]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
19 years agoBecause code paths for I/O requests are quite complex, add comments above
pjd [Sat, 13 Aug 2005 17:45:37 +0000 (17:45 +0000)]
Because code paths for I/O requests are quite complex, add comments above
the functions which participate in I/O paths.

MFC after: 1 day

19 years agoadd list mac and mac:kick support
sam [Sat, 13 Aug 2005 17:38:09 +0000 (17:38 +0000)]
add list mac and mac:kick support

Submitted by: Michal Mertl (original version)
MFC after: 2 weeks

19 years agoExtend acl support to pass ioctl requests through and use this to
sam [Sat, 13 Aug 2005 17:31:48 +0000 (17:31 +0000)]
Extend acl support to pass ioctl requests through and use this to
add support for getting the current policy setting and collecting
the list of mac addresses in the acl table.

Submitted by: Michal Mertl (original version)
MFC after: 2 weeks

19 years agoAllow multiple commands to be specified on a single command line, e.g.,
cperciva [Sat, 13 Aug 2005 16:27:13 +0000 (16:27 +0000)]
Allow multiple commands to be specified on a single command line, e.g.,
"portsnap fetch update" or "portsnap -I cron update".  They will be
executed in the order that they appear, and duplicates are not removed
(so "portsnap fetch fetch fetch fetch" is meaningful, albeit rather
silly).

Requested by: Roman Divacky

19 years agoPut the comparison with PEOF into a new macro is_eof(). Don't use it if the
stefanf [Sat, 13 Aug 2005 15:47:13 +0000 (15:47 +0000)]
Put the comparison with PEOF into a new macro is_eof().  Don't use it if the
character comes from a string.

19 years agoDeclare echocmd() and testcmd().
stefanf [Sat, 13 Aug 2005 15:04:30 +0000 (15:04 +0000)]
Declare echocmd() and testcmd().

19 years agoUse assignment rather than comparison.
stefanf [Sat, 13 Aug 2005 15:00:54 +0000 (15:00 +0000)]
Use assignment rather than comparison.

19 years agoFix the progress statistics code by printing the right variable. I
cperciva [Sat, 13 Aug 2005 11:55:29 +0000 (11:55 +0000)]
Fix the progress statistics code by printing the right variable.  I
missed this when I changed the stats code in the ports -> base
transition.

Pointed out by: simon

19 years ago* Solve "No PHY found" problem for more Yukon Lite variants.
bz [Sat, 13 Aug 2005 11:13:12 +0000 (11:13 +0000)]
* Solve "No PHY found" problem for more Yukon Lite variants.
* Catch a bus attach error.
* Improve locking.

MFC after: 6 days

19 years ago* Add dynamic sysctl for net.inet6.ip6.fw.
bz [Sat, 13 Aug 2005 11:02:34 +0000 (11:02 +0000)]
* Add dynamic sysctl for net.inet6.ip6.fw.
* Correct handling of IPv6 Extension Headers.
* Add unreach6 code.
* Add logging for IPv6.

Submitted by: sysctl handling derived from patch from ume needed for ip6fw
Obtained from: is_icmp6_query and send_reject6 derived from similar
functions of netinet6,ip6fw
Reviewed by: ume, gnn; silence on ipfw@
Test setup provided by: CK Software GmbH
MFC after: 6 days

19 years ago- Provide all initialisers for a struct event.
stefanf [Sat, 13 Aug 2005 08:38:02 +0000 (08:38 +0000)]
- Provide all initialisers for a struct event.
- Include init.h in the generated init.c.
- Generate prototypes and put the brace on the next line.

19 years agoFirst declare the functions to pacify -Wmissing-prototypes.
stefanf [Sat, 13 Aug 2005 08:31:37 +0000 (08:31 +0000)]
First declare the functions to pacify -Wmissing-prototypes.

19 years agoInclude missing headers.
stefanf [Sat, 13 Aug 2005 08:26:58 +0000 (08:26 +0000)]
Include missing headers.

19 years agoUse prototypes in the MKINIT lines collected by mkinit.
stefanf [Sat, 13 Aug 2005 08:12:18 +0000 (08:12 +0000)]
Use prototypes in the MKINIT lines collected by mkinit.

