, 1 Jan 2007 00:50:25 +0000 (00:50 +0000)]
MFp4: Fix bit name for SPI SR register

17 years agoMFp4: Remove watchdog timeout that appears to be unused.
Warner Losh [Mon, 1 Jan 2007 00:48:25 +0000 (00:48 +0000)]
MFp4: Remove watchdog timeout that appears to be unused.

17 years agoMerge from FreeBSD-tsf-6 by way of p4:
Warner Losh [Mon, 1 Jan 2007 00:46:54 +0000 (00:46 +0000)]
Merge from FreeBSD-tsf-6 by way of p4:
correct values for PIO registers

submitted by: patrick schweiger

17 years agoFix typo in comment.
Warner Losh [Mon, 1 Jan 2007 00:35:34 +0000 (00:35 +0000)]
Fix typo in comment.

Submitted by: remko

17 years agoAdd comment about udp checksums being off in BSD 4.2 compatibility mode.
Warner Losh [Sun, 31 Dec 2006 21:34:53 +0000 (21:34 +0000)]
Add comment about udp checksums being off in BSD 4.2 compatibility mode.

Submitted by: Dr. Markus Waldeck
PR: kern/106657

17 years agoOnly signal the CV indicating that the MAC Framework is available for
Robert Watson [Sun, 31 Dec 2006 20:26:20 +0000 (20:26 +0000)]
Only signal the CV indicating that the MAC Framework is available for
exclusive access if there is at least one thread waiting for it to
become available.  This may significantly reduce overhead by reducing
the number of unnecessary wakeups issued whenever the framework becomes
idle.

Annotate that we still signal the CV more than necessary and should
fix this.

Obtained from: TrustedBSD Project
Reviewed by: csjp
Tested by: csjp

17 years agoremove delays that have been unnecessary since 2002... The iicbb driver
John-Mark Gurney [Sun, 31 Dec 2006 19:42:47 +0000 (19:42 +0000)]
remove delays that have been unnecessary since 2002...  The iicbb driver
has the proper delays...

17 years agoClarify a comment.
Ceri Davies [Sun, 31 Dec 2006 19:12:00 +0000 (19:12 +0000)]
Clarify a comment.

17 years agoFix typo.
Ceri Davies [Sun, 31 Dec 2006 18:53:45 +0000 (18:53 +0000)]
Fix typo.

17 years agoWelcome to 2007
Xin LI [Sun, 31 Dec 2006 16:35:29 +0000 (16:35 +0000)]
Welcome to 2007

17 years agoPrefer a more traditional spelling of inhibited in comments and panic
Robert Watson [Sun, 31 Dec 2006 15:56:04 +0000 (15:56 +0000)]
Prefer a more traditional spelling of inhibited in comments and panic
messages.

17 years agoFix wrong markup and some wordsmithing.
Hiroki Sato [Sun, 31 Dec 2006 15:36:44 +0000 (15:36 +0000)]
Fix wrong markup and some wordsmithing.

Submitted by: ru

17 years agoregen after addition of linux_utimes and linux_rt_sigtimedwait
Alexander Leidinger [Sun, 31 Dec 2006 13:20:31 +0000 (13:20 +0000)]
regen after addition of linux_utimes and linux_rt_sigtimedwait

17 years agoMFp4 (111746, 108671, 108945, 112352):
Alexander Leidinger [Sun, 31 Dec 2006 13:16:00 +0000 (13:16 +0000)]
MFp4 (111746, 108671, 108945, 112352):
 - add linux utimes syscall [1]
 - add linux rt_sigtimedwait syscall [2]

Submitted by: "Scot Hetzel" <swhetzel@gmail.com> [1]
Submitted by: Bruce Becker <hostmaster@whois.gts.net> [2]
PR: 93199 [2]

17 years agoMFp4:
Alexander Leidinger [Sun, 31 Dec 2006 12:42:55 +0000 (12:42 +0000)]
MFp4:
 - semi-automatic style fixes

17 years agoMFp4 (111746+):
Alexander Leidinger [Sun, 31 Dec 2006 12:39:10 +0000 (12:39 +0000)]
MFp4 (111746+):
  Redo the checking for 2.6 emulation. We now cache the value of
  use26 and replace calls to linux_get_osrelease() + parsing with
  a call to linux_use26(). Typical path is lockless now.

