]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
17 years agoGrammar nits.
ceri [Thu, 15 Feb 2007 21:49:32 +0000 (21:49 +0000)]
Grammar nits.

17 years agoPut one $FreeBSD$ keyword in standard location, bump copyright date.
bmah [Thu, 15 Feb 2007 20:14:59 +0000 (20:14 +0000)]
Put one $FreeBSD$ keyword in standard location, bump copyright date.

17 years agoAdd supported processors and motherboards sections from the MD
bmah [Thu, 15 Feb 2007 20:12:43 +0000 (20:12 +0000)]
Add supported processors and motherboards sections from the MD
hardware notes documents in  */proc-*.sgml.

17 years agoAdd selected bits from */article.sgml, common/artheader.sgml, and
bmah [Thu, 15 Feb 2007 19:48:36 +0000 (19:48 +0000)]
Add selected bits from */article.sgml, common/artheader.sgml, and
common/intro.sgml to make a valid article.

17 years agoFirst step in consolidating the various MD bits of the hardware notes.
bmah [Thu, 15 Feb 2007 19:33:47 +0000 (19:33 +0000)]
First step in consolidating the various MD bits of the hardware notes.
Convert MD references in the supported devices section from arch=""
attributes for conditional compilation to entities that will just
print architecture names.  (The entities aren't defined yet...this
will happen in a future commit.)

17 years agoForced commit to note a repo-copy of this file from
bmah [Thu, 15 Feb 2007 19:10:23 +0000 (19:10 +0000)]
Forced commit to note a repo-copy of this file from
src/release/doc/en_US.ISO8859-1/hardware/common/dev.sgml.

17 years agoAdd examples for configuring and mounting geom_uzip(4) based disks'
matteo [Thu, 15 Feb 2007 18:07:52 +0000 (18:07 +0000)]
Add examples for configuring and mounting geom_uzip(4) based disks'
image and sliced/partitioned memory disks.
MFC after: 1 week

17 years agoremove some leftover text and update the text to match the current version
luigi [Thu, 15 Feb 2007 17:52:17 +0000 (17:52 +0000)]
remove some leftover text and update the text to match the current version
of the code.

17 years agoCleanup and document the implementation of firmware(9) based on
luigi [Thu, 15 Feb 2007 17:21:31 +0000 (17:21 +0000)]
Cleanup and document the implementation of firmware(9) based on
a version that i posted earlier on the -current mailing list,
and subsequent feedback received.

The core of the change is just in sys/firmware.h and kern/subr_firmware.c,
while other files are just adaptation of the clients to the ABI change
(const-ification of some parameters and hiding of internal info,
so this is fully compatible at the binary level).

In detail:
- reduce the amount of information exported to clients in struct firmware,
  and constify the pointer;

- internally, document and simplify the implementation of the various
  functions, and make sure error conditions are dealt with properly.

The diffs are large, but the code is really straightforward now (i hope).

Note also that there is a subtle issue with the implementation of
firmware_register(): currently, as in the previous version, we just
store a reference to the 'imagename' argument, but we should rather
copy it because there is no guarantee that this is a static string.
I realised this while testing this code, but i prefer to fix it in
a later commit -- there is no regression with respect to the past.

Note, too, that the version in RELENG_6 has various bugs including
missing locks around the module release calls, mishandling of modules
loaded by /boot/loader, and so on, so an MFC is absolutely necessary
there.  I was just postponing it until this cleanup to avoid doing
things twice.

MFC after: 1 week

17 years agofix comment about what pnpinfo is set
imp [Thu, 15 Feb 2007 16:38:10 +0000 (16:38 +0000)]
fix comment about what pnpinfo is set

17 years agoFix spurious I/O errors when under high load.
scottl [Thu, 15 Feb 2007 15:36:15 +0000 (15:36 +0000)]
Fix spurious I/O errors when under high load.

