]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
18 years agoSynchronize accesses to the PTEG table.
alc [Sun, 25 Jun 2006 19:07:01 +0000 (19:07 +0000)]
Synchronize accesses to the PTEG table.

Add many lock assertions.

Tested by: grehan@

18 years agoThe common UID/GID space implementation. It has been discussed on -arch
babkin [Sun, 25 Jun 2006 18:37:44 +0000 (18:37 +0000)]
The common UID/GID space implementation. It has been discussed on -arch
in 1999, and there are changes to the sysctl names compared to PR,
according to that discussion. The description is in sys/conf/NOTES.
Lines in the GENERIC files are added in commented-out form.
I'll attach the test script I've used to PR.

PR: kern/14584
Submitted by: babkin

18 years agoReplace the three copies of the list of pci ids with a single centralized list.
anholt [Sun, 25 Jun 2006 14:44:43 +0000 (14:44 +0000)]
Replace the three copies of the list of pci ids with a single centralized list.
Add the i945 PCI IDs commented out -- I think it should just work, but it hasn't
been tested yet.

18 years agoUse suser_cred() instead of a direct comparison of cr_uid with 0 in
rwatson [Sun, 25 Jun 2006 13:54:59 +0000 (13:54 +0000)]
Use suser_cred() instead of a direct comparison of cr_uid with 0 in
rip6_output().

MFC after: 1 week

18 years agoIf linker_release_module() fails then we still hold a reference on
iedowse [Sun, 25 Jun 2006 12:36:21 +0000 (12:36 +0000)]
If linker_release_module() fails then we still hold a reference on
the linker_file, so record this by restoring the linker_file pointer
in fp->file.

18 years agoNew release notes:
hrs [Sun, 25 Jun 2006 12:33:57 +0000 (12:33 +0000)]
New release notes:
apm(4) suspend/resume support improvement,
read(2) vn lock now protected by FOFFSET_LOCKED flag,
cardbus(4), pccard(4), pccbb(4), and exca(4) buildable as KLDs,
midi(4) driver added again,
ath(4) and related drivers included in GENERIC by default,
wi(4) builable as KLD,
wlan_*(4) crypto drivers included in GENERIC by default,
ipfw(4) tag/untag keyword support,
the default retransmit timer for NFS over TCP changed to 60s,
the default value of vfs.nfs.iodmin changed to 0,
numeric arguments for find(1) -user/-group now work as expected,
hccontrol(8) HCI node autodetection,
id(1) prints euid with gid,
kdump(1) can print syscall flags in symbol names,
"mount -u -o rw" on read-only mount now works as expected,
sed(1) incorrect calculation of pattern space length fixed,
time(1) SIGINFO support added,
rc.d/mdconfig added,
DRM updated to a snapshot as of 20060517,
/media directory for removable media mount points added,
pc98 release CDROMs now boolable, and
kernel subsystem API doc generating framework with doxygen added.

Reviewed in part by: bmah

18 years agoDecrement the global syncache counter in syncache_expand() when the entry
andre [Sun, 25 Jun 2006 11:11:33 +0000 (11:11 +0000)]
Decrement the global syncache counter in syncache_expand() when the entry
is removed from the bucket.  This fixes the syncache statistics.

18 years agoMajor Fixes:
mjacob [Sun, 25 Jun 2006 04:23:26 +0000 (04:23 +0000)]
Major Fixes:

  Don't enable/disable I/O space except for SAS adapters.
  This fixes a problem with VMware 4.5 Workstation.

  Fix an egregious bug introduced to target mode so it actually
  will not panic when you first enable a lun.

Minor fixes:

  Take more infor from port facts and configuration pages.

MFC after: 1 week

18 years agoSimplify the code and remove two mutex operations.
pjd [Sat, 24 Jun 2006 22:55:43 +0000 (22:55 +0000)]
Simplify the code and remove two mutex operations.

MFC after: 2 weeks

18 years agoUpdate to SDM 2.2:
marcel [Sat, 24 Jun 2006 19:21:11 +0000 (19:21 +0000)]
Update to SDM 2.2:
o  Add tf (test feature) instruction,
o  Add vmsw (VM switch) instruction.

While here, update copyright.

