]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
17 years agoFix "pkg_version -Io" coredump, if some packages have no origin at
krion [Sat, 12 Aug 2006 18:32:51 +0000 (18:32 +0000)]
Fix "pkg_version -Io" coredump, if some packages have no origin at
all.

PR: bin/101932
Submitted by: novel, sat
MFC after: 5 days

17 years agoOops. Remove accidentally committed change.
pjd [Sat, 12 Aug 2006 18:29:49 +0000 (18:29 +0000)]
Oops. Remove accidentally committed change.

Noticed by: marck

17 years agoRestore the "plumb" compatibility command, which was lost
yar [Sat, 12 Aug 2006 18:07:17 +0000 (18:07 +0000)]
Restore the "plumb" compatibility command, which was lost
due to the recent change by sam@ to clone operations in ifconfig(8).

17 years agoFix incorrect .Xr syntax I introduced in the previous commit.
yar [Sat, 12 Aug 2006 17:11:43 +0000 (17:11 +0000)]
Fix incorrect .Xr syntax I introduced in the previous commit.

Noticed by: ru
X-MFC with: rev. 1.31

17 years agoHandle MSDOS file systems properly. Before the change file systems
pjd [Sat, 12 Aug 2006 15:34:15 +0000 (15:34 +0000)]
Handle MSDOS file systems properly. Before the change file systems
created on Windows XP (and others maybe) were not detected.
We detected only those created with newfs_msdos(8).

Submitted by: Tobias Reifenberger <treif@mayn.de>
style(9)ified by: pjd

17 years agoVerify if a label doesn't point to the parent directory.
pjd [Sat, 12 Aug 2006 15:30:24 +0000 (15:30 +0000)]
Verify if a label doesn't point to the parent directory.

17 years agoThe strstr() function is in the libkern now.
pjd [Sat, 12 Aug 2006 15:29:12 +0000 (15:29 +0000)]
The strstr() function is in the libkern now.

17 years agoAdd strstr() function to the libkern.
pjd [Sat, 12 Aug 2006 15:28:39 +0000 (15:28 +0000)]
Add strstr() function to the libkern.

17 years agoI don't know where I got the idea that the VIA C3 has 3DNow!; it doesn't.
des [Sat, 12 Aug 2006 09:46:43 +0000 (09:46 +0000)]
I don't know where I got the idea that the VIA C3 has 3DNow!; it doesn't.
It does have MMX (though MMX support is reputed to be incomplete in early
generations), and later generations have SSE.

MFC after: 2 weeks

17 years agoBatch of changes:
imp [Sat, 12 Aug 2006 09:06:55 +0000 (09:06 +0000)]
Batch of changes:
o when turning off the socket for a 16-bit card, write 0 to INTR register
  rather than just tying to just clear the rest bit.  this seems to fix
  card insert detection after an eject on TI bridges (ricoh bridges work
  either way, apparently).  This is a MFp4.
o Cope better with TOPIC95 bridges on powerup.  According to NetBSD driver,
  these bridges don't set POWER_STATE, so cope accordingly in our power
  code.  They also need a little extra time to settle, so do that as well.
o It appears that we need to turn on/off one of the clocks to the card
  when we power up/down that socket on a TOPIC97, also from NetBSD.
o TOPIC97 bridges need to specifically enable LV card support.  Unconditionally
  do this in the hopes that all laptops that have these chips support LV
  voltages (they should, since they are required for CardBus).
o TOPIC register name regularization.  Registers specific to models of TOPIC
  are now called out as such.

# I need a machine with a TOPIC95 for testing.

17 years agoThe TOPIC97 and TOPIC100 seem to have a special register in the exca
imp [Sat, 12 Aug 2006 08:53:54 +0000 (08:53 +0000)]
The TOPIC97 and TOPIC100 seem to have a special register in the exca
space that enables low voltage operation (and maybe other stuff).
Enable the bits in this register so low voltage 16-bit cards may work.

Existance noticed in NetBSD driver.

17 years agoDon't reset Tx threshold value whenever xl_init_locked() is called.
yongari [Sat, 12 Aug 2006 02:05:13 +0000 (02:05 +0000)]
Don't reset Tx threshold value whenever xl_init_locked() is called.
Instead the threshould is initialized in device attach. Later the
threshold could be increased in Tx underrun error and the new
threshold should be used in xl_init_locked().

