]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoAdd documentation for log(9).
joerg [Tue, 10 Aug 2004 09:19:08 +0000 (09:19 +0000)]
Add documentation for log(9).

MFC after: 1 week

20 years ago - Use a new flag, KEF_XFERABLE, to record with certainty that this kse had
jeff [Tue, 10 Aug 2004 07:52:21 +0000 (07:52 +0000)]
 - Use a new flag, KEF_XFERABLE, to record with certainty that this kse had
   contributed to the transferable load count.  This prevents any potential
   problems with sched_pin() being used around calls to setrunqueue().
 - Change the sched_add() load balancing algorithm to try to migrate on
   wakeup.  This attempts to place threads that communicate with each other
   on the same CPU.
 - Don't clear the idle counts in kseq_transfer(), let the cpus do that when
   they call sched_add() from kseq_assign().
 - Correct a few out of date comments.
 - Make sure the ke_cpu field is correct when we preempt.
 - Call kseq_assign() from sched_clock() to catch any assignments that were
   done without IPI.  Presently all assignments are done with an IPI, but I'm
   trying a patch that limits that.
 - Don't migrate a thread if it is still runnable in sched_add().  Previously,
   this could only happen for KSE threads, but due to changes to
   sched_switch() all threads went through this path.
 - Remove some code that was added with preemption but is not necessary.

20 years agoFavor building the cdrtools port over fetching the package. There
marcel [Tue, 10 Aug 2004 05:43:40 +0000 (05:43 +0000)]
Favor building the cdrtools port over fetching the package. There
hasn't been any package build for ia64 for a long time and in this
case it means that the cdrtools package on the FTP sites is one that
doesn't have mkisofs. This breaks the release build. By trying to
build the port first, we protect ourselves a bit more against this.

20 years agoUpdate wording regarding geography to match the man page and web site.
murray [Tue, 10 Aug 2004 05:34:23 +0000 (05:34 +0000)]
Update wording regarding geography to match the man page and web site.
Specifically, change the second level menu title from 'Country' to
'Country or Region', since e.g. Hong Kong is not a country.

Submitted by: Xin LI <delphij@frontfree.net>

20 years agoDisable the snmp_atm module until the source gets fixed. This will unbreak
scottl [Tue, 10 Aug 2004 04:56:39 +0000 (04:56 +0000)]
Disable the snmp_atm module until the source gets fixed.  This will unbreak
this portion of the build.

20 years agoDon't acquire Giant in cryptof_close(), as the code is intended to be
rwatson [Tue, 10 Aug 2004 03:26:17 +0000 (03:26 +0000)]
Don't acquire Giant in cryptof_close(), as the code is intended to be
able to run MPsafe (and appears to be MPsafe).

Discussed with (some time ago): sam

20 years agoInline umich license from COPYRIGHT to make it clear what license the
rwatson [Tue, 10 Aug 2004 03:23:05 +0000 (03:23 +0000)]
Inline umich license from COPYRIGHT to make it clear what license the
umich copyright is asserting.

Clarify that the copyright I'm asserting is the standard Berkeley
license.

Remove Giant assertions from AARP and DDP input routines.

20 years agoInitial pass at shifting sysinstall(8) to install X.org instead of XFree86.
kensmith [Tue, 10 Aug 2004 02:18:28 +0000 (02:18 +0000)]
Initial pass at shifting sysinstall(8) to install X.org instead of XFree86.
There are still a few nits to work out (graphical config non-functional).

Patches by: anholt@, adapted slightly
Reviewed by: 'make release'

20 years agoCatch up with change to <machine/pte.h>.
marcel [Tue, 10 Aug 2004 02:08:57 +0000 (02:08 +0000)]
Catch up with change to <machine/pte.h>.

20 years agoBy default, the watch utility will attempt to open /dev/snp0, if
csjp [Tue, 10 Aug 2004 01:49:46 +0000 (01:49 +0000)]
By default, the watch utility will attempt to open /dev/snp0, if
another process already has /dev/snp0 open, the snp(4) will return
EBUSY, in which case watch will try to open /dev/snp1..9. Currently
watch does not check errno to see if the failure was a result of EBUSY.

This results in watch making futile attempts to open snp0..snp9 even
though devices may not exist or the caller does not have permissions
to access the device.