MFC after: 1 week

18 years agoremove duplicate ifname in debug msg
sam [Sat, 24 Jun 2006 18:00:17 +0000 (18:00 +0000)]
remove duplicate ifname in debug msg

MFC after: 1 week

18 years agoVMware disk volumes are only on LUN 0, and
mjacob [Sat, 24 Jun 2006 16:40:59 +0000 (16:40 +0000)]
VMware disk volumes are only on LUN 0, and
some VMware HBAs have known bugs with commands
sent to other than LUN 0.

MFC after: 1 day

18 years agoSync up with SDM 2.1:
marcel [Sat, 24 Jun 2006 01:19:52 +0000 (01:19 +0000)]
Sync up with SDM 2.1:
o  Add nop/hint formats F16, I18, M48 and X5,
o  Add format M47 for ptc.e,
o  Add hint instruction,
o  Fix decoding of cmp8xchg16.

18 years agoBackout previous commit, Warner committed at91_pio.c...
cognet [Fri, 23 Jun 2006 23:07:11 +0000 (23:07 +0000)]
Backout previous commit, Warner committed at91_pio.c...

18 years agoThere's no need to allocate that much phdr/shdr from the stack.
cognet [Fri, 23 Jun 2006 22:45:35 +0000 (22:45 +0000)]
There's no need to allocate that much phdr/shdr from the stack.

18 years agoAdd the arm9_setup() prototype.
cognet [Fri, 23 Jun 2006 22:37:15 +0000 (22:37 +0000)]
Add the arm9_setup() prototype.

18 years agoComment out at91_pio.c, it's not in CVS.
cognet [Fri, 23 Jun 2006 22:30:55 +0000 (22:30 +0000)]
Comment out at91_pio.c, it's not in CVS.

18 years agoFix the name of the data set item for the SYSUNINIT in RW_SYSINIT to use
jhb [Fri, 23 Jun 2006 19:36:50 +0000 (19:36 +0000)]
Fix the name of the data set item for the SYSUNINIT in RW_SYSINIT to use
'rw' instead of 'mtx'.  This should only be a cosmetic change rather than
a functional one.

Submitted by: Alex Lyashkov <shadow AT itt dot net dot ru>

18 years agoThe linux times syscall can be called with a NULL pointer, so keep cool
netchild [Fri, 23 Jun 2006 18:49:38 +0000 (18:49 +0000)]
The linux times syscall can be called with a NULL pointer, so keep cool
and don't panic.

This fix is different from the patch submitted as it not only prevents
a NULL-pointer dereference, but also skips some work in this case.

Noticed by: Dmitry Ganenko <dima@apk-inform.com>
Reviewed by: rdivacky (the original version as in emulation@)
MFC after: 1 week
Security: This is a RELENG_x_y candidate (local DoS).
Go ahead by: secteam (cperciva)

18 years agodump new cabq stats
sam [Fri, 23 Jun 2006 18:48:44 +0000 (18:48 +0000)]
dump new cabq stats

18 years agonew stats
sam [Fri, 23 Jun 2006 18:48:08 +0000 (18:48 +0000)]
new stats

MFC after: 1 month

18 years agoremove display of the ERP ie from the list sta output (it's always
sam [Fri, 23 Jun 2006 17:22:03 +0000 (17:22 +0000)]
remove display of the ERP ie from the list sta output (it's always
zero); replace it with station capabilities

MFC after: 1 month

18 years agoMove the code to handle the vm.blacklist tunable up a layer into
jhb [Fri, 23 Jun 2006 16:44:24 +0000 (16:44 +0000)]
Move the code to handle the vm.blacklist tunable up a layer into
vm_page_startup().  As a result, we now only lookup the tunable once
instead of looking it up once for every physical page of memory in the
system.  This cuts out about a 1 second or so delay in boot on x86
systems.  The delay is much larger and more noticable on sun4v apparently.

Reported by: kmacy
MFC after: 1 week

18 years agotwalk() expects an `action' function not a comparison function.
keramida [Fri, 23 Jun 2006 13:36:33 +0000 (13:36 +0000)]
twalk() expects an `action' function not a comparison function.
The text is correct in the "DESCRIPTION" section, so fix "SYNOPSIS"
to use the correct name.