17 years agoMake sure to check frames in Tx queue are empty before clearing
yongari [Sat, 12 Aug 2006 01:55:51 +0000 (01:55 +0000)]
Make sure to check frames in Tx queue are empty before clearing
watchdog timer.

17 years agoDue to the poor PHY documentation from RealTek I can't sure but I
yongari [Sat, 12 Aug 2006 01:38:49 +0000 (01:38 +0000)]
Due to the poor PHY documentation from RealTek I can't sure but I
think the RealTek PHY needs driver to set RGEPHY_BMCR_AUTOEN bit of
RGEPHY_MII_BMCR register and proper ANAR register setting for manual
media type selection.
This fixes long standing manual media type selection bug in rgephy(4).

Reported by: Jelte Jansen <jelte AT NLnetLabs DOT nl>
Tested by: Jelte Jansen <jelte AT NLnetLabs DOT nl>

17 years agoFix invalid reference of mbuf chains.
yongari [Sat, 12 Aug 2006 01:30:38 +0000 (01:30 +0000)]
Fix invalid reference of mbuf chains.
Use proper pointer dereference to inform modified mbuf chains to
caller.

While I'm here perform checksum offload setup after loading DMA
maps.

In collaboration with:  glebius

17 years agoFix invalid reference of mbuf chains.
yongari [Sat, 12 Aug 2006 01:24:37 +0000 (01:24 +0000)]
Fix invalid reference of mbuf chains.
Use proper pointer dereference to inform modified mbuf chains to
caller.

In collaboration with: glebius

17 years agoFix invalid reference of mbuf chains.
yongari [Sat, 12 Aug 2006 01:19:37 +0000 (01:19 +0000)]
Fix invalid reference of mbuf chains.
Use proper pointer dereference to inform modified mbuf chains to
caller.

While I'm here perform checksum offload setup after loading DMA
maps as m_defrag(9) can return new mbuf chains.

In collaboration with: glebius

17 years agoRemove extraneous trailing \0 in string literal.
thomas [Fri, 11 Aug 2006 23:03:16 +0000 (23:03 +0000)]
Remove extraneous trailing \0 in string literal.

MFC after: 2 weeks

17 years agoBefore performing a sodealloc() when pru_attach() fails, assert that
rwatson [Fri, 11 Aug 2006 23:03:10 +0000 (23:03 +0000)]
Before performing a sodealloc() when pru_attach() fails, assert that
the socket refcount remains 1, and then drop to 0 before freeing the
socket.

PR: 101763
Reported by: Gleb Kozyrev <gkozyrev at ukr dot net>

17 years agoThe groff_mm(7) locale files are *not* obsolete; they are just
ru [Fri, 11 Aug 2006 22:02:16 +0000 (22:02 +0000)]
The groff_mm(7) locale files are *not* obsolete; they are just
not touched during the installation if they already exists.

Noticed by: Dmitry Pryanishnikov

17 years ago(sm_mon_1_svc): Fix debugging output: when establising a monitoring
thomas [Fri, 11 Aug 2006 21:44:53 +0000 (21:44 +0000)]
(sm_mon_1_svc): Fix debugging output: when establising a monitoring
 request, correctly report the location (usually localhost) to which
 a callback will be made when a notification is received for the
 monitored host. Previsouly, the name of the monitored host was
 reported instead.

MFC after: 2 weeks

17 years agoUnbreak nForce4 SATA support.
sos [Fri, 11 Aug 2006 21:19:29 +0000 (21:19 +0000)]
Unbreak nForce4 SATA support.
Hopefully I dont break something else this time.....

17 years agoFixes an edge case bug in timewait handling where ticks rolling over causing
mohans [Fri, 11 Aug 2006 21:15:23 +0000 (21:15 +0000)]
Fixes an edge case bug in timewait handling where ticks rolling over causing
the timewait expiry to be exactly 0 corrupts the timewait queues (and that entry).
Reviewed by: silby