  Pointed out by: kib

This allows to ship RELENG_7_0 with a default osrelease of 2.4.2 and the
possibility to enable 2.6.x emulation without the possible performance
impact of the previous version of the check.

Submitted by: rdivacky

17 years agoMFp4:
Alexander Leidinger [Sun, 31 Dec 2006 11:56:16 +0000 (11:56 +0000)]
MFp4:
 - semi-automatic style fixes
 - spelling fixes in comments
 - add some comments

17 years agoo Fix a comment: refer to the write community name by the OID not by
Maxim Konovalov [Sun, 31 Dec 2006 11:19:48 +0000 (11:19 +0000)]
o Fix a comment: refer to the write community name by the OID not by
the line number which is incorrect now and probably will shift in future.

17 years agoAllow for /usr/bin/env when parsing the shebang line from an
Yaroslav Tykhiy [Sun, 31 Dec 2006 11:07:29 +0000 (11:07 +0000)]
Allow for /usr/bin/env when parsing the shebang line from an
interpreted $command.  Some "portable" sofware packages use such a
line to skip the task of figuring out the absolute pathname of the
interpreter at install time, e.g.:

#!/usr/bin/env python

It is insecure, but a popular book on Python seems to have advised
it to a wide audience.  Hence a number of such scripts in the ports,
mostly written in Python.

PR: bin/100287
MFC after: 1 week

17 years agoUse $required_modules wherever suitable. Use load_kld() in special
Yaroslav Tykhiy [Sun, 31 Dec 2006 10:37:18 +0000 (10:37 +0000)]
Use $required_modules wherever suitable.  Use load_kld() in special
cases.  So we get rid of quite a few lines of duplicated code.

17 years agoFix a typo in a warning message.
Yaroslav Tykhiy [Sun, 31 Dec 2006 10:12:53 +0000 (10:12 +0000)]
Fix a typo in a warning message.

17 years agoUdate info after official patch
Andrey A. Chernov [Sun, 31 Dec 2006 09:50:44 +0000 (09:50 +0000)]
Udate info after official patch

17 years agoMerge conflict after official patch
Andrey A. Chernov [Sun, 31 Dec 2006 09:47:34 +0000 (09:47 +0000)]
Merge conflict after official patch

17 years agoChange RL_LIBRARY_VERSION
Andrey A. Chernov [Sun, 31 Dec 2006 09:27:16 +0000 (09:27 +0000)]
Change RL_LIBRARY_VERSION

17 years agoUpdate config
Andrey A. Chernov [Sun, 31 Dec 2006 09:26:00 +0000 (09:26 +0000)]
Update config

17 years agoMerge conflicts
Andrey A. Chernov [Sun, 31 Dec 2006 09:22:31 +0000 (09:22 +0000)]
Merge conflicts

17 years agoUpdate instructions
Andrey A. Chernov [Sun, 31 Dec 2006 09:12:19 +0000 (09:12 +0000)]
Update instructions

17 years agoThis commit was generated by cvs2svn to compensate for changes in r165670,
Andrey A. Chernov [Sun, 31 Dec 2006 09:06:30 +0000 (09:06 +0000)]
This commit was generated by cvs2svn to compensate for changes in r165670,
which included commits to RCS files with non-trunk default branches.

17 years agoVirgin import of GNU Readline 5.2
Andrey A. Chernov [Sun, 31 Dec 2006 09:06:30 +0000 (09:06 +0000)]
Virgin import of GNU Readline 5.2

17 years agoo Grammar: is appears -> appears.
Maxim Konovalov [Sun, 31 Dec 2006 07:22:55 +0000 (07:22 +0000)]
o Grammar: is appears -> appears.