PR: docs/90498
Submitted by: Vasil Dimov
MFC after: 3 days

18 years agoRemove some unused variables
brian [Fri, 23 Jun 2006 01:42:03 +0000 (01:42 +0000)]
Remove some unused variables

18 years agoSize the NFS server dupreq cache on the basis of nmbclusters. On servers
mohans [Fri, 23 Jun 2006 00:42:26 +0000 (00:42 +0000)]
Size the NFS server dupreq cache on the basis of nmbclusters. On servers
with low nmbclusters, we tie up too many mbclusters in the NFS duplicate
request cache. This change limits the size of the dupreq cache to 1/2
the nmbclusters (and flaots in a range of [64, 2048]).

MFC after 2 weeks.

Reported by: Steve Kargl, David O'Brien
Tested by:   Steve Kargl

18 years agoarm9_setup() is now needed even if we're not using a gzipped kernel, so move
cognet [Thu, 22 Jun 2006 22:33:21 +0000 (22:33 +0000)]
arm9_setup() is now needed even if we're not using a gzipped kernel, so move
it outside the #ifdef KZIP

Pointy Hat to: cognet

18 years agoImprove support for Intel based Macs:
marcel [Thu, 22 Jun 2006 22:22:32 +0000 (22:22 +0000)]
Improve support for Intel based Macs:
o  Accept hfs as partition type.
o  Print Apple HFS partitions using a friendly name.

18 years agoAdd the UUID of Apple's HFS file system as can be found in the Intel
marcel [Thu, 22 Jun 2006 22:11:12 +0000 (22:11 +0000)]
Add the UUID of Apple's HFS file system as can be found in the Intel
based Macs.

18 years agoMove the duplicated logic of parsing partition types into a new
marcel [Thu, 22 Jun 2006 22:05:28 +0000 (22:05 +0000)]
Move the duplicated logic of parsing partition types into a new
function called parse_uuid().

18 years ago- Convert to new SYNOPSIS layout
brueffer [Thu, 22 Jun 2006 20:03:35 +0000 (20:03 +0000)]
- Convert to new SYNOPSIS layout
- New Sentence, new line
- Remove eol whitespace

MFC after: 3 days

18 years agoPer weak consensus on this topic, remove suggestion to set
imp [Thu, 22 Jun 2006 16:52:53 +0000 (16:52 +0000)]
Per weak consensus on this topic, remove suggestion to set
HISTORICAL_MAKE_WORLD from the text that's output.  This was committed
against the previous consensus.  Leave the documentation in this file
as a compromose.  The HISTORICAL_MAKE_WORLD knob is intentionally
obfuscated and we only trust people smart enough to read the Makefile
to use it.  All others have no business using it due to its danger,
unless DESTDIR is set.

Dissentors: grog, obrien, trhodes

18 years agoMove the syncookie MD5 context from globals to the stack to make it MP safe.
andre [Thu, 22 Jun 2006 15:07:45 +0000 (15:07 +0000)]
Move the syncookie MD5 context from globals to the stack to make it MP safe.

18 years agoAdd myself to the freebsd calendar.
bsam [Thu, 22 Jun 2006 14:22:20 +0000 (14:22 +0000)]
Add myself to the freebsd calendar.

Approved by: netchild (mentor)

18 years ago- Pullup even when the extention header is unknown, to prevent
ume [Thu, 22 Jun 2006 13:22:54 +0000 (13:22 +0000)]
- Pullup even when the extention header is unknown, to prevent
  infinite loop with net.inet6.ip6.fw.deny_unknown_exthdrs=0.
- Teach ipv6 and ipencap as they appear in an IPv4/IPv6 over IPv6
  tunnel.
- Test the next extention header even when the routing header type
  is unknown with net.inet6.ip6.fw.deny_unknown_exthdrs=0.

Found by: xcast-fan-club
MFC after: 1 week

18 years agoMove conditional preprocessing out of the SYSCTL_ADD_STRING macro
dds [Thu, 22 Jun 2006 13:11:36 +0000 (13:11 +0000)]
Move conditional preprocessing out of the SYSCTL_ADD_STRING macro
invocation.  Per C99 6.10.3 paragraph 11 preprocessing directives
appearing inside macro arguments yield undefined behavior.

18 years agoClean up style and some printf messages. Note that returning anything other
njl [Thu, 22 Jun 2006 06:34:05 +0000 (06:34 +0000)]
Clean up style and some printf messages.  Note that returning anything other
than 0 for probe seems to cause a panic somewhere in sysctl kern.