17 years agoFirst pass at allowing memory to be mapped using cache modes other than
jhb [Fri, 11 Aug 2006 19:22:57 +0000 (19:22 +0000)]
First pass at allowing memory to be mapped using cache modes other than
WB (write-back) on x86 via control bits in PTEs and PDEs (including making
use of the PAT MSR).  Changes include:
- A new pmap_mapdev_attr() function for amd64 and i386 which takes an
  additional parameter (relative to pmap_mapdev()) specifying the cache
  mode for this mapping.  Note that on amd64 only WB mappings are done with
  the direct map, all other modes result in a private mapping.
- pmap_mapdev() on i386 and amd64 now defaults to using UC (uncached)
  mappings rather than WB.  Previously we relied on the BIOS setting up
  MTRR's to enforce memio regions being treated as UC.  This might make
  hw.cbb_start_memory unnecessary in some cases now for example.
- A new pmap_mapbios()/pmap_unmapbios() API has been added to allow places
  that used pmap_mapdev() to map non-device memory (such as ACPI tables)
  to do so using WB as before.
- A new pmap_change_attr() function for amd64 and i386 that changes the
  caching mode for a range of KVA.

Reviewed by: alc

17 years agoExplicitly set v3 mode only when it is requested. Don't bother otherwise.
jkim [Fri, 11 Aug 2006 19:16:50 +0000 (19:16 +0000)]
Explicitly set v3 mode only when it is requested.  Don't bother otherwise.

17 years agoBefore using byte offset for IV creation, covert it to little endian.
pjd [Fri, 11 Aug 2006 19:09:12 +0000 (19:09 +0000)]
Before using byte offset for IV creation, covert it to little endian.
This way one will be able to use provider encrypted on eg. i386 on
eg. sparc64. This doesn't really buy us much today, because UFS isn't
endian agnostic.

We retain backward compatibility by setting G_ELI_FLAG_NATIVE_BYTE_ORDER
flag on devices with version number less than 2 and not converting the
offset.

17 years agoForgot to bump version number after G_ELI_FLAG_READONLY flag addition.
pjd [Fri, 11 Aug 2006 18:39:58 +0000 (18:39 +0000)]
Forgot to bump version number after G_ELI_FLAG_READONLY flag addition.

17 years agoReintroduce CSTD, which allows a Makefile to specify the precise dialect
des [Fri, 11 Aug 2006 17:28:59 +0000 (17:28 +0000)]
Reintroduce CSTD, which allows a Makefile to specify the precise dialect
of C in which the program or library is written.

Note that this is *not* intended to be used across the whole tree.  It
is intended to be used for individual libraries or programs which use
specific language features which the compiler must know about in order
to produce correct warnings at high WARNS levels.

MFC after: 1 month

17 years agoEnsure that the page's new field for object-synchronized flags is always
alc [Fri, 11 Aug 2006 17:18:58 +0000 (17:18 +0000)]
Ensure that the page's new field for object-synchronized flags is always
initialized to zero.

Call vm_page_sleep_if_busy() instead of duplicating its implementation in
vm_page_grab().

17 years agoCosmetic: fix indentation.
des [Fri, 11 Aug 2006 17:17:14 +0000 (17:17 +0000)]
Cosmetic: fix indentation.

17 years agoRaise the quality of the HPET timer to 2000 so it will be the preferred
des [Fri, 11 Aug 2006 17:12:16 +0000 (17:12 +0000)]
Raise the quality of the HPET timer to 2000 so it will be the preferred
choice on systems which support it.

No objection by: phk

17 years agoOptionally pad outgoing frames to the minimum of 60 bytes (excl. FCS)
yar [Fri, 11 Aug 2006 17:09:27 +0000 (17:09 +0000)]
Optionally pad outgoing frames to the minimum of 60 bytes (excl. FCS)
before tagging them.  This can help to work around brain-damage in some
switches that fail to pad a frame after untagging it if its length drops
below the minimum.  This option is blessed by IEEE Std 802.1Q (2003 Ed.),
paragraph C.4.4.3.b.  It's controlled by sysctl net.link.vlan.soft_pad.

Idea by: az
MFC after: 1 week

17 years agoAdditional debugging stuff I had in my tree.
des [Fri, 11 Aug 2006 17:03:33 +0000 (17:03 +0000)]
Additional debugging stuff I had in my tree.

17 years agoSpell 'connection options' correctly.
keramida [Fri, 11 Aug 2006 12:35:22 +0000 (12:35 +0000)]
Spell 'connection options' correctly.