Submitted by: Erich Chen

17 years agoCorrect -c and -d description.
matteo [Thu, 15 Feb 2007 14:46:04 +0000 (14:46 +0000)]
Correct -c and -d description.

Other requests made in the PR were already solved in the past.

PR: bin/66763
MFC after: 1 week

17 years agoExpand history and authors section of mbuf.9 man page to discuss recent
rwatson [Thu, 15 Feb 2007 14:44:46 +0000 (14:44 +0000)]
Expand history and authors section of mbuf.9 man page to discuss recent
transition to mbuma (FreeBSD 5.3) and the fact that mbufs are now limited
almost entirely to packet storage, with straight UMA zones being used for
most other network data types.

17 years agoMention the nat command in the synopsis and in the action section.
piso [Thu, 15 Feb 2007 14:32:26 +0000 (14:32 +0000)]
Mention the nat command in the synopsis and in the action section.

Approved by: glebius (mentor)

17 years agoEnhances mdmfs(8) to mount md-based device such as uzip.
matteo [Thu, 15 Feb 2007 13:49:44 +0000 (13:49 +0000)]
Enhances mdmfs(8) to mount md-based device such as uzip.
Examples of use can be found in the PR text.

PR: 103501
MFC after: 1 week

17 years agoCatch up file descriptor printing function in DDB to the addition of kqueues
rwatson [Thu, 15 Feb 2007 10:55:43 +0000 (10:55 +0000)]
Catch up file descriptor printing function in DDB to the addition of kqueues
and POSIX message queues.

17 years agoBreak file descriptor printing logic out of db_show_files() into
rwatson [Thu, 15 Feb 2007 10:50:48 +0000 (10:50 +0000)]
Break file descriptor printing logic out of db_show_files() into
db_print_file(), and add a new "show file <ptr>" DDB command, which can
be used to print out file descriptors referenced in stack traces.

17 years agoRename somaxconn_sysctl() to sysctl_somaxconn() so that I will be able to
rwatson [Thu, 15 Feb 2007 10:11:00 +0000 (10:11 +0000)]
Rename somaxconn_sysctl() to sysctl_somaxconn() so that I will be able to
claim that sofoo() functions all accept a socket as their first argument.

17 years agoIf both ISDOTDOT and NOCROSSMOUNT are set then lookup() might breaks out
kib [Thu, 15 Feb 2007 09:53:49 +0000 (09:53 +0000)]
If both ISDOTDOT and NOCROSSMOUNT are set then lookup() might breaks out
of the special handling for ".." and perform an ISDOTDOT VOP_LOOKUP()
for a filesystem root vnode. Handle this case inside lookup().

Submitted by: tegge
PR: 92785
MFC after: 1 week

17 years agoStyle(9).
kib [Thu, 15 Feb 2007 09:24:58 +0000 (09:24 +0000)]
Style(9).

17 years agoIt turns out that it is easier to not NULL out pccard and cardbus
imp [Thu, 15 Feb 2007 07:22:27 +0000 (07:22 +0000)]
It turns out that it is easier to not NULL out pccard and cardbus
device pointers.  They don't change as the children device drivers
come and go.  Rather, check to see if the device is attached where we
would have checked ! NULL.  This solves many asymmetries in the code
that likely could lead to crashes when loading/unloading cbb without
one or more of the expected children's driver not present.

17 years agoFix three bugs:
imp [Thu, 15 Feb 2007 07:13:38 +0000 (07:13 +0000)]
Fix three bugs:
o When detaching all children, try really hard to get all the children
  list before giving up.  This is based on an observation by hans petter
  selasky in his usb p4 branch.
o When rescanning devices after a driver is added, abort if we can't get
  the child list with a message.
o when rescanning devices, if the reprobe/attach is successful, save the
  device for cardbus/pccard.

17 years agopkill(1) and pgrep(1) have been moved to /bin so that they are
yar [Thu, 15 Feb 2007 06:51:31 +0000 (06:51 +0000)]
pkill(1) and pgrep(1) have been moved to /bin so that they are
available to rc.d scripts early in the boot sequence.