In addition to this, it attempts to setup the screen for snooping even
though it may not ever get an snp device.

So this patch does two things
1) Checks errno for EBUSY, if open(2) fails for another reason
   print that reason and exit.
2) setup the terminal for snooping after the snp descriptor has
   been obtained.

Approved by: bmilekic (mentor)

20 years agoSkip the syncing disks loop if there are no dirty buffers. Remove a
njl [Tue, 10 Aug 2004 01:32:05 +0000 (01:32 +0000)]
Skip the syncing disks loop if there are no dirty buffers.  Remove a
variable used to flag the initial printf.

Submitted by: truckman (earlier version)

20 years agoIntroduce ng_hci_inquiry_response structure and use it in the hccontrol(8)
emax [Tue, 10 Aug 2004 00:38:50 +0000 (00:38 +0000)]
Introduce ng_hci_inquiry_response structure and use it in the hccontrol(8)

20 years agoAdd a temporary debugging hack to detect a deadlock in setrunqueue(). This
scottl [Tue, 10 Aug 2004 00:26:25 +0000 (00:26 +0000)]
Add a temporary debugging hack to detect a deadlock in setrunqueue().  This
is here so that we can gather stats on the nature of the recent rash of
hard lockups, and in this particular case panic the machine instead of
letting it deadlock forever.

20 years agoIn syscall, always make a copy of parameters from trapframe, this
davidxu [Mon, 9 Aug 2004 23:57:59 +0000 (23:57 +0000)]
In syscall, always make a copy of parameters from trapframe, this
becauses some syscalls using set_mcontext can sneakily change
parameters and later when those syscalls references parameters,
they will wrongly use register values in mcontext_t.

Approved by: peter

20 years agoClean up and add coments.
julian [Mon, 9 Aug 2004 23:24:22 +0000 (23:24 +0000)]
Clean up and add coments.
reserve some bits in flag words for upcoming work.
Note the unused bits so they can be found easier.

20 years agoAdd twa(4) to the Hardware Notes.
simon [Mon, 9 Aug 2004 22:33:20 +0000 (22:33 +0000)]
Add twa(4) to the Hardware Notes.

20 years ago- Add a HARDWARE section which lists supported devices.
simon [Mon, 9 Aug 2004 22:27:33 +0000 (22:27 +0000)]
- Add a HARDWARE section which lists supported devices.

20 years agoSlight changes to comments and some whitespace changes.
julian [Mon, 9 Aug 2004 21:57:30 +0000 (21:57 +0000)]
Slight changes to comments and some whitespace changes.

20 years agoDuring FTP installs, when installing Postfix as an alternate MTA, the
simon [Mon, 9 Aug 2004 21:50:10 +0000 (21:50 +0000)]
During FTP installs, when installing Postfix as an alternate MTA, the
install of Postfix fails since Postfix 1 is picked from INDEX instead
of Postfix 2.

The problem is that the package_add function matches multiple Postfix
packages and then installs a "random" version (the last one found in
INDEX).  This does not occur with a CD-ROM install since there is only
one Postfix package on the CD-ROM.

The correct solution to this would be to check the origin field from
INDEX instead of the package name, but due to the way sysinstall is
made that is not trivial, so instead work around the problem by
hard-coding the Postfix major and minor revision (for the current
stable version) to install directly into the package_add call.

PR: misc/65426
Reported by: Harold Kachelmyer <bugs@princessharold.net>
Approved by: trhodes

20 years agoThere was a star danced, and under that was I born.
oliver [Mon, 9 Aug 2004 21:32:25 +0000 (21:32 +0000)]
There was a star danced, and under that was I born.
 -- Shakespeare in "Much Ado About Nothing"

Add my birthday in case someone cares someday...

20 years agoMinimal fix to prevent crashes when an AGP v2 card is used with the new v3 VIA
anholt [Mon, 9 Aug 2004 21:01:49 +0000 (21:01 +0000)]
Minimal fix to prevent crashes when an AGP v2 card is used with the new v3 VIA
chipsets, based on Linux's via-agp.c.  On boot, the system selects which AGP
version to use based on the inserted card.  If v2 was chosen, the chipset
needs to be programmed with the v2 registers still.  Also included in kern/69953
are changes to make the programming of the v3 registers match linux, but that
will be left out until the need to do so is confirmed (want specs or a tester).