Submitted by:   Leonidas Tsampros <ltsampros@upnet.gr>

17 years ago- em(4) updated to version 6.1.4
brueffer [Fri, 11 Aug 2006 12:32:43 +0000 (12:32 +0000)]
- em(4) updated to version 6.1.4
- stge(4) merged

17 years agoMerge in new driver from Intel, version 6.1.4. It adds support for
glebius [Fri, 11 Aug 2006 10:58:24 +0000 (10:58 +0000)]
Merge in new driver from Intel, version 6.1.4. It adds support for
82571EB quad port copper NIC and has few minor fixes.

Details:
  - if_em.c. Merged manually, viewing diff between new vendor
    driver and previous one.
  - if_em_hw.c. Dropped in from vendor, and then restored
    revision 1.15.

17 years agoAdd some more errno mappings (bsd -> linux) and a comment about the status..
netchild [Thu, 10 Aug 2006 22:05:25 +0000 (22:05 +0000)]
Add some more errno mappings (bsd -> linux) and a comment about the status..

Submitted by: "Intron" <mag@intron.ac>

17 years agoFox two typos.
netchild [Thu, 10 Aug 2006 20:03:52 +0000 (20:03 +0000)]
Fox two typos.

Noticed by: simon

17 years agoMFp4: Integrate a boatload of bug fixes from p4. We're right on the
imp [Thu, 10 Aug 2006 19:55:52 +0000 (19:55 +0000)]
MFp4: Integrate a boatload of bug fixes from p4.  We're right on the
8k boundary with this program still.

   text    data     bss     dec     hex filename
   7925       4    4476   12405    3075 bootiic.out

so we have like 293 bytes left before we have to play games.  There
may be ways to reduce that somewhat, but they start to be very board
specific.

17 years agoAdd some more files...
netchild [Thu, 10 Aug 2006 19:42:24 +0000 (19:42 +0000)]
Add some more files...

List provided by: kris

17 years agocheckpoint new version
sam [Thu, 10 Aug 2006 19:01:16 +0000 (19:01 +0000)]
checkpoint new version

17 years agoMFp4:
imp [Thu, 10 Aug 2006 18:22:56 +0000 (18:22 +0000)]
MFp4:

Reach over into the arm tree to grab some at91 definitions.
Better warnings
Lots of build tweaks

17 years agoMFp4: remove obsolete files
imp [Thu, 10 Aug 2006 18:19:45 +0000 (18:19 +0000)]
MFp4: remove obsolete files

17 years agoMFp4:
imp [Thu, 10 Aug 2006 18:14:35 +0000 (18:14 +0000)]
MFp4:
Remove obsolete files in list
Add spi flash reading routines

17 years agoMFp4: Numerous changes from p4 to try to improve tftp downloading,
imp [Thu, 10 Aug 2006 18:11:22 +0000 (18:11 +0000)]
MFp4: Numerous changes from p4 to try to improve tftp downloading,
reduce code footprint, etc.  While some problems still remain, the
reliability of tftp is much improved.

17 years agoMFp4: consolidate #include files down to lib.h
imp [Thu, 10 Aug 2006 18:07:49 +0000 (18:07 +0000)]
MFp4: consolidate #include files down to lib.h

17 years agoMFp4: These will migrate to the boot specific directories because they
imp [Thu, 10 Aug 2006 18:07:01 +0000 (18:07 +0000)]
MFp4: These will migrate to the boot specific directories because they
are too hard to share between the different boot loaders.

17 years agoMFp4: Make it clearer that the address passed to the eeprom routines is
imp [Thu, 10 Aug 2006 18:03:50 +0000 (18:03 +0000)]
MFp4: Make it clearer that the address passed to the eeprom routines is
really an offset within the eeprom device, and not a IIC address.

17 years agoMFp4: Catchup with migration of some defines.
imp [Thu, 10 Aug 2006 18:02:27 +0000 (18:02 +0000)]
MFp4: Catchup with migration of some defines.

17 years agoMFp4: Tweaks for our board.
imp [Thu, 10 Aug 2006 17:59:22 +0000 (17:59 +0000)]
MFp4: Tweaks for our board.

17 years agoMFp4: Increase character timeout to 10. Make it a #define for easier
imp [Thu, 10 Aug 2006 17:54:51 +0000 (17:54 +0000)]
MFp4: Increase character timeout to 10.  Make it a #define for easier
changes in the future.  This helps with getting started and to
overcome the really sucky level of granuality this timeout has in
getc.  A timeout of 1 means 'wait until top of next second' rather
than 'wait for at least a second'.