17 years agoDon't be paranoid about hostname(1) and order the things logically.
yar [Thu, 15 Feb 2007 06:46:33 +0000 (06:46 +0000)]
Don't be paranoid about hostname(1) and order the things logically.

Pointed out by: ceri

17 years agoDon't nag about unset $hostname if DHCP is in use.
yar [Thu, 15 Feb 2007 06:42:42 +0000 (06:42 +0000)]
Don't nag about unset $hostname if DHCP is in use.

Pointed out by: ceri

17 years agoUnbreak non-H/W VLAN extraction case.
yongari [Thu, 15 Feb 2007 06:21:34 +0000 (06:21 +0000)]
Unbreak non-H/W VLAN extraction case.
Unlike other GigEs Yukon II always set VLAN bit when it detects VLAN
tagged packet regardless of H/W VLAN processing configuration state.
So it need to check IFCAP_VLAN_HWTAGGING bit to know whether driver
is configured to take advantage of H/W VLAN processing. If H/W VLAN
processing was disabled don't adjust received packet length such that
subsequent validation logic works for software VLAN processing.

Reported by: bms
Tested by: bms

17 years agoRelax the page queue lock assertions in vm_page_remove() and
alc [Thu, 15 Feb 2007 05:43:38 +0000 (05:43 +0000)]
Relax the page queue lock assertions in vm_page_remove() and
vm_page_free_toq() to account for recent changes that allow
vm_page_free_toq() to be called on some pages without the page queues lock
being held, specifically, pages that are not contained in a vm object and
not a member of a page queue.  (Examples of such pages include page table
pages, pv entry pages, and uma small alloc pages.)

17 years agoKill blank line at EOF.
trhodes [Thu, 15 Feb 2007 02:45:14 +0000 (02:45 +0000)]
Kill blank line at EOF.

17 years agoHook ypclnt.3 up to the build.
trhodes [Thu, 15 Feb 2007 02:43:14 +0000 (02:43 +0000)]
Hook ypclnt.3 up to the build.

17 years agoAdd a ypclnt.3 manual page referenced by various other YP based manual pages.
trhodes [Thu, 15 Feb 2007 02:40:31 +0000 (02:40 +0000)]
Add a ypclnt.3 manual page referenced by various other YP based manual pages.

PR: 108980
Obtained from: OpenBSD (minimal changes for mdoc(7) style)

17 years agoTeach DDB how to print sockets, socket buffers, protosw's, and domain
rwatson [Thu, 15 Feb 2007 01:28:22 +0000 (01:28 +0000)]
Teach DDB how to print sockets, socket buffers, protosw's, and domain
structures given pointers to them.

17 years agoFix accidental removal of an empty line from the previous commit.
jkim [Thu, 15 Feb 2007 01:20:43 +0000 (01:20 +0000)]
Fix accidental removal of an empty line from the previous commit.

17 years agoRegen.
jkim [Thu, 15 Feb 2007 01:15:31 +0000 (01:15 +0000)]
Regen.

17 years agoMFP4: 113033
jkim [Thu, 15 Feb 2007 01:13:36 +0000 (01:13 +0000)]
MFP4: 113033

Port iopl(2) from i386.  This fixes LTP iopl01 and iopl02 on amd64.

17 years agoRegen.
jkim [Thu, 15 Feb 2007 00:57:04 +0000 (00:57 +0000)]
Regen.

17 years agoMFP4: 113025, 113146, 113177, 113203, 113500, 113546, 113570
jkim [Thu, 15 Feb 2007 00:54:40 +0000 (00:54 +0000)]
MFP4: 113025, 113146, 113177, 113203, 113500, 113546, 113570

- PROT_READ, PROT_WRITE, or PROT_EXEC implies PROT_READ and PROT_EXEC.
Linux/ia64's i386 emulation layer does this and it complies with Linux
header files.  This fixes mmap05 LTP test case on amd64.
- Do not adjust stack size when failure has occurred.
- Synchronize i386 mmap/mprotect with amd64.