PR: kern/69953
Submitted by: Oleg Sharoiko <os@rsu.ru>
Tested by: Oleg Sharoiko <os@rsu.ru>, Geoff Speicher <geoff@speicher.org>
(full version from PR)

20 years agoBetter preserve the original protection for the mappings we maintain.
marcel [Mon, 9 Aug 2004 20:44:41 +0000 (20:44 +0000)]
Better preserve the original protection for the mappings we maintain.
The hardware always gives read access for privilege level 0, which
means that we cannot use the hardware access rights and privilege
level in the PTE to test whether there's a change in protection.  So,
we save the original vm_prot_t in the PTE as well.
Add pmap_pte_prot() to set the proper access rights and privilege
level on the PTE given a pmap and the requested protection.

The above allows us to compare the protection in pmap_extract_and_hold()
which was missing. While in pmap_extract_and_hold(), add pmap locking.

While here, clean up most (i.e. all but one) PTE macros we inherited
from alpha. They were either unused, used inconsistently, badly named
or simply weren't beneficial. We save the wired and managed state of
the PTE in distinct (bit) fields.

While in pte.h, s/u_int64_t/uint64_t/g

pmap locking obtained from: alc@
feedback & review by: alc@

20 years agoMake kg->kg_runnable actually count runnable threads in the ksegrp run queue
julian [Mon, 9 Aug 2004 20:36:03 +0000 (20:36 +0000)]
Make kg->kg_runnable actually count runnable threads in the ksegrp run queue
instead of only doing it sometimes.. This is not used outdide of debugging code
in the current code, but that will probably change.

20 years agoFix rl(4)'s lock behavior upon deinitialization. I would get a panic
green [Mon, 9 Aug 2004 20:22:17 +0000 (20:22 +0000)]
Fix rl(4)'s lock behavior upon deinitialization.  I would get a panic
when kldunloading due to its private locking being acquired recursively.

20 years agoRemove typos on KASSERT messages.
julian [Mon, 9 Aug 2004 20:13:07 +0000 (20:13 +0000)]
Remove typos on KASSERT messages.

20 years agoRevamp VM map wiring.
green [Mon, 9 Aug 2004 19:52:29 +0000 (19:52 +0000)]
Revamp VM map wiring.

* Allow no-fault wiring/unwiring to succeed for consistency;
  however, the wired count remains at zero, so it's a special case.

* Fix issues inside vm_map_wire() and vm_map_unwire() where the
  exact state of user wiring (one or zero) and system wiring
  (zero or more) could be confused; for example, system unwiring
  could succeed in removing a user wire, instead of being an
  error.

* Require all mappings to be unwired before they are deleted.
  When VM space is still wired upon deletion, it will be waited
  upon for the following unwire.  This makes vslock(9) work
  rather than allowing kernel-locked memory to be deleted
  out from underneath of its consumer as it would before.

20 years agoShift to the same version of perl that portmgr@ has as their default.
kensmith [Mon, 9 Aug 2004 19:36:53 +0000 (19:36 +0000)]
Shift to the same version of perl that portmgr@ has as their default.

Pointed out by: jhb

20 years agoFix a 4.X -> 5.X upgrade problem: crunchgen used to emit a line 'MAKE=make'
harti [Mon, 9 Aug 2004 19:18:19 +0000 (19:18 +0000)]
Fix a 4.X -> 5.X upgrade problem: crunchgen used to emit a line 'MAKE=make'
or 'env MAKEOBJDIRPREFIX=... make' depending on the setting of
MAKEOBJDIRPREFIX in the environment. In any case this line kills the
original value of ${MAKE}. When during buildworld a new make is built (as
is the case during the upgrade) this causes a wrong make to be picked up
(the first one in the path). Use the same technique as Makefile.inc1:
create a MAKEENV variable and a CRUNCHMAKE that calls ${MAKE} with that
MAKEENV prefixed. Use CRUNCHMAKE instead of MAKE throughout the generated
makefile. This leaves the original ${MAKE} undisturbed.

20 years agoMake two changes to vm_fault().
alc [Mon, 9 Aug 2004 18:46:39 +0000 (18:46 +0000)]
Make two changes to vm_fault().
1. Move a comment to its proper place, updating it.  (Except for white-
   space, this comment had been unchanged since revision 1.1!)
2. Remove spl calls.