PR: docs/107306
Submitted by: Tomas Mozes
MFC after: 1 week

17 years ago"block is set to UTC" -> "clock is set to UTC"
Colin Percival [Sun, 31 Dec 2006 05:18:35 +0000 (05:18 +0000)]
"block is set to UTC" -> "clock is set to UTC"

Submitted by: csjp

17 years agoInvert the logic inside of two KASSERTS which resulted in two kernel panics
Christian S.J. Peron [Sun, 31 Dec 2006 02:50:07 +0000 (02:50 +0000)]
Invert the logic inside of two KASSERTS which resulted in two kernel panics
for circumstances which are quite normal.

Discussed with: kmacy

17 years agoEliminate global symbols starting with an underscore from rc.d
Yaroslav Tykhiy [Sat, 30 Dec 2006 22:53:20 +0000 (22:53 +0000)]
Eliminate global symbols starting with an underscore from rc.d
scripts, except for mdconfig* and jail.  Such symbols are reserved
for the rc.subr internals.  Most scripts can be fixed by just
declaring _foo symbols as local: few scripts actually need them to
be global.

Discussed with: dougb in freebsd-rc

17 years agoAdd the fortunes from the PR that were not already present.
Doug Barton [Sat, 30 Dec 2006 21:54:41 +0000 (21:54 +0000)]
Add the fortunes from the PR that were not already present.
Delete some duplicates found while double checking the new ones,
and fix a typo.

These haven't been sorted yet, but will be in a future commit.

PR: ports/40273
Submitted by: Achim Patzner <ap@proxon.bnc.net>

17 years ago- Don't defer the removal of an 802.1q header for no real reason.
Yaroslav Tykhiy [Sat, 30 Dec 2006 21:10:25 +0000 (21:10 +0000)]
- Don't defer the removal of an 802.1q header for no real reason.
- Micro-optimize the addition of an 802.1q header to match the removal code.
- Consistently check for interfaces being up and running.
- Consistently use NULL instead of 0 with pointers.

17 years agonow that MFS_IMAGE is in the FULLKERNEL section, use FULLKERNEL instead of
John-Mark Gurney [Sat, 30 Dec 2006 20:27:24 +0000 (20:27 +0000)]
now that MFS_IMAGE is in the FULLKERNEL section, use FULLKERNEL instead of
KERNEL_KO, this fixes MFS_IMAGE on a debugging kernel...

Submitted by: Neelkanth Natu

17 years agoRemove alpha-specific stuff from the hardware notes (these are mostly
Bruce A. Mah [Sat, 30 Dec 2006 18:22:09 +0000 (18:22 +0000)]
Remove alpha-specific stuff from the hardware notes (these are mostly
in attributes for generating MD hardware, as well as a few alpha-only
paragraphs here and there).

While here, bump copyright date in dev.archlist.txt.

17 years agoWhitespace fix and remove an extra cast.
John Baldwin [Sat, 30 Dec 2006 17:53:28 +0000 (17:53 +0000)]
Whitespace fix and remove an extra cast.

17 years agoBe consistent with the spelling of "dependent" in user-visible places.
Ceri Davies [Sat, 30 Dec 2006 11:55:47 +0000 (11:55 +0000)]
Be consistent with the spelling of "dependent" in user-visible places.

PR: kern/27429
Submitted by: T. William Wells

17 years agoMFC:
Daniel Gerzo [Fri, 29 Dec 2006 22:45:12 +0000 (22:45 +0000)]
MFC:
- add some files to FILES section and provide their description
- reference devfs.conf and devfs.rules in SEE ALSO section

Approved by: keramida (mentor), trhodes (mentor)

17 years agoSome how the d slipped out in the last commit, add it back...
John-Mark Gurney [Fri, 29 Dec 2006 22:30:10 +0000 (22:30 +0000)]
Some how the d slipped out in the last commit, add it back...