17 years agoDefault output to stdout as the man page suggests.
brian [Wed, 14 Feb 2007 23:22:33 +0000 (23:22 +0000)]
Default output to stdout as the man page suggests.

17 years agoOnly go through our cylinder group and inode info when we need to.
brian [Wed, 14 Feb 2007 23:20:30 +0000 (23:20 +0000)]
Only go through our cylinder group and inode info when we need to.
This allows ``ffsinfo -o - -l1 /tmp'' to run a lot quicker.

17 years agoRe-indent main() in preparation for further changes.
brian [Wed, 14 Feb 2007 23:12:32 +0000 (23:12 +0000)]
Re-indent main() in preparation for further changes.

17 years agoAdjust the global MSI blacklisting strategy so we don't have to explicitly
jhb [Wed, 14 Feb 2007 22:36:27 +0000 (22:36 +0000)]
Adjust the global MSI blacklisting strategy so we don't have to explicitly
blacklist a bunch of old chipsets.  If a system contains a PCI-PCI bridge
that supports PCI-X, assume the chipset supports PCI-X.  If a system
contains a PCI-express root port, assume the chipset supports PCI-express.
If the chipset doesn't support either PCI-X or PCI-express, then blacklist
it by default.  We should now only need to explicitly blacklist PCI-X or
PCI-express chipsets that don't properly handle MSI.

17 years ago- Fix an off by one error in pci_remap_msix_method() that effectively
jhb [Wed, 14 Feb 2007 22:32:55 +0000 (22:32 +0000)]
- Fix an off by one error in pci_remap_msix_method() that effectively
  broke the method as all the MSI-X table indices were off by one in
  the backend MD code.
- Fix a cosmetic nit in the bootverbose printf in pci_alloc_msix_method().

17 years agoCatch up to MSI-X API changes. Tested with both MSI and MSI-X.
jhb [Wed, 14 Feb 2007 22:31:21 +0000 (22:31 +0000)]
Catch up to MSI-X API changes.  Tested with both MSI and MSI-X.

17 years agoUse bge_writereg_ind() to do global reset as we did before 1.159 for certain
jkim [Wed, 14 Feb 2007 19:44:15 +0000 (19:44 +0000)]
Use bge_writereg_ind() to do global reset as we did before 1.159 for certain
chipsets.  It was causing 'firmware handshake timed out' errors for some
chips.

Discussed with: scottl

17 years agoFix two typos in comments.
jkim [Wed, 14 Feb 2007 19:01:33 +0000 (19:01 +0000)]
Fix two typos in comments.

17 years agoFix a typo from the previous commit.
jkim [Wed, 14 Feb 2007 18:21:32 +0000 (18:21 +0000)]
Fix a typo from the previous commit.

Pointed out by: brad@openbsd.org

17 years agoAdd missing 'break' that in this case is harmless.
jhb [Wed, 14 Feb 2007 17:02:15 +0000 (17:02 +0000)]
Add missing 'break' that in this case is harmless.

17 years agoFix compilation for statically linked snd_envy24{ht}/spicds. Use explicit
ariff [Wed, 14 Feb 2007 15:23:44 +0000 (15:23 +0000)]
Fix compilation for statically linked snd_envy24{ht}/spicds. Use explicit
struct mtx rather than void pointer.

PR: kern/109147

17 years agoMinor rearrangement of global variables, comments, etc, in UNIX domain
rwatson [Wed, 14 Feb 2007 15:05:40 +0000 (15:05 +0000)]
Minor rearrangement of global variables, comments, etc, in UNIX domain
sockets.

17 years agoRetire most of the classful network behaviour of netstat -r output, for IPv4.
bms [Wed, 14 Feb 2007 14:17:01 +0000 (14:17 +0000)]
Retire most of the classful network behaviour of netstat -r output, for IPv4.