20 years agoNormalize the VM wiring done with SPARSE_MAPPING: check for errors, and
green [Mon, 9 Aug 2004 18:46:13 +0000 (18:46 +0000)]
Normalize the VM wiring done with SPARSE_MAPPING: check for errors, and
unmap when done.  For whatever reason, SPARSE_MAPPING is not even a
config option, so this is dead code.

20 years agoIncrease the amount of data exported by KTR in the KTR_RUNQ setting.
julian [Mon, 9 Aug 2004 18:21:12 +0000 (18:21 +0000)]
Increase the amount of data exported by KTR in the KTR_RUNQ setting.
This extra data is needed to really follow what is going on in the
threaded case.

20 years agoRemove a stale comment from vm_map_lookup() that pertains to share maps.
alc [Mon, 9 Aug 2004 18:15:46 +0000 (18:15 +0000)]
Remove a stale comment from vm_map_lookup() that pertains to share maps.
(The last vestiges of the share map code were removed in revisions 1.153
and 1.159.)

20 years agoRemove commented out pcic driver. It is too broken to work (even if you
imp [Mon, 9 Aug 2004 17:36:19 +0000 (17:36 +0000)]
Remove commented out pcic driver.  It is too broken to work (even if you
fix the obvious bugs, nastier ones reside below the surfac), and having
it commented out here just encourages people to try it.

# I'm not removing it from the base system, yet.

20 years agoLink the snmp_atm module to the build unless NOATM is defined.
harti [Mon, 9 Aug 2004 16:40:39 +0000 (16:40 +0000)]
Link the snmp_atm module to the build unless NOATM is defined.

20 years agoNow that gensnmptree can read and merge more than one tree make
harti [Mon, 9 Aug 2004 16:37:25 +0000 (16:37 +0000)]
Now that gensnmptree can read and merge more than one tree make
the rule so that the tree .c and .h files can be produced from several
trees.

20 years agoA module for bsnmpd(1) that exports a MIB that contains information about
harti [Mon, 9 Aug 2004 16:36:04 +0000 (16:36 +0000)]
A module for bsnmpd(1) that exports a MIB that contains information about
the local ATM interfaces and allows to change some parameters via SNMP.

20 years agoMake a comment that IP source routing is not SMP and PREEMPTION safe.
andre [Mon, 9 Aug 2004 16:17:37 +0000 (16:17 +0000)]
Make a comment that IP source routing is not SMP and PREEMPTION safe.

20 years agoMake a comment that "ipfw forward" is not SMP and PREEMPTION safe.
andre [Mon, 9 Aug 2004 16:16:10 +0000 (16:16 +0000)]
Make a comment that "ipfw forward" is not SMP and PREEMPTION safe.

20 years agoMake explicit that MAKEOBJDIRPREFIX and MAKOBJDIR must be set
harti [Mon, 9 Aug 2004 16:13:54 +0000 (16:13 +0000)]
Make explicit that MAKEOBJDIRPREFIX and MAKOBJDIR must be set
as environment variables and should not be set on make's command
line. They happen to work accidentially as command line variables
too when none of the sub-makes wants to play games with them (because
make is putting command line variables into the environment and will
find them there later on). Makefile.inc1 wants to change
MAKEOBJDIRPREFIX. In this case one cannot set it on the command line.

20 years agoNew ipfw option "antispoof":
andre [Mon, 9 Aug 2004 16:12:10 +0000 (16:12 +0000)]
New ipfw option "antispoof":

 For incoming packets, the packet's source address is checked if it
 belongs to a directly connected network.  If the network is directly
 connected, then the interface the packet came on in is compared to
 the interface the network is connected to.  When incoming interface
 and directly connected interface are not the same, the packet does
 not match.

Usage example:

 ipfw add deny ip from any to any not antispoof in

Manpage education by: ru

20 years agomdoc(7) and english grammar makeover.
mlaier [Mon, 9 Aug 2004 15:58:39 +0000 (15:58 +0000)]
mdoc(7) and english grammar makeover.

Submitted by: ru

20 years agoPer letter dated July 22, 1999 remove 3rd clause of Berkeley derived
dds [Mon, 9 Aug 2004 15:29:41 +0000 (15:29 +0000)]
Per letter dated July 22, 1999 remove 3rd clause of Berkeley derived
software (original contributor).