19 years agoUse prototypes in the generated builtins.{c,h} files.
stefanf [Sat, 13 Aug 2005 08:09:22 +0000 (08:09 +0000)]
Use prototypes in the generated builtins.{c,h} files.

19 years agoPrune empty directories in /usr
phk [Sat, 13 Aug 2005 08:06:18 +0000 (08:06 +0000)]
Prune empty directories in /usr

Move argv parsing.

19 years agoForced commit due to wrong commit-log file:
stefanf [Sat, 13 Aug 2005 07:59:46 +0000 (07:59 +0000)]
Forced commit due to wrong commit-log file:
- Provide missing prototypes and variable declarations in arith.h, include
  arith.h.
- Make arith_assign() and yyerror() static.

19 years agocvs -d stefanf@ncvs.FreeBSD.org:/home/ncvs
stefanf [Sat, 13 Aug 2005 07:54:23 +0000 (07:54 +0000)]
cvs -d stefanf@ncvs.FreeBSD.org:/home/ncvs

19 years agoAdd an implementation of the semi-standard wcsdup() function, as found
tjr [Sat, 13 Aug 2005 05:54:33 +0000 (05:54 +0000)]
Add an implementation of the semi-standard wcsdup() function, as found
on Microsoft and GNU systems.

19 years agosave+restore wpa, privacy, and roaming settings
sam [Sat, 13 Aug 2005 04:23:33 +0000 (04:23 +0000)]
save+restore wpa, privacy, and roaming settings

MFC after: 2 weeks

19 years agoChange OUT from -2 to CHAR_MIN-1, making it impossible for it to
tjr [Sat, 13 Aug 2005 02:30:15 +0000 (02:30 +0000)]
Change OUT from -2 to CHAR_MIN-1, making it impossible for it to
inadvertently match a negative char in the RE being compiled.

This fixes compilation of "\376" (as an ERE) and "\376\376" (as a BRE).

PR: 84740
MFC after: 1 week

19 years agofix handling of the current channel (missed in previous commit)
sam [Sat, 13 Aug 2005 00:30:26 +0000 (00:30 +0000)]
fix handling of the current channel (missed in previous commit)

19 years agofix media_status so devd recognizes the device as an 802.11 interface
sam [Sat, 13 Aug 2005 00:04:42 +0000 (00:04 +0000)]
fix media_status so devd recognizes the device as an 802.11 interface

Submitted by: Robert C. Noland III

19 years agocorrect CTS duration calculation; SIFS+ACK should use the xmit rate not
sam [Fri, 12 Aug 2005 23:11:44 +0000 (23:11 +0000)]
correct CTS duration calculation; SIFS+ACK should use the xmit rate not
the rate for CTS

MFC after: 3 days
Obtained from: Atheros

19 years agoPromise PDC4071[89] support.
brueffer [Fri, 12 Aug 2005 23:05:48 +0000 (23:05 +0000)]
Promise PDC4071[89] support.

19 years agoAdd missing links from getgrent_r.3, getgrnam_r.3, and getgrgid_r.3 to
simon [Fri, 12 Aug 2005 22:46:15 +0000 (22:46 +0000)]
Add missing links from getgrent_r.3, getgrnam_r.3, and getgrgid_r.3 to
getgrent.3.

Submitted by: Ulf Lilleengen <lulf@kerneled.org>
MFC after: 3 days

19 years agoCheck for marker pages when scanning active and inactive page queues.
tegge [Fri, 12 Aug 2005 18:17:40 +0000 (18:17 +0000)]
Check for marker pages when scanning active and inactive page queues.

Reviewed by: alc

19 years agoAdd support for the Promise PDC4071[89] chips used on fx the Fasttrak TX4300.
sos [Fri, 12 Aug 2005 16:54:11 +0000 (16:54 +0000)]
Add support for the Promise PDC4071[89] chips used on fx the Fasttrak TX4300.
Docs kindly provided by Promise.