 Without -n, we now only print a "network name" without the prefix length
 under the following conditions:
  1) the network address and mask matches a classful network prefix;
  2) getnetbyaddr(3) returns a network name for this network address.

 With -n, we unconditionally print the full unabbreviated CIDR network
 prefix in the form "a.b.c.d/p". 0.0.0.0/0 is still printed as "default".

This change is in preparation for changes such as equal-cost multipath, and
to more generally assist operational deployment of FreeBSD as a modern IPv4
router. There are currently no plans to backport this change.

Discussed on: freebsd-net

17 years agoChange unp_mtx to supporting recursion, and do not drop the unp_mtx over
rwatson [Wed, 14 Feb 2007 12:22:11 +0000 (12:22 +0000)]
Change unp_mtx to supporting recursion, and do not drop the unp_mtx over
sonewconn() in unp_connect().  This avoids a race that occurs due to
v_socket being an uncounted reference, as the lock was being released in
order to call sonewconn(), which otherwise recurses into the UNIX domain
socket code via pru_attach, as well as holding the lock over a sleeping
memory allocation in uipc_attach().  Switch to a non-sleeping memory
allocation during UNIX domain socket attach.

This fix non-ideal in that it requires enabling recursion, but is a much
smaller change than moving to using true references for v_socket.  The
reported panic occurs in unp_connect() following the return of
sonewconn().

Update copyright year.

Panic reported by:      jhb

17 years agoThe return value of aac_alloc_command() was misinterpreted in aac_ioctl_event().
luoqi [Wed, 14 Feb 2007 09:10:37 +0000 (09:10 +0000)]
The return value of aac_alloc_command() was misinterpreted in aac_ioctl_event().
Once triggered this would leak away all available commands and starve the rest
of the driver.

Reviewed by: scottl

17 years agoCorrectly handle writes beyond the end of the archive entry
kientzle [Wed, 14 Feb 2007 08:16:08 +0000 (08:16 +0000)]
Correctly handle writes beyond the end of the archive entry
(as determined by the initial size given to the header).
Libarchive recently changed to correctly return the amount
of data actually consumed in this case, which revealed this
bug in bsdtar.

17 years agoAdd some entries to log unknown devices, based on the bus they appear
joerg [Wed, 14 Feb 2007 08:13:54 +0000 (08:13 +0000)]
Add some entries to log unknown devices, based on the bus they appear
on.

Some things appear to be completely missing though, like the USB vendor
and product strings.

17 years agoNew sentence -> new line. While here, fix apostrophe abuse.
brueffer [Wed, 14 Feb 2007 07:38:39 +0000 (07:38 +0000)]
New sentence -> new line.  While here, fix apostrophe abuse.

17 years agoAvoid the unnecessary acquisition of the free page queues lock when a page
alc [Wed, 14 Feb 2007 07:05:55 +0000 (07:05 +0000)]
Avoid the unnecessary acquisition of the free page queues lock when a page
is actually being added to the hold queue, not the free queue.  At the same
time, avoid unnecessary tests to wake up threads waiting for free memory
and the idle thread that zeroes free pages.  (These tests will be performed
later when the page finally moves from the hold queue to the free queue.)

17 years agoOptimize bitcount32 by replacing 6 logical operations with 2. The key
cperciva [Wed, 14 Feb 2007 05:21:22 +0000 (05:21 +0000)]
Optimize bitcount32 by replacing 6 logical operations with 2.  The key
observation here is that it doesn't matter what garbage accumulates in
bits which we're going to end up masking away anyway, as long as the
garbage doesn't overflow into bits which we care about.

This improved version may not be the fastest possible on all systems,
but it's certainly going to be better than what was here before.

17 years agoAdd KTR tracing
kevlo [Wed, 14 Feb 2007 04:41:28 +0000 (04:41 +0000)]
Add KTR tracing

17 years agoAnother crashdump fix: nfiles was renamed to openfiles in 5.x.
jhb [Wed, 14 Feb 2007 04:20:41 +0000 (04:20 +0000)]
Another crashdump fix: nfiles was renamed to openfiles in 5.x.