Reviewed by: imp

20 years agoThe multiplier prefix is actually a multiplier suffix.
des [Mon, 9 Aug 2004 14:43:50 +0000 (14:43 +0000)]
The multiplier prefix is actually a multiplier suffix.

20 years agoFix bad free() usage.
sos [Mon, 9 Aug 2004 14:22:58 +0000 (14:22 +0000)]
Fix bad free() usage.

Submitted by: Pawel Jakub Dawidek

20 years agoWrite my country name in standard form
den [Mon, 9 Aug 2004 13:45:43 +0000 (13:45 +0000)]
Write my country name in standard form

Requested by: ru

20 years agoChange the behaviour of `-v' so that, e.g., stepping a month back
yar [Mon, 9 Aug 2004 13:43:39 +0000 (13:43 +0000)]
Change the behaviour of `-v' so that, e.g., stepping a month back
on March 31 won't take you to March 2 or 3 (now the result will
be the last day of February.)

In general, now stepping by months from the last days of the current
month A will take you to the very last day of the target month B if
B is shorter than A.

The previous version would just step to March 31 and rely on mktime(3)
to correct the date.  Despite its simplicity, such way was counter-intuitive
to users and caused pain to shell script writers.

Noticed by: Igor Timkin <ivt at gamma dot ru>
Approved by: brian
MFC after: 2 weeks

20 years agoRemove another mention of "make world".
ru [Mon, 9 Aug 2004 13:25:06 +0000 (13:25 +0000)]
Remove another mention of "make world".

20 years agoFix a documentation bug: "make release" uses "make buildworld"
ru [Mon, 9 Aug 2004 13:22:44 +0000 (13:22 +0000)]
Fix a documentation bug: "make release" uses "make buildworld"
not "make world", as the first step of building the release.
While here, remove another mention of "make world", as the
latter is no longer supported by src/Makefile.

20 years agoAdd myself to USSR count :)
den [Mon, 9 Aug 2004 12:59:43 +0000 (12:59 +0000)]
Add myself to USSR count :)

20 years agoClose a race in ata_reinit().
sos [Mon, 9 Aug 2004 12:02:32 +0000 (12:02 +0000)]
Close a race in ata_reinit().

20 years agoMake make recurse into sub-directories and sub-makes when given
harti [Mon, 9 Aug 2004 11:38:41 +0000 (11:38 +0000)]
Make make recurse into sub-directories and sub-makes when given
two -n flags.  If only one -n flag is given the old behaviour
is retained (POLA).  In order to make this working for installworld
change the IMAKEENV in this case so that the tools are found
(we have no temporary installation environment in this case).

Submitted by: ru (IMAKEENV part)

20 years agoAnd here I am. The first person from Poland in here.
pjd [Mon, 9 Aug 2004 11:37:14 +0000 (11:37 +0000)]
And here I am. The first person from Poland in here.

20 years agoTurn on the magic.
obrien [Mon, 9 Aug 2004 11:30:36 +0000 (11:30 +0000)]
Turn on the magic.

20 years ago- Introduce option for hardcoding providers' names into metadata.
pjd [Mon, 9 Aug 2004 11:29:42 +0000 (11:29 +0000)]
- Introduce option for hardcoding providers' names into metadata.
  It allows to fix problems when last provider's sector is shared between few
  providers.
- Bump version number for CONCAT and STRIPE and add code for backward
  compatibility.
- Do not bump version number of MIRROR, as it wasn't officially introduced yet.
  Even if someone started to play with it, there is no big deal, because
  wrong MD5 sum of metadata will deny those providers.
- Update manual pages.
- Add version history to g_(stripe|concat).h files.

20 years agoAdd myself here. Increase USSR count ;-)
rik [Mon, 9 Aug 2004 11:17:59 +0000 (11:17 +0000)]
Add myself here. Increase USSR count ;-)

20 years agoDo not use g_wither_geom(9). I doesn't work in the way which is expected
pjd [Mon, 9 Aug 2004 11:14:25 +0000 (11:14 +0000)]
Do not use g_wither_geom(9). I doesn't work in the way which is expected
here anymore (after g_wither_washer() was introduced), i.e. geom and consumer
will not be immediately destroyed if possible.