18 years agoIdentify the cual-core Montecito.
marcel [Thu, 22 Jun 2006 00:56:58 +0000 (00:56 +0000)]
Identify the cual-core Montecito.

MFC after: 3 days

18 years agoNitsville: the routine is called initarm, not init_arm, correct it in
imp [Wed, 21 Jun 2006 23:47:25 +0000 (23:47 +0000)]
Nitsville: the routine is called initarm, not init_arm, correct it in
a comment.

18 years agoReplace the kld_mtx mutex with a kld_sx sx lock and expand it's scope to
jhb [Wed, 21 Jun 2006 20:42:08 +0000 (20:42 +0000)]
Replace the kld_mtx mutex with a kld_sx sx lock and expand it's scope to
protect all linker-related data structures including the contents of
linker file objects and the any linker class data as well. Considering how
rarely the linker is used I just went with the simple solution of
single-threading the whole thing rather than expending a lot of effor on
something more fine-grained and complex.  Giant is still explicitly
acquired while registering and deregistering sysctl's as well as in the
elf linker class while calling kmupetext().  The rest of the linker runs
without Giant unless it has to acquire Giant while loading files from a
non-MPSAFE filesystem.

18 years agoAdd a sx_xlocked() macro which returns true if the current thread holds an
jhb [Wed, 21 Jun 2006 20:38:29 +0000 (20:38 +0000)]
Add a sx_xlocked() macro which returns true if the current thread holds an
exclusive lock on the specified sx lock.

18 years ago- Push down Giant in kldfind() and kldsym().
jhb [Wed, 21 Jun 2006 20:15:36 +0000 (20:15 +0000)]
- Push down Giant in kldfind() and kldsym().
- Remove several goto's by either using direct return's or else clauses.

18 years agoWhoops, revert accidental commit.
jhb [Wed, 21 Jun 2006 17:48:59 +0000 (17:48 +0000)]
Whoops, revert accidental commit.

18 years agoFix two comments and a style fix.
jhb [Wed, 21 Jun 2006 17:48:03 +0000 (17:48 +0000)]
Fix two comments and a style fix.

18 years agoVarious whitespace fixes.
jhb [Wed, 21 Jun 2006 17:47:45 +0000 (17:47 +0000)]
Various whitespace fixes.

18 years agoFix the VLAN_ARRAY case, mostly regarding improper use of atomic(9)
yar [Wed, 21 Jun 2006 13:48:34 +0000 (13:48 +0000)]
Fix the VLAN_ARRAY case, mostly regarding improper use of atomic(9)
in place of conventional rw locking.  Alas, atomic(9) can't buy us
lockless operation so easily.

18 years agoMake the mincore(2) return ENOMEM when requested range is not fully mapped.
kib [Wed, 21 Jun 2006 12:59:05 +0000 (12:59 +0000)]
Make the mincore(2) return ENOMEM when requested range is not fully mapped.

Requested by: Bruno Haible <bruno at clisp org>
Reviewed by: alc
Approved by: pjd (mentor)
MFC after: 1 month

18 years agoDocument the fact that 'true' and 'false' are among sh(1) built-in commands.
yar [Wed, 21 Jun 2006 12:01:52 +0000 (12:01 +0000)]
Document the fact that 'true' and 'false' are among sh(1) built-in commands.

MFC after: 3 days

18 years agoDon't forget to touch Dd.
yar [Wed, 21 Jun 2006 11:46:29 +0000 (11:46 +0000)]
Don't forget to touch Dd.

18 years agoDon't forget to define uart_sa1110_vaddr.
cognet [Wed, 21 Jun 2006 10:56:59 +0000 (10:56 +0000)]
Don't forget to define uart_sa1110_vaddr.

Submitted by: kevlo

18 years agoSince INSTALLS_SHLIB has been superseded by USE_LDCONFIG in bsd.port.mk, I
flz [Wed, 21 Jun 2006 10:22:44 +0000 (10:22 +0000)]
Since INSTALLS_SHLIB has been superseded by USE_LDCONFIG in bsd.port.mk, I
expect to see quite a few files appearing in libdata/ldconfig directories.
This change avoids the screen to be filled with the names of those ldconfig
files and replace them by the actual non-default directories they contain.
Most of them will be ${PREFIX}/lib so, 'sort -u' will help reducing the
output.