MFC after: 3 days

17 years agostyle(9) cleanup.
kevlo [Wed, 14 Feb 2007 01:25:41 +0000 (01:25 +0000)]
style(9) cleanup.

17 years agoIn sendsig:
kevlo [Wed, 14 Feb 2007 01:08:42 +0000 (01:08 +0000)]
In sendsig:

- Add sigacts locking.
- Add a mutex to struct sigacts that protects all the members of the struct.
- Create and log events via the CTRx macros.

Reviewed by: cognet

17 years agoChange the date.
julian [Tue, 13 Feb 2007 23:06:39 +0000 (23:06 +0000)]
Change the date.

17 years agoMake the kse man page reflect the removal of the KSEGRP kernel abstraction.
julian [Tue, 13 Feb 2007 23:02:15 +0000 (23:02 +0000)]
Make the kse man page reflect the removal of the KSEGRP kernel abstraction.

17 years agoSet UNP_CONNECTING when committing to moving ahead in unp_connect().
rwatson [Tue, 13 Feb 2007 21:00:57 +0000 (21:00 +0000)]
Set UNP_CONNECTING when committing to moving ahead in unp_connect().
This logic was lost when merging the remainder of these changes in
1.178.

17 years agoMark up lkm with .Nm, since lkm is name of an api.
brueffer [Tue, 13 Feb 2007 17:06:15 +0000 (17:06 +0000)]
Mark up lkm with .Nm, since lkm is name of an api.

Submitted by: ru

17 years agoMake sure the address is valid before mapping it.
cognet [Tue, 13 Feb 2007 15:35:57 +0000 (15:35 +0000)]
Make sure the address is valid before mapping it.

MFC after: 1 week

17 years agoFix typo: MacPPC -> ARM
kevlo [Tue, 13 Feb 2007 07:19:26 +0000 (07:19 +0000)]
Fix typo: MacPPC -> ARM

17 years agoRaise WARNS level to 6.
rodrigc [Tue, 13 Feb 2007 05:39:38 +0000 (05:39 +0000)]
Raise WARNS level to 6.

17 years agoGet the vfs giant lock before calling nfs_access.
mpp [Tue, 13 Feb 2007 03:27:45 +0000 (03:27 +0000)]
Get the vfs giant lock before calling nfs_access.

Reviewed by: mohan

17 years agoRepocopied from src/sys/isofs/cd9660 to src/sys/fs/cd9660.
rodrigc [Tue, 13 Feb 2007 02:03:31 +0000 (02:03 +0000)]
Repocopied from src/sys/isofs/cd9660 to src/sys/fs/cd9660.

17 years agoMake vfs_getopts() set *error to ENOENT if the option wasn't found, so that
cognet [Tue, 13 Feb 2007 01:28:48 +0000 (01:28 +0000)]
Make vfs_getopts() set *error to ENOENT if the option wasn't found, so that
consumers don't have to check for both error and the return value (some of
them actually don't do it).

MFC After: 1 week

17 years agoFix typos in comments while I am here.
jkim [Tue, 13 Feb 2007 00:34:32 +0000 (00:34 +0000)]
Fix typos in comments while I am here.

17 years agoAllow the -c -f file options to actually execute.
mpp [Tue, 13 Feb 2007 00:22:29 +0000 (00:22 +0000)]
Allow the -c -f file options to actually execute.

17 years agoDo not do a vn_close for all references to the ktraced file if we are
mpp [Tue, 13 Feb 2007 00:20:13 +0000 (00:20 +0000)]
Do not do a vn_close for all references to the ktraced file if we are
doing a CLEARFILE option.  Do a vrele instead.  This prevents
a panic later due to v_writecount being negative when the vnode
is taken off the freelist.