19 years agoAdd definition for SATAII 3Gb/s mode.
sos [Fri, 12 Aug 2005 16:51:39 +0000 (16:51 +0000)]
Add definition for SATAII 3Gb/s mode.

19 years ago- fix typo in comment.
ume [Fri, 12 Aug 2005 15:27:25 +0000 (15:27 +0000)]
- fix typo in comment.
- nuke unused code.

Submitted by: suz
Obtained from: KAME

19 years agoChange directory one level at a time, and use CDUP to back out. This is a
des [Fri, 12 Aug 2005 12:48:50 +0000 (12:48 +0000)]
Change directory one level at a time, and use CDUP to back out.  This is a
work in progress; it partially fixed bin/83278 and is a prerequisite to
fixing bin/83277.

PR: bin/83277, bin/83278

19 years agoIntroduce the vm.boot_pages tunable and sysctl, which controls the number
des [Fri, 12 Aug 2005 12:24:19 +0000 (12:24 +0000)]
Introduce the vm.boot_pages tunable and sysctl, which controls the number
of pages reserved to bootstrap the kernel memory allocator.

MFC after: 2 weeks

19 years agoEliminate an unnecessary bcopy().
des [Fri, 12 Aug 2005 12:22:05 +0000 (12:22 +0000)]
Eliminate an unnecessary bcopy().

19 years agoAxe ppp_for_tty(). Use tty->t_lsc pointer to store sc. This
glebius [Fri, 12 Aug 2005 08:27:15 +0000 (08:27 +0000)]
Axe ppp_for_tty(). Use tty->t_lsc pointer to store sc. This
also eliminates recursive use of ppp_softc_list_mtx.

PR: kern/84686
Reviewed by: phk
MFC after: 1 week

19 years agoGELI doesn't need cryptodev.
pjd [Fri, 12 Aug 2005 07:44:42 +0000 (07:44 +0000)]
GELI doesn't need cryptodev.

MFC after: 2 days

19 years agoWhen using cmp(1) to determine if the ports tree is already up to date,
cperciva [Fri, 12 Aug 2005 04:35:27 +0000 (04:35 +0000)]
When using cmp(1) to determine if the ports tree is already up to date,
we want the exit code, but not the actual output; add a redirect to
/dev/null.

Pointy hat to: cperciva

19 years agoAdd NATM_LOCK() and NATM_UNLOCK() in places where npcb_add() and
rodrigc [Fri, 12 Aug 2005 02:38:20 +0000 (02:38 +0000)]
Add NATM_LOCK() and NATM_UNLOCK() in places where npcb_add() and
npcb_free() are called, in order to eliminate witness panics.
This was overlooked in removal of GIANT from ATM.

Reviewed by: rwatson

19 years agoProvide more complete "How to add a new file system to glabel." list.
pjd [Fri, 12 Aug 2005 00:34:45 +0000 (00:34 +0000)]
Provide more complete "How to add a new file system to glabel." list.

MFC after: 1 week

19 years agoAdd code for Ext2FS and ReiserFS labels recognition.
pjd [Fri, 12 Aug 2005 00:27:45 +0000 (00:27 +0000)]
Add code for Ext2FS and ReiserFS labels recognition.

Submitted by: Stanislav Sedov <stas@310.ru>
PR: kern/84638
MFC after: 1 week

19 years agoAvoid creating directories in devfs by changing all '/' in labels to '_'.
pjd [Fri, 12 Aug 2005 00:05:09 +0000 (00:05 +0000)]
Avoid creating directories in devfs by changing all '/' in labels to '_'.

Idea from: Stanislav Sedov <stas@310.ru>
MFC after: 3 days

19 years agoEliminate unneeded diagnostic code.
alc [Thu, 11 Aug 2005 23:38:02 +0000 (23:38 +0000)]
Eliminate unneeded diagnostic code.

Eliminate an unused #include.  (Kernel stack allocation and deallocation
long ago migrated to the machine-independent code.)

19 years agoEl_parse's third parameter is now const-qualified, add a cast.
stefanf [Thu, 11 Aug 2005 20:28:26 +0000 (20:28 +0000)]
El_parse's third parameter is now const-qualified, add a cast.