Approved by: cperciva (implicit)
MFC after: 1 week

18 years agoNote a short-cut in the build procedure.
obrien [Wed, 21 Jun 2006 09:53:42 +0000 (09:53 +0000)]
Note a short-cut in the build procedure.

18 years agoSet an example of using load_kld() from rc.subr.
yar [Wed, 21 Jun 2006 09:53:25 +0000 (09:53 +0000)]
Set an example of using load_kld() from rc.subr.

18 years agoFix type warnings on 64-bit machines.
obrien [Wed, 21 Jun 2006 09:52:37 +0000 (09:52 +0000)]
Fix type warnings on 64-bit machines.

18 years agoQuite a number of rc.d scripts try to load kernel modules. Many
yar [Wed, 21 Jun 2006 09:42:55 +0000 (09:42 +0000)]
Quite a number of rc.d scripts try to load kernel modules.  Many
of them do that conditionally depending on kldstat.  The code is
duplicated all over, but bugs can be uniqie.

To make the things more consistent, introduce a new rc.subr function,
load_kld, which takes care of loading a kernel module conditionally.

(Found this lying for a while in my p4 branch for various hacks.)

18 years agoThis commit was generated by cvs2svn to compensate for changes in r159825,
obrien [Wed, 21 Jun 2006 09:28:00 +0000 (09:28 +0000)]
This commit was generated by cvs2svn to compensate for changes in r159825,
which included commits to RCS files with non-trunk default branches.

18 years agoFix segfault when '/etc/malloc.conf@ -> AJ' due to differnce between
obrien [Wed, 21 Jun 2006 09:28:00 +0000 (09:28 +0000)]
Fix segfault when '/etc/malloc.conf@ -> AJ' due to differnce between
calloc & malloc as agreed by vendor.  Also remove an unused variable.

18 years agoCommit the DUMMY stuff (printing messages for missing syscalls) for amd64 too.
netchild [Wed, 21 Jun 2006 08:45:40 +0000 (08:45 +0000)]
Commit the DUMMY stuff (printing messages for missing syscalls) for amd64 too.

Submitted by: rdivacky
Sponsored by: Google SoC 2006
Noticed by: jkim
Pointyhat to: netchild

18 years agoTrack interface department events and detach vlans from
yar [Wed, 21 Jun 2006 07:29:44 +0000 (07:29 +0000)]
Track interface department events and detach vlans from
departing trunk so that we don't get into trouble later
by dereferencing a stale pointer to dead trunk's things.

Prodded by: oleg
Sponsored by: RiNet (Cronyx Plus LLC)
MFC after: 1 week

18 years ago- First initialize ifnet, and then insert it into global
glebius [Wed, 21 Jun 2006 06:02:35 +0000 (06:02 +0000)]
- First initialize ifnet, and then insert it into global
  list.
- First remove from global list, then start destroying.

PR: kern/97679
Submitted by: Alex Lyashkov <shadow itt.net.ru>
Reviewed by: rwatson, brooks

18 years agoCompute physmem so we can print it correctly on boot.
imp [Tue, 20 Jun 2006 23:40:04 +0000 (23:40 +0000)]
Compute physmem so we can print it correctly on boot.
Slightly optimize while I'm here.

18 years agoFix a potential problem when mfi_get_log_state and only
ps [Tue, 20 Jun 2006 23:08:35 +0000 (23:08 +0000)]
Fix a potential problem when mfi_get_log_state and only
release a command if one was allocated.
Also release the command in mfi_shutdown.

18 years agoInstead of using scsi probes to do device discovery, use the firmware
ps [Tue, 20 Jun 2006 22:41:44 +0000 (22:41 +0000)]
Instead of using scsi probes to do device discovery, use the firmware
commands to grab the device listing.  This resolves issues using
multiple volumes, where each volume was actually internally pointing
to target 0.

18 years agoFix a typo when getting the log_state and set the flags on the
ps [Tue, 20 Jun 2006 22:17:54 +0000 (22:17 +0000)]
Fix a typo when getting the log_state and set the flags on the
command to signify it is being polled and expecting data to be dma'd
from the card.

18 years agoo Remove LK_REENABLE and LK_NOPAUSE lockinit(9) flags, add LK_NOSHARE.
maxim [Tue, 20 Jun 2006 21:41:14 +0000 (21:41 +0000)]
o Remove LK_REENABLE and LK_NOPAUSE lockinit(9) flags, add LK_NOSHARE.