Submitted by: jhb

17 years agoAdd BCM5701 A0/B0 CRC bug workaround. Magic values taken from Linux driver.
jkim [Mon, 12 Feb 2007 23:58:52 +0000 (23:58 +0000)]
Add BCM5701 A0/B0 CRC bug workaround.  Magic values taken from Linux driver.

17 years agoFix style(9).
jkim [Mon, 12 Feb 2007 23:33:05 +0000 (23:33 +0000)]
Fix style(9).

Pointed out by: many

17 years ago- Copyright updates (aka 2007)
rrs [Mon, 12 Feb 2007 23:24:31 +0000 (23:24 +0000)]
- Copyright updates (aka 2007)
- ZONE get now also take a type cast so it does the
  cast like mtod does.
- New macro SCTP_LIST_EMPTY, which in bsd is just
  LIST_EMPTY
- Removal of const in some of the static hmac functions
  (not needed)
- Store length changes to allow for new fields in auth
- Auth code updated to current draft (this should be the
  RFC version we think).
- use uint8_t instead of u_char in LOOPBACK address comparison
- Some u_int32_t converted to uint32_t (in crc code)
- A bug was found in the mib counts for ordered/unordered
  count, this was fixed (was referencing a freed mbuf).
- SCTP_ASOCLOG_OF_TSNS added (code will probably disappear
  after my testing completes. It allows us to keep a
  small log on each assoc of the last 40 TSN's in/out and
  stream assignment. It is NOT in options and so is only
  good for private builds.
- Some CMT changes in prep for Jana fixing his problem
  with reneging when CMT is enabled (Concurrent Multipath
  Transfer = CMT).
- Some missing mib stats added.
- Correction to number of open assoc's count in mib
- Correction to os_bsd.h to get right sha2 macros
- Add of special AUTH_04 flags so you can compile the code
  with the old format (in case the peer does not yet support
  the latest auth code).
- Nonce sum was incorrectly being set in when ecn_nonce was
  NOT on.
- LOR in listen with implicit bind found and fixed.
- Moved away from using mbuf's for socket options to using
  just data pointers. The mbufs were used to harmonize
  NetBSD code since both Net and Open used this method. We
  have decided to move away from that and more conform to
  FreeBSD style (which makes more sense).
- Very very nasty bug found in some of my "debug" code. The
  cookie_how collision case tracking had an endless loop in
  it if you got a second retransmission of a cookie collision
  case. This would lock up  a CPU .. ugly..
- auth function goes to using size_t instead of int which
  conforms to socketapi better
- Found the nasty bug that happens after 9 days of testing.. you
  get the data chunk, deliver it and due to the reference to a ch->
  that every now and then has been deleted (depending on the postion
  in the mbuf) you have an invalid ch->ch.flags.. and thus you don't
  advance the stream sequence number.. so you block the stream
  permanently. The fix is to make local variables of these guys
  and set them up before you have any chance of trimming the
  mbuf.
- style fix in sctp_util.h, not sure how this got bad maybe in
  the last patch? (aka it may not be in the real source).
- Found interesting bug when using the extended snd/rcv info where
  we would get an error on receiving with this. Thats because
  it was NOT padded to the same size as the snd_rcv info. We
  increase (add the pad) so the two structs are the same size
  in sctp_uio.h
- In sctp_usrreq.c one of the most common things we did for
  socket options was to cast the pointer and validate the size.
  This as been macro-ized to help make the code more readable.
- in sctputil.c two things, the socketapi class found a missing
  flag type (the next msg is a notification) and a missing
  scope recovery was also fixed.

Reviewed by: gnn

17 years agoAdd a VNASSERT to vn_close to detect if v_writecount is going
mpp [Mon, 12 Feb 2007 22:53:01 +0000 (22:53 +0000)]
Add a VNASSERT to vn_close to detect if v_writecount is going
to become negative.  This will detect the underflow when it
happens, instead of having it discovered when the vnode is
taken off the freelist, long after the offending process is long
gone.