17 years agoAdd libcrypto.so.4 and libssl.so.4 to OLD_LIBS since they are obsoleted
simon [Thu, 10 Aug 2006 16:50:50 +0000 (16:50 +0000)]
Add libcrypto.so.4 and libssl.so.4 to OLD_LIBS since they are obsoleted
by the OpenSSL 0.9.8b import.

Reminded by: kris

17 years agoFix typo.
obrien [Thu, 10 Aug 2006 16:39:55 +0000 (16:39 +0000)]
Fix typo.

Submitted by: kevlo

17 years agoo add noise floor to stats
sam [Thu, 10 Aug 2006 16:31:37 +0000 (16:31 +0000)]
o add noise floor to stats
o include current tx rate in stats so athstats gets a consistent
  snapshot and doesn't have to make an extra ioctl
o record tx rate for raw frames

MFC after: 3 weeks

17 years agoImprove the way we'll detect video devices as per ACPI 3.0.
bruno [Thu, 10 Aug 2006 13:18:02 +0000 (13:18 +0000)]
Improve the way we'll detect video devices as per ACPI 3.0.

PR: 100271
Requested by: john AT utzweb DOT net
Submitted by: hrs
Reviewed by: njl
Approved by: njl
MFC after: 3 days

17 years agoSome perfectionizm against last revision.
glebius [Thu, 10 Aug 2006 11:07:11 +0000 (11:07 +0000)]
Some perfectionizm against last revision.

Submitted by: ru

17 years agoAdd support for splitting at gigabyte boundaries. [1]
keramida [Thu, 10 Aug 2006 10:41:47 +0000 (10:41 +0000)]
Add support for splitting at gigabyte boundaries. [1]

Also make both lowercase and uppercase suffix letters work
as byte-count suffixes, i.e. the following two commands are
equivalent now:

    % split -b 4m foo
    % split -b 4M foo

Submitted by: Roman Divacky [1]
Lots of help by: cperciva
Reviewed by: cperciva
MFC after: 1 week

17 years agogetnewvnode() can be called with NULL mp.
pjd [Thu, 10 Aug 2006 08:56:03 +0000 (08:56 +0000)]
getnewvnode() can be called with NULL mp.

Found by: Coverity Prevent (tm)
Coverity ID: 1521
Confirmed by: phk

17 years agoThe acpi module is no different from others. If it is loaded while the kernel
njl [Thu, 10 Aug 2006 06:32:16 +0000 (06:32 +0000)]
The acpi module is no different from others.  If it is loaded while the kernel
has it compiled in, it will be rejected with "module acpi already present!"
and no strange behavior.

MFC after: 1 day

17 years agoDon't need to special case arm here anymore
imp [Thu, 10 Aug 2006 06:29:43 +0000 (06:29 +0000)]
Don't need to special case arm here anymore

17 years agoAdd a dummy makefile to keep build happy
imp [Thu, 10 Aug 2006 06:29:15 +0000 (06:29 +0000)]
Add a dummy makefile to keep build happy

17 years agofixup list station support:
sam [Thu, 10 Aug 2006 06:06:13 +0000 (06:06 +0000)]
fixup list station support:
o add sanity check to avoid possible looping
o use intended api for IEEE80211_IOC_STA_INFO
o when operating in sta mode get the sta info for the ap

17 years agoMore statistics fixups:
sam [Thu, 10 Aug 2006 06:04:00 +0000 (06:04 +0000)]
More statistics fixups:
o change rssi to be signed in ieee80211_nodestats
o add noise floor in ieee80211_nodestats (use an implicit hole to
  preserve layout); return it as zero until we can update the api's
  so the driver can provide noise floor data
o add a bandaid so IEEE80211_IOC_STA_STATS works for sta mode; when
  all nodes are in the station table this will no longer be needed
o fix braino in IEEE80211_IOC_STA_INFO implementation; was supposed
  to take a mac address and return info for that sta or all stations
  if ff:ff:ff:ff:ff was supplied--but somehow this didn't get implemented;
  implement the intended semantics and leave a compat shim at the old
  ioctl number for the previous api

Reviewed by: mlaier
MFC after: 3 weeks

17 years agominor fixups:
sam [Thu, 10 Aug 2006 05:54:36 +0000 (05:54 +0000)]
minor fixups:
o add some missing stats to the global stat structure
o move accounting work for data frame rx into ieee80211_deliver_data
o add per-sta stats for rx ucast/mcast frames
o set rcvif in ieee80211_deliver_data so callers don't need to