MFC after: 1 week

18 years agoConditionally acquire Giant around VFS operations.
jhb [Tue, 20 Jun 2006 21:31:38 +0000 (21:31 +0000)]
Conditionally acquire Giant around VFS operations.

18 years agoAllow gif interfaces to be added as span ports, the user may want to send a
thompsa [Tue, 20 Jun 2006 21:28:18 +0000 (21:28 +0000)]
Allow gif interfaces to be added as span ports, the user may want to send a
copy of all packets to the other side of the world.

18 years agoCleanup dcmd firmware processing into a single function, mfi_dcmd_command
ps [Tue, 20 Jun 2006 21:06:05 +0000 (21:06 +0000)]
Cleanup dcmd firmware processing into a single function, mfi_dcmd_command
to avoid duplication and mistakes when setting up firmware commands
for submission.

Reviewed by: scottl

18 years agoAdd a new section in this file for functions that are only exported by the
jhb [Tue, 20 Jun 2006 20:59:55 +0000 (20:59 +0000)]
Add a new section in this file for functions that are only exported by the
linker for use in the linker class handlers.  Move linker_add_class(),
linker_file_unload(), linker_load_dependencies(), and linker_make_file()
into this section.

18 years ago- Push Giant down into linker_reference_module().
jhb [Tue, 20 Jun 2006 20:54:13 +0000 (20:54 +0000)]
- Push Giant down into linker_reference_module().
- Add a new function linker_release_module() as a more intuitive complement
  to linker_reference_module() that wraps linker_file_unload().
  linker_release_module() can either take the module name and version info
  passed to linker_reference_module() or it can accept the linker file
  object returned by linker_reference_module().

18 years agoChange get_pv_entry() such that the call to vm_page_alloc() specifies
alc [Tue, 20 Jun 2006 20:52:11 +0000 (20:52 +0000)]
Change get_pv_entry() such that the call to vm_page_alloc() specifies
VM_ALLOC_NORMAL instead of VM_ALLOC_SYSTEM when try is TRUE.  In other
words, when get_pv_entry() is permitted to fail, it no longer tries as
hard to allocate a page.

Change pmap_enter_quick_locked() to fail rather than wait if it is
unable to allocate a page table page.  This prevents a race between
pmap_enter_object() and the page daemon.  Specifically, an inactive
page that is a successor to the page that was given to
pmap_enter_quick_locked() might become a cache page while
pmap_enter_quick_locked() waits and later pmap_enter_object() maps
the cache page violating the invariant that cache pages are never
mapped.  Similarly, change
pmap_enter_quick_locked() to call pmap_try_insert_pv_entry() rather
than pmap_insert_entry().  Generally speaking,
pmap_enter_quick_locked() is used to create speculative mappings.  So,
it should not try hard to allocate memory if free memory is scarce.

Add an assertion that the object containing m_start is locked in
pmap_enter_object().  Remove a similar assertion from
pmap_enter_quick_locked() because that function no longer accesses the
containing object.

Remove a stale comment.

Reviewed by: ups@

18 years agoClean up some comments and device printf failures in mfi_get_log_state
ps [Tue, 20 Jun 2006 20:41:54 +0000 (20:41 +0000)]
Clean up some comments and device printf failures in mfi_get_log_state
and mfi_get_controller_info.

18 years agoregen after change to syscalls.master
netchild [Tue, 20 Jun 2006 20:41:29 +0000 (20:41 +0000)]
regen after change to syscalls.master

18 years agoMake linker_find_file_by_name() and linker_find_file_by_id() static to
jhb [Tue, 20 Jun 2006 20:41:15 +0000 (20:41 +0000)]
Make linker_find_file_by_name() and linker_find_file_by_id() static to
simplify linker locking.  The only external consumers now use
linker_file_foreach().

18 years agoSwitch to using the DUMMY infrastructure instead of UNIMPL for the new
netchild [Tue, 20 Jun 2006 20:38:44 +0000 (20:38 +0000)]
Switch to using the DUMMY infrastructure instead of UNIMPL for the new
syscalls. This way there will be a log message printed to the console
(this time for real).

Note: UNIMPL should be used for syscalls we do not implement ever, e.g.
syscalls to load linux kernel modules.