Pointed out by: mdoc-police (ru)

17 years agoSummer of Code 2005: improve libalias - part 2 of 2
Paolo Pisati [Fri, 29 Dec 2006 21:59:17 +0000 (21:59 +0000)]
Summer of Code 2005: improve libalias - part 2 of 2

With the second (and last) part of my previous Summer of Code work, we get:

-ipfw's in kernel nat

-redirect_* and LSNAT support

General information about nat syntax and some examples are available
in the ipfw (8) man page. The redirect and LSNAT syntax are identical
to natd, so please refer to natd (8) man page.

To enable in kernel nat in rc.conf, two options were added:

o firewall_nat_enable: equivalent to natd_enable

o firewall_nat_interface: equivalent to natd_interface

Remember to set net.inet.ip.fw.one_pass to 0, if you want the packet
to continue being checked by the firewall ruleset after being
(de)aliased.

NOTA BENE: due to some problems with libalias architecture, in kernel
nat won't work with TSO enabled nic, thus you have to disable TSO via
ifconfig (ifconfig foo0 -tso).

Approved by: glebius (mentor)

17 years agoa) macro-ization of all mbuf and random number
Randall Stewart [Fri, 29 Dec 2006 20:21:42 +0000 (20:21 +0000)]
a) macro-ization of all mbuf and random number
   access plus timers. This makes the code
   more portable and able to change out the
   mbuf or timer system used more easily ;-)
b) removal of all use of pkt-hdr's until only
   the places we need them (before ip_output routines).
c) remove a bunch of code not needed due to <b> aka
   worrying about pkthdr's :-)
d) There was one last reorder problem it looks where
   if a restart occur's and we release and relock (at
   the point where we setup our alias vtag) we would
   end up possibly getting the wrong TSN in place. The
   code that fixed the TSN's just needed to be shifted
   around BEFORE the release of the lock.. also code that
   set the state (since this also could contribute).
Approved by: gnn

17 years agoSlightly resort functions in file so that no forward function prototypes
Robert Watson [Fri, 29 Dec 2006 20:21:21 +0000 (20:21 +0000)]
Slightly resort functions in file so that no forward function prototypes
are required.

Obtained from: TrustedBSD Project

17 years agoRe-add include of opt_mac.h in mac_framework.c, which was improperly
Robert Watson [Fri, 29 Dec 2006 20:16:29 +0000 (20:16 +0000)]
Re-add include of opt_mac.h in mac_framework.c, which was improperly
removed from this file.  It is required to pick up the definition of
MAC_STATIC.

17 years agoRe-connect snmp_bridge to the build. It survived a make universe
Bjoern A. Zeeb [Fri, 29 Dec 2006 19:24:51 +0000 (19:24 +0000)]
Re-connect snmp_bridge to the build. It survived a make universe
(apart from powerpc where the build broke earlier).

17 years agobridge_snmp.h
Bjoern A. Zeeb [Fri, 29 Dec 2006 19:23:38 +0000 (19:23 +0000)]
bridge_snmp.h
* Change the API of bridge_get_basemac to take a maximum buffer length.

bridge_if.c
* Adopt to new API.
* In bridge_attach_newif() remove an additional pointer to the buffer
  by shuffling the code a bit. Also makes the code more readable.

bridge_sys.c
* bridge_get_basemac():
  - Adopt to the new API.
  - Change check for error code of getifaddrs().
  - First check for sa_family != AF_LINK.
  - Copy sockaddr_dl * to get around alignment constraints on some
    platforms.
  - Use strcmp instead of strncmp so that "foo11" != "foo1".
* other functions:
  - Allocate n times of the struct we need instead of arbitrary len,
    cast to the type we want it to be and pass around struct *s instead
    of char *s. This gets us around alignment restrictions on some
    platforms and in addition it is more clear what data is passed around.
  - Name variables for same types consistently.