19 years agoForced commit to note rev 1.15 was
njl [Thu, 11 Aug 2005 18:42:45 +0000 (18:42 +0000)]
Forced commit to note rev 1.15 was

Submitted by: Philip S. Schulz <ph.schulz / gmx.de>

19 years agoThe "lowest" sysctl setting makes more sense as the lowest one to use, so
njl [Thu, 11 Aug 2005 18:40:58 +0000 (18:40 +0000)]
The "lowest" sysctl setting makes more sense as the lowest one to use, so
discard all levels less than this setting, not less than/equal to.

MFC after: 1 day

19 years agoEliminate unneeded diagnostic code.
alc [Thu, 11 Aug 2005 17:43:28 +0000 (17:43 +0000)]
Eliminate unneeded diagnostic code.

Reviewed by: tegge

19 years agoModify the fix in rev 1.16 slightly. Instead of setting the cap to zero,
njl [Thu, 11 Aug 2005 17:12:57 +0000 (17:12 +0000)]
Modify the fix in rev 1.16 slightly.  Instead of setting the cap to zero,
just skip summing it if it's unknown.

MFC after: 1 day

19 years ago- Clarify how events are used and how the different event sysctl correlate
markus [Thu, 11 Aug 2005 15:51:11 +0000 (15:51 +0000)]
- Clarify how events are used and how the different event sysctl correlate
- Document how an event looks like that is received by devd(8) from acpi_ibm(4)
- Document the event codes as they occur on a T41p
- Add a note about sysctl.conf and an Xref to it
- Add an example section providing an example script and an entry for
  devd.conf. This is an adapted version of the one in acpi_fujitsu(4).
- s/Fn-F4/Fn+F4/ for consistency sake
- Add a button description where refering to a T41p, remove it otherwise,
  since it might vary from model to model
- Bump .Dd

Thanks for improvements to: brueffer, ru

Approved by: brueffer

19 years agoIf we fail in chown(2), try to just change the group and continue on to
brian [Thu, 11 Aug 2005 15:43:55 +0000 (15:43 +0000)]
If we fail in chown(2), try to just change the group and continue on to
change the permissions.  Failures are still recorded.

This allows mtree to do a generally better job of things when uid != 0.

Sponsored by: ActiveState/Sophos
Partially submitted by: neilw at ActiveState dot com
Reviewed by: neilw at ActiveState dot com
MFC after: 3 weeks

19 years agoEmbellish comment.
obrien [Thu, 11 Aug 2005 15:35:59 +0000 (15:35 +0000)]
Embellish comment.

Submitted by: Rostislav Krasny <rosti.bsd@gmail.com>

19 years agoPut rev. 1.65 into usage().
obrien [Thu, 11 Aug 2005 15:23:23 +0000 (15:23 +0000)]
Put rev. 1.65 into usage().

19 years agoGELI doesn't need cryptodev.
pjd [Thu, 11 Aug 2005 14:52:27 +0000 (14:52 +0000)]
GELI doesn't need cryptodev.

MFC after: 3 days

19 years agoFix stand-alone (or MODULES_WITH_WORLD=1) build of
yar [Thu, 11 Aug 2005 14:45:10 +0000 (14:45 +0000)]
Fix stand-alone (or MODULES_WITH_WORLD=1) build of
the procfs module by creating opt_compat.h with
appropriate compatibility options: COMPAT_43 on all
arch's and COMPAT_IA32 in addition on amd64.

Pointy hat to: peter
MFC after: 3 days

19 years agoIf the ports tree is already up to date when we run `portsnap update`,
cperciva [Thu, 11 Aug 2005 13:48:13 +0000 (13:48 +0000)]
If the ports tree is already up to date when we run `portsnap update`,
exit early; this avoids spending five seconds rebuilding the INDEX
files.

Requested by: somebody on IRC, but I can't remember who

19 years agoCorrect some grammar.
ceri [Thu, 11 Aug 2005 10:32:05 +0000 (10:32 +0000)]
Correct some grammar.