MFC after: 2 weeks

17 years agoadd per-sta ucast/mcast stats
sam [Thu, 10 Aug 2006 05:37:49 +0000 (05:37 +0000)]
add per-sta ucast/mcast stats

MFC after: 1 week

17 years agoChange vm_page_cowfault() so that it doesn't allocate a pre-busied page.
alc [Thu, 10 Aug 2006 04:48:29 +0000 (04:48 +0000)]
Change vm_page_cowfault() so that it doesn't allocate a pre-busied page.

17 years agoo Spell.
maxim [Thu, 10 Aug 2006 01:13:38 +0000 (01:13 +0000)]
o Spell.

Submitted by: ru

17 years agoEliminate one set of XBOX #ifdefs. The Xbox code just needs to set a
imp [Wed, 9 Aug 2006 23:47:38 +0000 (23:47 +0000)]
Eliminate one set of XBOX #ifdefs.  The Xbox code just needs to set a
different TIMER_FREQ value than default.  Accomplish this via the
config file rather than via an #ifdef.

17 years agoMinor style(9) nit.
imp [Wed, 9 Aug 2006 23:37:30 +0000 (23:37 +0000)]
Minor style(9) nit.

17 years ago- Added ability to provide (optional) path to the kernel sources
rik [Wed, 9 Aug 2006 21:23:42 +0000 (21:23 +0000)]
- Added ability to provide (optional) path to the kernel sources
- Added check of dirs used by example generator
- Fixed path for ${s}/conf/files.FOO and include it via files
  directive from config file
- Changed kernel configuration example with the driver: it is
  not produced by copying Generic but by including it
- KDB is added to config (for DDB)
- Added module building instead and fixed kernel building

Reviewed by: julian@

17 years agoHook into the watchdog device, if present. Also, turn off the
imp [Wed, 9 Aug 2006 20:58:55 +0000 (20:58 +0000)]
Hook into the watchdog device, if present.  Also, turn off the
watchdog timer stuff when we boot because the boot blocks are turning
it on...

17 years agoAdded copyright templete
rik [Wed, 9 Aug 2006 20:57:24 +0000 (20:57 +0000)]
Added copyright templete
Fixed compilation

Reviewed by: julian@

17 years agoStrengthen the check for a PMBR:
marcel [Wed, 9 Aug 2006 20:53:01 +0000 (20:53 +0000)]
Strengthen the check for a PMBR:
o PMBR partitions count to the number of partitions on the disk, which
  means that if a PMBR entry is invalid we will not treat the MBR as a
  PMBR by virtue of it not describing any partitions.
  Previously the checks were inconsistent in that an invalid PMBR entry
  would be harmless when no other partitions exist (we would treat the
  MBR as a PMBR by virtue of it being empty), but it would be fatal when
  there is at least one other partition.
o The partition size of a PMBR partition is one less than the media size
  because the GPT starts at the second sector (LBA 1) and extends to
  the end of the media. For backward bug-compatibility we accept a size
  that's exactly the media size (FreeBSD bug).
  Also, when the partition size can not be represented in a 32-bit
  integral, the partition size in the MBR is to be set to 0xFFFFFFFF.
  Accept this as a valid size, even if the size can be represented.

17 years agoFix a bug in the size of the PMBR partition. Since the partition starts
marcel [Wed, 9 Aug 2006 20:19:15 +0000 (20:19 +0000)]
Fix a bug in the size of the PMBR partition. Since the partition starts
at LBA 1, the size is not the mediasize in sectors, but one less.

17 years ago10/100 PHY shouldn't support gigabit media types.
pdeuskar [Wed, 9 Aug 2006 20:10:35 +0000 (20:10 +0000)]
10/100 PHY shouldn't support gigabit media types.

Submitted by: brad (brad@comstyle.com)
Obtained from: OpenBSD
MFC after: 1 week

17 years agoo Strip eol whitespaces.
maxim [Wed, 9 Aug 2006 19:41:34 +0000 (19:41 +0000)]
o Strip eol whitespaces.

17 years agoo Simple strcpy/strcat replacement.
maxim [Wed, 9 Aug 2006 19:12:10 +0000 (19:12 +0000)]
o Simple strcpy/strcat replacement.