Submitted by: rdivacky
Sponsored by: Goole SoC 2006
P4 IDs: 99600, 99602

18 years agoAdd a missing case for the switch statement in arena_run_reg_dalloc(). [1]
jasone [Tue, 20 Jun 2006 20:38:25 +0000 (20:38 +0000)]
Add a missing case for the switch statement in arena_run_reg_dalloc(). [1]

Fix a leak in chunk_dealloc(). [2]

Reported by: [1] djam8193ah@hotmail.com,
[2] Ville-Pertti Keinonen <will@exomi.com>

18 years ago- Add a new linker_file_foreach() function that walks the list of linker
jhb [Tue, 20 Jun 2006 20:37:17 +0000 (20:37 +0000)]
- Add a new linker_file_foreach() function that walks the list of linker
  file objects calling a user-specified predicate function on each object.
  The iteration terminates either when the entire list has been iterated
  over or the predicate function returns a non-zero value.
  linker_file_foreach() returns the value returned by the last invocation
  of the predicate function.  It also accepts a void * context pointer that
  is passed to the predicate function as well.  Using an iterator function
  avoids exposing linker internals to the rest of the kernel making locking
  simpler.
- Use linker_file_foreach() instead of walking the list of linker files
  manually to lookup ndis files in ndis(4).
- Use linker_file_foreach() to implement linker_hwpmc_list_objects().

18 years agoMake linker_file_add_dependency() and linker_load_module() static since
jhb [Tue, 20 Jun 2006 20:18:42 +0000 (20:18 +0000)]
Make linker_file_add_dependency() and linker_load_module() static since
only the linker uses them.

18 years agoProbe the memory size of the board better. Look at the bus width,
imp [Tue, 20 Jun 2006 20:13:40 +0000 (20:13 +0000)]
Probe the memory size of the board better.  Look at the bus width,
number of banks, rows and columns the SDRAMC is programmed to access
to determine the RAM size for the board, rather than hard-wiring it to
be 32MB.  My company's board with 64MB now probes correctly, as does
the KB9202 with only 32MB.  This means that to detect the right memory
size, our boot loader must correctly initialize these values.  This is
a fairly safe assumption because the boot loader has to initialize
SDRAM already, and it isn't really possible to change this register
after we've accessed SDRAM.

18 years agoDon't check if malloc(M_WAITOK) returns NULL.
jhb [Tue, 20 Jun 2006 20:11:00 +0000 (20:11 +0000)]
Don't check if malloc(M_WAITOK) returns NULL.

18 years agoUse 'else' to remove another goto.
jhb [Tue, 20 Jun 2006 19:49:28 +0000 (19:49 +0000)]
Use 'else' to remove another goto.

18 years ago- Remove some useless variable initializations.
jhb [Tue, 20 Jun 2006 19:32:10 +0000 (19:32 +0000)]
- Remove some useless variable initializations.
- Make some conditional free()'s where the condition was always true
  unconditional.

18 years agoWe no longer need to disable interrupts in MD trap machinery
yar [Tue, 20 Jun 2006 12:44:21 +0000 (12:44 +0000)]
We no longer need to disable interrupts in MD trap machinery
when we're about to call kdb_trap() because the latter MI
function can disable interrupts by itself now.

Pointed out by: bde
X-MFC remark: depends on kern/subr_kdb.c#1.18
Sponsored by: RiNet (Cronyx Plus LLC)

18 years agoProperly cast the values of valsize (the size of the value passed in)
gnn [Tue, 20 Jun 2006 12:36:40 +0000 (12:36 +0000)]
Properly cast the values of valsize (the size of the value passed in)
in setsockopt so that they can be compared correctly against negative
values.  Passing in a negative value had a rather negative effect
on our socket code, making it impossible to open new sockets.

PR: 98858
Submitted by: James.Juran@baesystems.com
MFC after: 1 week

18 years ago+To create quality software, the ability to say no is usually far
phk [Tue, 20 Jun 2006 08:34:36 +0000 (08:34 +0000)]
+To create quality software, the ability to say no is usually far
+more important than the ability to say yes.
+               -- Michi Henning
+%

18 years agoAllocate a zero'ed syncache hashtable. mtx_init() tests the supplied
andre [Tue, 20 Jun 2006 08:11:30 +0000 (08:11 +0000)]
Allocate a zero'ed syncache hashtable.  mtx_init() tests the supplied
memory location for already existing/initialized mutexes.  With random
data in the memory location this fails (ie. after a soft reboot).