19 years agoo To prevent a race between RTM_DELETE message and
glebius [Thu, 11 Aug 2005 08:26:31 +0000 (08:26 +0000)]
o To prevent a race between RTM_DELETE message and
  arptimer() deleting stale entry, we need to lock
  rtentry before unlocking radix head.

Reviewed by: sam

19 years agoo Fix a race between three threads: output path,
glebius [Thu, 11 Aug 2005 08:25:48 +0000 (08:25 +0000)]
o Fix a race between three threads: output path,
  incoming ARP packet and route request adding/removing
  ARP entries. The root of the problem is that
  struct llinfo_arp was accessed without any locks.
  To close race we will use locking provided by
  rtentry, that references this llinfo_arp:
  - Make arplookup() return a locked rtentry.
  - In arpresolve() hold the lock provided by
    rt_check()/arplookup() until the end of function,
    covering all accesses to the rtentry itself and
    llinfo_arp it refers to.
  - In in_arpinput() do not drop lock provided by
    arplookup() during first part of the function.
  - Simplify logic in the first part of in_arpinput(),
    removing one level of indentation.
  - In the second part of in_arpinput() hold rtentry
    lock while copying address.

o Fix a condition when route entry is destroyed, while
  another thread is contested on its lock:
  - When storing a pointer to rtentry in llinfo_arp list,
    always add a reference to this rtentry, to prevent
    rtentry being destroyed via RTM_DELETE request.
  - Remove this reference when removing entry from
    llinfo_arp list.

o Further cleanup of arptimer():
  - Inline arptfree() into arptimer().
  - Use official queue(3) way to pass LIST.
  - Hold rtentry lock while reading its structure.
  - Do not check that sdl_family is AF_LINK, but
    assert this.

Reviewed by: sam
Stress test: http://www.holm.cc/stress/log/cons141.html
Stress test: http://people.freebsd.org/~pho/stress/log/cons144.html

19 years agoo Make rt_check() function more strict:
glebius [Thu, 11 Aug 2005 08:14:53 +0000 (08:14 +0000)]
o Make rt_check() function more strict:
  - rt0 passed to rt_check() must not be NULL, assert this.
  - rt returned by rt_check() must be valid locked rtentry,
    if no error occured.
o Modify callers, so that they never pass NULL rt0
  to rt_check().

Reviewed by: sam, ume (nd6.c)

19 years agocreate sysctl tree dynamically. it is required to share
ume [Thu, 11 Aug 2005 07:28:01 +0000 (07:28 +0000)]
create sysctl tree dynamically.  it is required to share
net.inet6.ip6.fw with upcomming ipfw2 improvement for IPv6.

Requested by: bz

19 years agoDecouple the unrefing of a page table page from the removal of a pv entry.
alc [Thu, 11 Aug 2005 02:22:55 +0000 (02:22 +0000)]
Decouple the unrefing of a page table page from the removal of a pv entry.
In other words, change pmap_remove_entry() such that it no longer unrefs
the page table page.  Now, it only removes the pv entry.

Reviewed by: tegge

19 years agoDocument methods people use to load linux shared libraries in FreeBSD
imp [Wed, 10 Aug 2005 21:44:26 +0000 (21:44 +0000)]
Document methods people use to load linux shared libraries in FreeBSD
binaries.  A very neat trick.

19 years agoUse device_printf() and if_printf() and remove dc_unit from softc.
jhb [Wed, 10 Aug 2005 20:33:46 +0000 (20:33 +0000)]
Use device_printf() and if_printf() and remove dc_unit from softc.

19 years agoUse if_printf() and device_printf() and axe sf_unit from the softc as a
jhb [Wed, 10 Aug 2005 20:24:39 +0000 (20:24 +0000)]
Use if_printf() and device_printf() and axe sf_unit from the softc as a
result.

19 years ago- Remove pre-new-bus code under #if 0 and some other rotted code under
jhb [Wed, 10 Aug 2005 20:21:29 +0000 (20:21 +0000)]
- Remove pre-new-bus code under #if 0 and some other rotted code under
  #if 0.
- Use pci_enable_busmaster() to enable busmastering instead of frobbing
  the command register directly.
- Don't check to see if memory or I/O can be enabled by writing to the
  command register.  The PCI bus driver's bus_alloc_resource() method
  already checks this and will fail if it can't enable the bit.