PR: bin/101575
Founded by: Dan Lukes
Obtained from: OpenBSD, rev. 1.11 by deraadt
MFC after: 2 weeks

17 years agoDefine NO_TLS on PowerPC.
marcel [Wed, 9 Aug 2006 19:01:27 +0000 (19:01 +0000)]
Define NO_TLS on PowerPC.

See also: PR ia64/91846

17 years agoo New sentence, new line.
maxim [Wed, 9 Aug 2006 18:35:31 +0000 (18:35 +0000)]
o New sentence, new line.
o Touch Dd for -r.

17 years agoAdd pc98 specific code to adjust the firmware geometry when it differs
imp [Wed, 9 Aug 2006 18:25:07 +0000 (18:25 +0000)]
Add pc98 specific code to adjust the firmware geometry when it differs
from the actual geometry.  This enables support of disks larger than
~120GB on pc98 boxes.  They make great little network appliances.
I've been using these changes for the past year or so on my network
storage pc98 box :-).

17 years agoMost platforms map the actual drive geometry to the firmware's notion
imp [Wed, 9 Aug 2006 18:23:47 +0000 (18:23 +0000)]
Most platforms map the actual drive geometry to the firmware's notion
of geometry.  However, some platforms have a more complicated mapping
of the firmware values to the actual values.  pc98 is the only
platform that currently does this.  This mapping is necessary for
large disks connected to pc98 boxes, as the firmware labels require do
special hacks to the actual geometry for interoperability.  We cannot
do this all in the geom layer because of initialization issues (geom
looks for an already initialized pc98 label, but we need the geometry
information prior to initialization, classic chicken and egg problem).
We pass the disk and the device_t to this function because the
geometry mapping depends on what kind of controller is used.

This hook allows platforms that want to override things to do so, and
has 0 overhead on all other platforms.  These patches have been in use
locally for a long time, and received good feedback from the pc98
community and sos@ at various times during their development.

MFC After: 1 week

17 years agoAllow geli to operate on read-only providers.
pjd [Wed, 9 Aug 2006 18:11:14 +0000 (18:11 +0000)]
Allow geli to operate on read-only providers.

Initial patch from: vd
MFC after: 2 weeks

17 years agoo It speaks about struct msqid_ds not struct shmid_ds.
maxim [Wed, 9 Aug 2006 18:05:44 +0000 (18:05 +0000)]
o It speaks about struct msqid_ds not struct shmid_ds.

PR: docs/101314
Submitted by: Vasil Dimov
MFC after: 1 week

17 years agoIntroduce a field to struct vm_page for storing flags that are
alc [Wed, 9 Aug 2006 17:43:27 +0000 (17:43 +0000)]
Introduce a field to struct vm_page for storing flags that are
synchronized by the lock on the object containing the page.

Transition PG_WANTED and PG_SWAPINPROG to use the new field,
eliminating the need for holding the page queues lock when setting
or clearing these flags.  Rename PG_WANTED and PG_SWAPINPROG to
VPO_WANTED and VPO_SWAPINPROG, respectively.

Eliminate the assertion that the page queues lock is held in
vm_page_io_finish().

Eliminate the acquisition and release of the page queues lock
around calls to vm_page_io_finish() in kern_sendfile() and
vfs_unbusy_pages().

17 years agoSince bpf_allocbufs() uses malloc() with M_WAITOK, don't check return
rwatson [Wed, 9 Aug 2006 16:30:26 +0000 (16:30 +0000)]
Since bpf_allocbufs() uses malloc() with M_WAITOK, don't check return
values for NULL or return an error state.  Assert that all three bpf
buffer pointers are NULL before starting.

MFC after: 1 week

17 years agoAdd a bandaid to avoid a deadlock in a situation, when we are trying to suspend
pjd [Wed, 9 Aug 2006 12:47:30 +0000 (12:47 +0000)]
Add a bandaid to avoid a deadlock in a situation, when we are trying to suspend
a file system, but need to obtain a vnode. We may not be able to do it, because
all vnodes could be already in use and other processes cannot release them,
because they are waiting in "suspfs" state.

In such situation, we allow to allocate a vnode anyway.

This is a temporary fix - there is no backpressure to free vnodes allocated in
those circumstances.