Reported by: brueffer, YAMAMOTO Shigeru
Submitted by: YAMAMOTO Shigeru <shigeru-at-iij.ad.jp>

18 years agoDocument that 'bad cksum 0' is expected on NICs with checksum off-loading.
obrien [Tue, 20 Jun 2006 05:04:05 +0000 (05:04 +0000)]
Document that 'bad cksum 0' is expected on NICs with checksum off-loading.

18 years agoipw statistics tool
sam [Tue, 20 Jun 2006 00:26:35 +0000 (00:26 +0000)]
ipw statistics tool

18 years agoAdd variable cpu_mxcsr_mask to save valid bits of mxcsr register.
davidxu [Mon, 19 Jun 2006 22:59:28 +0000 (22:59 +0000)]
Add variable cpu_mxcsr_mask to save valid bits of mxcsr register.

18 years agoMFi386:
davidxu [Mon, 19 Jun 2006 22:36:01 +0000 (22:36 +0000)]
MFi386:
Use the method described in IA-32 Intel Architecture Software
Developer's Manual chapter 11.6.6 to get valid mxcsr bits,
use the mxcsr mask to clear invalid bits passed by user code.

18 years agoImport interface groups from OpenBSD. This allows to group interfaces in
mlaier [Mon, 19 Jun 2006 22:20:45 +0000 (22:20 +0000)]
Import interface groups from OpenBSD.  This allows to group interfaces in
order to - for example - apply firewall rules to a whole group of
interfaces.  This is required for importing pf from OpenBSD 3.9

Obtained from: OpenBSD (with changes)
Discussed on: -net (back in April)

18 years ago- Move the gigabit NICs that use miibus into the miibus section to match
jhb [Mon, 19 Jun 2006 22:11:44 +0000 (22:11 +0000)]
- Move the gigabit NICs that use miibus into the miibus section to match
  GENERIC.
- Add bce(4) and bfe(4).

18 years agoThere is a plenty of arch's out there where sizeof(size_t) != sizeof(int).
yar [Mon, 19 Jun 2006 17:12:25 +0000 (17:12 +0000)]
There is a plenty of arch's out there where sizeof(size_t) != sizeof(int).
A field width in printf(3) must be int, so cast return value from strlen()
to the type.

Noticed by: Andrzej Tobola <ato iem pw edu pl>; tinderbox
Pointy hat to: yar
X-MFC with: the rest of the new asf code

18 years agoAdd a forgotten ","
matteo [Mon, 19 Jun 2006 16:31:58 +0000 (16:31 +0000)]
Add a forgotten ","

18 years agoAdd Hamlet eXagerate XURS232 USB-Serial adapter
matteo [Mon, 19 Jun 2006 16:30:35 +0000 (16:30 +0000)]
Add Hamlet eXagerate XURS232 USB-Serial adapter

PR: usb/98983
Submitted by: Massimo Lusetti
MFC after: 1 week

18 years agoBoji Tony Kannanthanam no longer supports this driver, Achim Leubner
brueffer [Mon, 19 Jun 2006 15:22:34 +0000 (15:22 +0000)]
Boji Tony Kannanthanam no longer supports this driver, Achim Leubner
however still does.  Confirmed by email contact with both of them.

MFC after: 3 days

18 years agoWhen we receive an out-of-window SYN for an "ESTABLISHED" connection,
dwmalone [Mon, 19 Jun 2006 12:33:52 +0000 (12:33 +0000)]
When we receive an out-of-window SYN for an "ESTABLISHED" connection,
ACK the SYN as required by RFC793, rather than ignoring it. NetBSD
have had a similar change since 1999.

PR: 93236
Submitted by: Grant Edwards <grante@visi.com>
MFC after: 1 month

18 years agoFix type casts so that we get at the right response queue.
sos [Mon, 19 Jun 2006 11:47:57 +0000 (11:47 +0000)]
Fix type casts so that we get at the right response queue.

Submitted by: glebius+ru

18 years agoTrival whitespace change.
obrien [Mon, 19 Jun 2006 11:30:36 +0000 (11:30 +0000)]
Trival whitespace change.

18 years agoUpdate for the 'file' 4.17 import.
obrien [Mon, 19 Jun 2006 08:10:23 +0000 (08:10 +0000)]
Update for the 'file' 4.17 import.