20 years agoDon't try to programatically support running file(1) w/in /usr/obj or other
obrien [Mon, 9 Aug 2004 11:05:09 +0000 (11:05 +0000)]
Don't try to programatically support running file(1) w/in /usr/obj or other
non-standard install path.

Agreed with: ru

20 years agoUse the '+' flag to make make recurse into sub-directories even when
harti [Mon, 9 Aug 2004 10:54:05 +0000 (10:54 +0000)]
Use the '+' flag to make make recurse into sub-directories even when
given -n. For POLA reasons this behaviour is switched on only when
at least two -n flags are given to make. One -n flag keeps the old behaviour
of showing the shell command that would recurse into the sub-directories.

Discussed with: ru

20 years agoThe file(1) related build-tool moved to libmagic.
obrien [Mon, 9 Aug 2004 10:51:01 +0000 (10:51 +0000)]
The file(1) related build-tool moved to libmagic.

20 years agoUncomment the real paths and remove the local debugging paths.
obrien [Mon, 9 Aug 2004 10:47:50 +0000 (10:47 +0000)]
Uncomment the real paths and remove the local debugging paths.

Noticed by: ru (who's faster than 'make world' on my stock src test machine)

20 years agoUncomment the real paths and remove the local debugging paths.
obrien [Mon, 9 Aug 2004 10:32:16 +0000 (10:32 +0000)]
Uncomment the real paths and remove the local debugging paths.

Noticed by: ru (who's faster than 'make world' on my stock src test machine)

20 years agoadd myself. :)
leeym [Mon, 9 Aug 2004 10:22:52 +0000 (10:22 +0000)]
add myself. :)

20 years agoThe file consumer binary, which depends on the libmagic.
obrien [Mon, 9 Aug 2004 08:49:49 +0000 (08:49 +0000)]
The file consumer binary, which depends on the libmagic.

20 years agoBmake the library containing and processing the magic.
obrien [Mon, 9 Aug 2004 08:48:28 +0000 (08:48 +0000)]
Bmake the library containing and processing the magic.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r133359,
obrien [Mon, 9 Aug 2004 08:45:41 +0000 (08:45 +0000)]
This commit was generated by cvs2svn to compensate for changes in r133359,
which included commits to RCS files with non-trunk default branches.

20 years agoVirgin import of Christos Zoulas's FILE 4.10.
obrien [Mon, 9 Aug 2004 08:45:41 +0000 (08:45 +0000)]
Virgin import of Christos Zoulas's FILE 4.10.
*- file is now broken into a library containing and processing the magic
   and a consumer binary.

20 years agomake sure that ioctl's to ata use the proper cmd IOCATA...
jmg [Mon, 9 Aug 2004 06:47:53 +0000 (06:47 +0000)]
make sure that ioctl's to ata use the proper cmd IOCATA...

Reviewed by: sos

20 years agochange the name of the md module, to g_md, introduce a define with the
jmg [Mon, 9 Aug 2004 06:45:20 +0000 (06:45 +0000)]
change the name of the md module, to g_md, introduce a define with the
name MD_MODNAME, and make mdconfig use this new define...

20 years agoToo many versions.
phk [Mon, 9 Aug 2004 06:04:00 +0000 (06:04 +0000)]
Too many versions.

Spotted by: pjd

20 years agoMake two changes to vm_fault().
alc [Mon, 9 Aug 2004 06:01:46 +0000 (06:01 +0000)]
Make two changes to vm_fault().
1. Retain the map lock until after the calls to pmap_enter() and
   vm_fault_prefault().
2. Remove a stale comment.  Submitted by: tegge@

20 years agoadd option to automaticly mark core dumps with the nodump flag
jmg [Mon, 9 Aug 2004 05:46:46 +0000 (05:46 +0000)]
add option to automaticly mark core dumps with the nodump flag

PR: 57065
Submitted by: Walter C. Pelissero

20 years agofix bug which prevented programming function keys that were exactly 16
jmg [Mon, 9 Aug 2004 04:27:58 +0000 (04:27 +0000)]
fix bug which prevented programming function keys that were exactly 16
characters long..  strcpy was coping over the length...