19 years ago- Use if_printf() and device_printf() instead of outputting my%d:
jhb [Wed, 10 Aug 2005 20:15:56 +0000 (20:15 +0000)]
- Use if_printf() and device_printf() instead of outputting my%d:
  everywhere.  This means that my_unit is no longer used as well.  The
  watchdog routine now also prints 'my0: ...' rather than 'm0x0: ...'.
- Don't bzero the softc and don't try to free it in detach or if attach
  fails.
- A whitespace fix.

19 years ago- Use callout_init_mtx() to close races between hme_stop() and hme_tick().
jhb [Wed, 10 Aug 2005 20:12:25 +0000 (20:12 +0000)]
- Use callout_init_mtx() to close races between hme_stop() and hme_tick().
- Use the driver lock instead of Giant in a bus dma callback.
- Clear IFF_DRV_(RUNNING|OACTIVE) in hme_stop() instead of just clearing
  RUNNING in hme_ioctl() to be more like other ethernet drivers.
- Lock the driver lock around mii operations.
- Remove spls.
- Cleanup locking in hme_ioctl().

MFC after: 1 week

19 years agoDon't use ic_curmode to decide when to do 11g station accounting,
sam [Wed, 10 Aug 2005 17:42:13 +0000 (17:42 +0000)]
Don't use ic_curmode to decide when to do 11g station accounting,
use the station channel properties.  Fixes assert failure/bogus
operation when an ap is operating in 11a and has associated stations
then switches to 11g.

Noticed by: Michal Mertl
Reviewed by: avatar
MFC after: 2 weeks

19 years agoremoved RFC1885-related code. it was obsoleted by RFC2463, and the
ume [Wed, 10 Aug 2005 17:30:10 +0000 (17:30 +0000)]
removed RFC1885-related code.  it was obsoleted by RFC2463, and the
code was #ifdef'ed out for a long time.

Submitted by: suz
Obtained from: KAME

19 years agoBe a nice community member and honor the request for backout after
schweikh [Wed, 10 Aug 2005 17:18:07 +0000 (17:18 +0000)]
Be a nice community member and honor the request for backout after
stirring the hornet's nest. This issue will be resolved by core.

19 years agofix typo.
ume [Wed, 10 Aug 2005 17:09:08 +0000 (17:09 +0000)]
fix typo.

Submitted by: suz
Obtained from: KAME

19 years agoClarify/fix handling of the current channel:
sam [Wed, 10 Aug 2005 16:22:30 +0000 (16:22 +0000)]
Clarify/fix handling of the current channel:
o add ic_curchan and use it uniformly for specifying the current
  channel instead of overloading ic->ic_bss->ni_chan (or in some
  drivers ic_ibss_chan)
o add ieee80211_scanparams structure to encapsulate scanning-related
  state captured for rx frames
o move rx beacon+probe response frame handling into separate routines
o change beacon+probe response handling to treat the scan table
  more like a scan cache--look for an existing entry before adding
  a new one; this combined with ic_curchan use corrects handling of
  stations that were previously found at a different channel
o move adhoc neighbor discovery by beacon+probe response frames to
  a new ieee80211_add_neighbor routine

Reviewed by: avatar
Tested by: avatar, Michal Mertl
MFC after: 2 weeks

19 years agoRecord an error message if there are write errors when extracting the
kientzle [Wed, 10 Aug 2005 15:02:53 +0000 (15:02 +0000)]
Record an error message if there are write errors when extracting the
content of an archive entry to a file descriptor.

19 years agoMinor configuration fix to disable ACL support on MacOS X (which
kientzle [Wed, 10 Aug 2005 15:01:03 +0000 (15:01 +0000)]
Minor configuration fix to disable ACL support on MacOS X (which
lacks ACL_USER).

Thanks to:  Marcus Geiger, Joe Esch, and Markus Slopianka

19 years agoRemove unused variable.
cperciva [Wed, 10 Aug 2005 13:45:00 +0000 (13:45 +0000)]
Remove unused variable.