MFC after: 1 week
Reviewed by: tegge

17 years agoReadd sade, the problem should be fixed.
netchild [Wed, 9 Aug 2006 11:26:11 +0000 (11:26 +0000)]
Readd sade, the problem should be fixed.

17 years agoRemove the keymap stuff too, it is not needed in sade.
netchild [Wed, 9 Aug 2006 11:25:10 +0000 (11:25 +0000)]
Remove the keymap stuff too, it is not needed in sade.

17 years agoRemove sade from build-tools, the tinderbox problem should be fixed now.
netchild [Wed, 9 Aug 2006 11:03:06 +0000 (11:03 +0000)]
Remove sade from build-tools, the tinderbox problem should be fixed now.

17 years ago- Added 'files "files.FOO"' to config for inclusion of user specific files.
rik [Wed, 9 Aug 2006 10:53:26 +0000 (10:53 +0000)]
- Added 'files "files.FOO"' to config for inclusion of user specific files.
- Added KDB option to config file (for DDB).
- Added ability to compile custom kernel.

Discussed with: julian@

17 years agoFix ng_pppoe(4) after turning off "autosrc feature" on ng_ether(4).
glebius [Wed, 9 Aug 2006 09:56:58 +0000 (09:56 +0000)]
Fix ng_pppoe(4) after turning off "autosrc feature" on ng_ether(4).

- Store the Ethernet header in node softc.
- Initialize header with dst addr and ethertype in node
  constructor method.
- In node connect method send NGM_ETHER_GET_ENADDR message
  downwards.
- If received reply from ng_ether(4) store the src addr
  in softc.
- Add NGM_PPPOE_SETENDADDR message that allows user to
  override the address with whatever he/she wants.

17 years agoNot only a request from us can be passed to g_{mirror,raid3}_worker()
pjd [Wed, 9 Aug 2006 09:41:53 +0000 (09:41 +0000)]
Not only a request from us can be passed to g_{mirror,raid3}_worker()
function, but also a request to us, in which case checking bio_cflags
is wrong, because the class above us is controling it, not we.

MFC after: 1 week

17 years agoGet rid of the rtermcap stuff (embedding termcaps into the program). This
netchild [Wed, 9 Aug 2006 09:12:01 +0000 (09:12 +0000)]
Get rid of the rtermcap stuff (embedding termcaps into the program). This
is needed in sysinstall on boot, but not in sade on a fully working system.

After this change I noticed no change in behavior on a vty when TERM is not
set.

17 years agoBump WARNS level to 3.
delphij [Wed, 9 Aug 2006 08:25:21 +0000 (08:25 +0000)]
Bump WARNS level to 3.

Tested with: make universe

17 years agoUnbreak sparc64 build.
delphij [Wed, 9 Aug 2006 08:24:46 +0000 (08:24 +0000)]
Unbreak sparc64 build.

Reported by: tinderbox

17 years agoFix PC98 build.
delphij [Wed, 9 Aug 2006 05:33:45 +0000 (05:33 +0000)]
Fix PC98 build.

Spotted by: make universe

17 years agodisconnect sade until someone fixes tinderbox builds
sam [Wed, 9 Aug 2006 05:26:19 +0000 (05:26 +0000)]
disconnect sade until someone fixes tinderbox builds

17 years agoAdd the sade to build-tools list. This fixes a tinderbox.
nyan [Wed, 9 Aug 2006 02:27:31 +0000 (02:27 +0000)]
Add the sade to build-tools list.  This fixes a tinderbox.

17 years agoAdd a new kernel environment variable "boot.netif.mtu" which is used to
brooks [Wed, 9 Aug 2006 01:56:17 +0000 (01:56 +0000)]
Add a new kernel environment variable "boot.netif.mtu" which is used to
set the MTU prior to mounting root via NFS.  This is required if the
server supports a higher than default MTU because the client will not
see the responses otherwise.

MFC after: 3 weeks

17 years agoThomas Wintergerst reports that when this tsleep went away, certain
imp [Wed, 9 Aug 2006 00:05:54 +0000 (00:05 +0000)]
Thomas Wintergerst reports that when this tsleep went away, certain
cards stopped working.  Specifically the AVM B1 PCMCIA Card no longer
detected.  Its CIS chain read back as all FF's.  Putting the delay
back solves those problems.  I've opted to put in a much shorter delay
because as far as I can tell, no delay is really needed here.  We'll
see how well this works in practice.