PR: 52960
Submitted by: Dmitry Sivachenko
MFC after: 1 week

20 years agoFix long standing mediaopt setting bugs seen on sparc64. Though
yongari [Mon, 9 Aug 2004 03:13:57 +0000 (03:13 +0000)]
Fix long standing mediaopt setting bugs seen on sparc64. Though
the bug exists in little-endian machine, it was not triggered due
to the difference of memory ordering between little/big endian
machines. Instead of relying on possibly modified value during
function invokcations, use saved copy of ifr.ifr_addr.sa_family.
Also add a comment at the top of ifconfig.c clarifying the issue
so the bug won't re-appear.

Approved by: jake
Reviewed by: yar

20 years agoI should have marked my previous commit with an XXX to indicate it's a
kensmith [Mon, 9 Aug 2004 01:42:32 +0000 (01:42 +0000)]
I should have marked my previous commit with an XXX to indicate it's a
workaround to a problem that shouldn't exist.

Noticed by: ru

20 years agodon't seg if the syscall is invalid for i386fbsd.
alfred [Sun, 8 Aug 2004 23:29:36 +0000 (23:29 +0000)]
don't seg if the syscall is invalid for i386fbsd.

20 years agoUse fallthrough to simplify the multiplier logic; optimistically add
des [Sun, 8 Aug 2004 23:14:44 +0000 (23:14 +0000)]
Use fallthrough to simplify the multiplier logic; optimistically add
support for the T multiplier; improve the error message for unrecognized
multipliers.

20 years agoFix some whitespace issues, and move a curly brace out of an #ifdef to
des [Sun, 8 Aug 2004 23:11:43 +0000 (23:11 +0000)]
Fix some whitespace issues, and move a curly brace out of an #ifdef to
avoid confusing auto-indenting editors.

20 years agoInitialize thread_db module.
davidxu [Sun, 8 Aug 2004 22:46:22 +0000 (22:46 +0000)]
Initialize thread_db module.

20 years ago1. Add some code check if thread suspending or resuming is failed.
davidxu [Sun, 8 Aug 2004 22:45:36 +0000 (22:45 +0000)]
1. Add some code check if thread suspending or resuming is failed.
2. Add code to retrieve thread tls address.

20 years agoCheck debugger suspending flag for system scope thread.
davidxu [Sun, 8 Aug 2004 22:42:11 +0000 (22:42 +0000)]
Check debugger suspending flag for system scope thread.

Reviewed by: deischen

20 years agoAuto generate device listings the following drivers: adv, amd, bt, ida,
simon [Sun, 8 Aug 2004 22:39:19 +0000 (22:39 +0000)]
Auto generate device listings the following drivers: adv, amd, bt, ida,
iir, ips, nsp, ohci, and stg.

20 years ago1.Use new way to check if a thread is in critical region, defer suspending
davidxu [Sun, 8 Aug 2004 22:37:53 +0000 (22:37 +0000)]
1.Use new way to check if a thread is in critical region, defer suspending
  if it is true.
2.Add thread_db api td_thr_tls_get_addr to get tls address, the real code
  is commented out util tls patch is committed.

Reviewed by: deischen

20 years ago1.Add KSE_INTR_DBSUSPEND command for kse_thr_interrupt to suspend a bound
davidxu [Sun, 8 Aug 2004 22:32:20 +0000 (22:32 +0000)]
1.Add KSE_INTR_DBSUSPEND command for kse_thr_interrupt to suspend a bound
thread, after the bound thread leaves critical region, the thread should
check debug flag may suspend itself by using the command.
2.Schedule upcall after thread is suspended by debugger
3.Wakeup upcall thread after process suspension.

Reviewed by: deischen

20 years agoCall thread_user_enter for M:N thread, ast() should be treated as another
davidxu [Sun, 8 Aug 2004 22:28:33 +0000 (22:28 +0000)]
Call thread_user_enter for M:N thread, ast() should be treated as another
entrance of kernel.

20 years agoBump document date for last commit.
simon [Sun, 8 Aug 2004 22:27:02 +0000 (22:27 +0000)]
Bump document date for last commit.

20 years agoAdd pl_flags to ptrace_lwpinfo, two flags PL_FLAG_SA and PL_FLAG_BOUND
davidxu [Sun, 8 Aug 2004 22:26:11 +0000 (22:26 +0000)]
Add pl_flags to ptrace_lwpinfo, two flags PL_FLAG_SA and PL_FLAG_BOUND
indicate that a thread is in UTS critical region.