Reported by: stefanf

19 years agoMove <sys/cdefs.h> up to reduce diff to NetBSD.
stefanf [Wed, 10 Aug 2005 13:41:31 +0000 (13:41 +0000)]
Move <sys/cdefs.h> up to reduce diff to NetBSD.

19 years agoSubmitted to and merged from NetBSD (rev. 1.23 and 1.24):
stefanf [Wed, 10 Aug 2005 13:38:01 +0000 (13:38 +0000)]
Submitted to and merged from NetBSD (rev. 1.23 and 1.24):
- Don't delete the current line when typing `yy'.
- Don't use a possibly stale pointer in cv_paste().
-

19 years agoDo not drop the vnode interlock if vdropl is called on already doomed vnode.
kan [Wed, 10 Aug 2005 11:46:03 +0000 (11:46 +0000)]
Do not drop the vnode interlock if vdropl is called on already doomed vnode.
vdropl callers expect it to return with interlock still being held.

MFC after: 2 days

19 years agosupports stealth forwarding in IPv6, as well as in IPv4
suz [Wed, 10 Aug 2005 09:13:35 +0000 (09:13 +0000)]
supports stealth forwarding in IPv6, as well as in IPv4

PR: kern/54625
MFC after: 1 week

19 years agoRemove public declarations of variables that were forgotten when they were
obrien [Wed, 10 Aug 2005 07:10:02 +0000 (07:10 +0000)]
Remove public declarations of variables that were forgotten when they were
made static.

19 years agoRemove the need to forward declare statics by moving them around.
obrien [Wed, 10 Aug 2005 07:08:14 +0000 (07:08 +0000)]
Remove the need to forward declare statics by moving them around.

19 years agoMatch IPv6 and use a static struct pr_usrreqs nousrreqs.
obrien [Wed, 10 Aug 2005 06:41:04 +0000 (06:41 +0000)]
Match IPv6 and use a static struct pr_usrreqs nousrreqs.

19 years agoStyle nit.
obrien [Wed, 10 Aug 2005 06:38:46 +0000 (06:38 +0000)]
Style nit.

19 years agoAdd an option to tell what version of config(8) this is.
obrien [Wed, 10 Aug 2005 06:36:44 +0000 (06:36 +0000)]
Add an option to tell what version of config(8) this is.

19 years agoUse the ISO standard function variable vs. a GCC'ism.
obrien [Wed, 10 Aug 2005 06:25:41 +0000 (06:25 +0000)]
Use the ISO standard function variable vs. a GCC'ism.

19 years agofixed a kernel crash at the start-up time of an IPv6 multicast daemons o
suz [Wed, 10 Aug 2005 05:28:11 +0000 (05:28 +0000)]
fixed a kernel crash at the start-up time of an IPv6 multicast daemons o
(e.g. pim6dd, pim6sd)

MFC after: 3 days

19 years agoStyle cleanup.
obrien [Wed, 10 Aug 2005 04:01:21 +0000 (04:01 +0000)]
Style cleanup.

19 years agoFix FBSDid style nit.
obrien [Wed, 10 Aug 2005 03:56:31 +0000 (03:56 +0000)]
Fix FBSDid style nit.

19 years agoAdd proper copyright attribution.
obrien [Wed, 10 Aug 2005 03:54:37 +0000 (03:54 +0000)]
Add proper copyright attribution.

19 years agoRemove a reference to compute_stats(), since it is no longer documented at
ken [Wed, 10 Aug 2005 00:51:36 +0000 (00:51 +0000)]
Remove a reference to compute_stats(), since it is no longer documented at
least.

Submitted by: osa
MFC after: 1 week

19 years agoDon't allow pagedaemon to skip pages while scanning PQ_ACTIVE or PQ_INACTIVE
tegge [Wed, 10 Aug 2005 00:17:36 +0000 (00:17 +0000)]
Don't allow pagedaemon to skip pages while scanning PQ_ACTIVE or PQ_INACTIVE
due to the vm object being locked.