Reviewed by:    syrinx

17 years agoFix an off-by-one which could mean writing beyond the end of the array
Bjoern A. Zeeb [Fri, 29 Dec 2006 18:44:04 +0000 (18:44 +0000)]
Fix an off-by-one which could mean writing beyond the end of the array
when copying the interface name.
This code part should probably be rewritten.

17 years agoFix fat-fingering in previous commit.
John Baldwin [Fri, 29 Dec 2006 16:38:22 +0000 (16:38 +0000)]
Fix fat-fingering in previous commit.

Pointy hat to: jhb

17 years ago- Fix markup.
Ruslan Ermilov [Fri, 29 Dec 2006 16:28:38 +0000 (16:28 +0000)]
- Fix markup.

- Somewhat improve wording.

- Change the layout of the EXAMPLES section so that descriptions
  come before example, as in most other manpages.

- Fix a bad example that edits a label using a `c' partition.

17 years agoAdd "nojail" keyword to prevent spurious error messages.
Hiroki Sato [Fri, 29 Dec 2006 15:56:21 +0000 (15:56 +0000)]
Add "nojail" keyword to prevent spurious error messages.

Spotted by: Keve Nagy
PR: conf/107083

17 years agoFix oops in previous commit.
Bruce Evans [Fri, 29 Dec 2006 15:48:18 +0000 (15:48 +0000)]
Fix oops in previous commit.

17 years agoFixed some style bugs (mainly assorted errors in comments, and inconsistent
Bruce Evans [Fri, 29 Dec 2006 15:29:49 +0000 (15:29 +0000)]
Fixed some style bugs (mainly assorted errors in comments, and inconsistent
spelling of `result').

17 years agoSome whitespace nits and remove a few casts.
John Baldwin [Fri, 29 Dec 2006 14:58:18 +0000 (14:58 +0000)]
Some whitespace nits and remove a few casts.

17 years agoFixed some style bugs (whitespace only).
Bruce Evans [Fri, 29 Dec 2006 14:28:23 +0000 (14:28 +0000)]
Fixed some style bugs (whitespace only).

17 years agoVarious bpf(4) related fixes to catch places up to the new bpf(4)
John Baldwin [Fri, 29 Dec 2006 13:59:50 +0000 (13:59 +0000)]
Various bpf(4) related fixes to catch places up to the new bpf(4)
semantics.
- Stop testing bpf pointers for NULL.  In some cases use
  bpf_peers_present() and then call the function directly inside the
  conditional block instead of the macro.
- For places where the entire conditional block is the macro, remove the
  test and make the macro unconditional.
- Use BPF_MTAP() in if_pfsync on FreeBSD instead of an expanded version of
  the old semantics.

Reviewed by: csjp (older version)

17 years agoWork around a long standing LOR with user/group rules by doing the socket
Max Laier [Fri, 29 Dec 2006 13:59:03 +0000 (13:59 +0000)]
Work around a long standing LOR with user/group rules by doing the socket
lookup early.  This has some performance implications and should not be
enabled by default, but might help greatly in certain setups.  After some
more testing this could be turned into a sysctl.

Tested by: avatar
LOR ids: 17, 24, 32, 46, 191 (conceptual)
MFC after: 6 weeks

17 years agoTry harder to garbage-collect the "LOCORE" (really asm) version of
Bruce Evans [Fri, 29 Dec 2006 13:36:26 +0000 (13:36 +0000)]
Try harder to garbage-collect the "LOCORE" (really asm) version of
MPLOCKED.  The cleaning in rev.1.25 was supposed to have been undone
by rev.1.26, but 1.26 could never have actually affected asm files
since atomic.h is full of C declarations so including it in asm files
would just give syntax errors.  The asm MPLOCKED is even less needed
than when misplaced definitions of it were first removed, and is now
unused in any asm file in the src tree except in anachronismns in
sys/i386/i386/support.s.