Reviewed by: deischen
Approved by: marcel

20 years ago- Add a HARDWARE section which lists supported devices.
simon [Sun, 8 Aug 2004 22:19:05 +0000 (22:19 +0000)]
- Add a HARDWARE section which lists supported devices.

20 years ago- Add a HARDWARE section which lists supported devices based on the
simon [Sun, 8 Aug 2004 21:47:40 +0000 (21:47 +0000)]
- Add a HARDWARE section which lists supported devices based on the
  manual page and the Hardware Notes.

20 years ago- Add support for the .Dq macro.
simon [Sun, 8 Aug 2004 21:28:00 +0000 (21:28 +0000)]
- Add support for the .Dq macro.

20 years agoImplement C99's standard header <tgmath.h>. It provides type-generic macros
stefanf [Sun, 8 Aug 2004 20:05:47 +0000 (20:05 +0000)]
Implement C99's standard header <tgmath.h>.  It provides type-generic macros
for the <math.h> and <complex.h> functions that have float, double and long
double implementations.  Such type-generic macros expand to an actual
function, depending on the types of the macro arguments, eg. if <tgmath.h>
is included, the invocation cos(1.0f) calls the function cosf().

20 years agos/FALLTHOUGH/FALLTHROUGH/
dwmalone [Sun, 8 Aug 2004 18:59:19 +0000 (18:59 +0000)]
s/FALLTHOUGH/FALLTHROUGH/

Submitted by: Xin LI <delphij@frontfree.net>

20 years agogtar uses -F for another purpose, so bsdtar no longer treats it as a
kientzle [Sun, 8 Aug 2004 17:20:23 +0000 (17:20 +0000)]
gtar uses -F for another purpose, so bsdtar no longer treats it as a
synonym for --format.  Update the man page to reflect this.  While
I'm here, change the man page to document "tar" rather than "bsdtar,"
update some comments about -l compatibility and fix a few grammar nits.

20 years agoAdd missing bit of last if_start workaround: mark scan callout
sam [Sun, 8 Aug 2004 17:10:54 +0000 (17:10 +0000)]
Add missing bit of last if_start workaround: mark scan callout
MPSAFE only debug_mpsafenet is 1 so callbacks to send management
frames hold Giant; this is another bandaid on the path to removing Giant.

20 years agoUse &url.*; for URLs in www.FreeBSD.org.
hrs [Sun, 8 Aug 2004 15:08:29 +0000 (15:08 +0000)]
Use &url.*; for URLs in www.FreeBSD.org.

20 years agoSimplify parameter entities in doctype declaration.
hrs [Sun, 8 Aug 2004 14:59:27 +0000 (14:59 +0000)]
Simplify parameter entities in doctype declaration.

20 years agouse bufdone() not biodone().
phk [Sun, 8 Aug 2004 13:23:05 +0000 (13:23 +0000)]
use bufdone() not biodone().

20 years agoUse bufdone(), not biodone().
phk [Sun, 8 Aug 2004 13:20:43 +0000 (13:20 +0000)]
Use bufdone(), not biodone().

20 years agoInclude the nanoBSD make.conf in the top-level nanoBSD Makefile, so
thomas [Sun, 8 Aug 2004 13:05:03 +0000 (13:05 +0000)]
Include the nanoBSD make.conf in the top-level nanoBSD Makefile, so
all configuration knobs (both system settings and nanoBSD-specific ones)
can be tuned by modifying a single file. Move KERNCONF default value
to make.conf.

Attempt to resolve CFGMASTER like CUSTOMIZE, first relative to ${.CURDIR},
then as an absolute path.

Allow optional additional arguments to be passed to the CUSTOMIZE script.

20 years agoLooks like the x.org manuals are in x11/xorg-manpages, not x11/xorg-manuals.
kensmith [Sun, 8 Aug 2004 10:50:19 +0000 (10:50 +0000)]
Looks like the x.org manuals are in x11/xorg-manpages, not x11/xorg-manuals.

20 years agoMake sure that AT_PHDR has a useful value even for static programs.
dfr [Sun, 8 Aug 2004 09:48:10 +0000 (09:48 +0000)]
Make sure that AT_PHDR has a useful value even for static programs.