When a process writes large amounts of data to a file, the vm object associated
with that file can contain most of the physical pages on the machine.  If the
process is preempted while holding the lock on the vm object, pagedaemon would
be able to move very few pages from PQ_INACTIVE to PQ_CACHE or from PQ_ACTIVE
to PQ_INACTIVE, resulting in unlimited cleaning of dirty pages belonging to
other vm objects.

Temporarily unlock the page queues lock while locking vm objects to avoid lock
order violation.  Detect and handle relevant page queue changes.

This change depends on both the lock portion of struct vm_object and normal
struct vm_page being type stable.

Reviewed by: alc

19 years agoDocument kvm(3)-related error constants, and correct minor formatting
rwatson [Tue, 9 Aug 2005 22:19:30 +0000 (22:19 +0000)]
Document kvm(3)-related error constants, and correct minor formatting
nits.

19 years agoCall tulip_start() rather than tulip_ifstart() from the interrupt handler
jhb [Tue, 9 Aug 2005 21:53:26 +0000 (21:53 +0000)]
Call tulip_start() rather than tulip_ifstart() from the interrupt handler
to avoid recursing on the driver lock.  Not sure why my test box didn't
catch this earlier.

MFC after: 3 days

19 years agoThere's no reason to check the valence. This allows ciss to work
ps [Tue, 9 Aug 2005 20:53:51 +0000 (20:53 +0000)]
There's no reason to check the valence.  This allows ciss to work
on the P600.

19 years agoAdd helper function ip_findmoptions(), which accepts an inpcb, and attempts
rwatson [Tue, 9 Aug 2005 17:19:21 +0000 (17:19 +0000)]
Add helper function ip_findmoptions(), which accepts an inpcb, and attempts
to atomically return either an existing set of IP multicast options for the
PCB, or a newlly allocated set with default values.  The inpcb is returned
locked.  This function may sleep.

Call ip_moptions() to acquire a reference to a PCB's socket options, and
perform the update of the options while holding the PCB lock.  Release the
lock before returning.

Remove garbage collection of multicast options when values return to the
default, as this complicates locking substantially.  Most applications
allocate a socket either to be multicast, or not, and don't tend to keep
around sockets that have previously been used for multicast, then used for
unicast.

This closes a number of race conditions involving multiple threads or
processes modifying the IP multicast state of a socket simultaenously.

MFC after: 7 days

19 years agoGrammar improvements.
murray [Tue, 9 Aug 2005 14:11:57 +0000 (14:11 +0000)]
Grammar improvements.

PR: docs/84619
Submitted by: Gary W. Swearingen <garys@opusnet.com>
MFC after: 3 days

19 years agoFix typo.
murray [Tue, 9 Aug 2005 13:55:21 +0000 (13:55 +0000)]
Fix typo.

PR: docs/84660
Submitted by: Dirk Gouders <gouders@et.bocholt.fh-ge.de>
MFC after: 3 days

19 years agoNetBSD merged our typo fixes, update $NetBSD$.
stefanf [Tue, 9 Aug 2005 13:37:59 +0000 (13:37 +0000)]
NetBSD merged our typo fixes, update $NetBSD$.

19 years agoMerge a change I missed in the last commit.
stefanf [Tue, 9 Aug 2005 13:35:48 +0000 (13:35 +0000)]
Merge a change I missed in the last commit.

19 years agonForce3 and nForce4 are supported as well.
brueffer [Tue, 9 Aug 2005 13:32:45 +0000 (13:32 +0000)]
nForce3 and nForce4 are supported as well.

MFC after: 3 days

19 years agoMerge NetBSD's 1.25 which fixes a small bug introduced in 1.24.
stefanf [Tue, 9 Aug 2005 13:31:59 +0000 (13:31 +0000)]
Merge NetBSD's 1.25 which fixes a small bug introduced in 1.24.

19 years agoAdd an order between UDP inpcb locks and the IPv4 multicast address
rwatson [Tue, 9 Aug 2005 13:27:50 +0000 (13:27 +0000)]
Add an order between UDP inpcb locks and the IPv4 multicast address
list lock, as there has been a report that an alternative lock order
is getting introduced.  This should help ferret it out.

Reported by: Ed Maste <emaste at phaedrus dot sandvine dot ca>