17 years agoAdd PHY DSP code for BCM5755M.
jkim [Mon, 12 Feb 2007 22:51:25 +0000 (22:51 +0000)]
Add PHY DSP code for BCM5755M.

Obtained from: OpenBSD

17 years agoTypo fix in a comment
brueffer [Mon, 12 Feb 2007 21:41:17 +0000 (21:41 +0000)]
Typo fix in a comment

17 years ago"options PIM" is gone.
brueffer [Mon, 12 Feb 2007 21:30:34 +0000 (21:30 +0000)]
"options PIM" is gone.

17 years agoThe KAME project has been dissolved and kame.4 isn't terribly useful.
brueffer [Mon, 12 Feb 2007 21:12:37 +0000 (21:12 +0000)]
The KAME project has been dissolved and kame.4 isn't terribly useful.
=> Sayonara

Approved by: gnn, suz

17 years agoUse the correct line to put into loader.conf. Duh!
brueffer [Mon, 12 Feb 2007 20:58:32 +0000 (20:58 +0000)]
Use the correct line to put into loader.conf. Duh!

Clue bat applied by: ru
Pointy hat to: brueffer

17 years agoBCM5701 PHY cannot read-modify-write. Just re-use the magic number from DSP
jkim [Mon, 12 Feb 2007 20:26:56 +0000 (20:26 +0000)]
BCM5701 PHY cannot read-modify-write.  Just re-use the magic number from DSP
init code.

17 years agoReplace magic numbers with corresponding definitions.
jkim [Mon, 12 Feb 2007 19:33:22 +0000 (19:33 +0000)]
Replace magic numbers with corresponding definitions.

17 years agoMLINK for bus_dmamap_load_mbuf_sg(9)
jhb [Mon, 12 Feb 2007 17:53:21 +0000 (17:53 +0000)]
MLINK for bus_dmamap_load_mbuf_sg(9)

17 years agoRearrange the SATA connect logic so that we so that we pickup ATAPI devices.
sos [Mon, 12 Feb 2007 17:17:31 +0000 (17:17 +0000)]
Rearrange the SATA connect logic so that we so that we pickup ATAPI devices.
The rest of the logic should be in place for most supporting chipsets.

17 years agoRemove the cast to caddr_t for sfp, they're not needed.
kevlo [Mon, 12 Feb 2007 08:59:33 +0000 (08:59 +0000)]
Remove the cast to caddr_t for sfp, they're not needed.

Reviewed by: marcel

17 years agoImprove a sentence.
jkoshy [Mon, 12 Feb 2007 03:26:22 +0000 (03:26 +0000)]
Improve a sentence.

Prodded by: billf
MFC after: 3 days

17 years agoUse uma_set_align().
cognet [Sun, 11 Feb 2007 22:24:54 +0000 (22:24 +0000)]
Use uma_set_align().

17 years agoAdd uma_set_align() interface, which will be called at most once during
rwatson [Sun, 11 Feb 2007 20:13:52 +0000 (20:13 +0000)]
Add uma_set_align() interface, which will be called at most once during
boot by MD code to indicated detected alignment preference.  Rather than
cache alignment being encoded in UMA consumers by defining a global
alignment value of (16 - 1) in UMA_ALIGN_CACHE, UMA_ALIGN_CACHE is now
a special value (-1) that causes UMA to look at registered alignment.  If
no preferred alignment has been selected by MD code, a default alignment
of (16 - 1) will be used.

Currently, no hardware platforms specify alignment; architecture
maintainers will need to modify MD startup code to specify an alignment
if desired.  This must occur before initialization of UMA so that all UMA
zones pick up the requested alignment.

Reviewed by: jeff, alc
Submitted by: attilio

17 years agoo s/cgetseq/cgetset/
maxim [Sun, 11 Feb 2007 18:14:49 +0000 (18:14 +0000)]
o s/cgetseq/cgetset/

Obtained from: DragonFly

17 years agoFix some spelling / markup / grammar.
mpp [Sun, 11 Feb 2007 18:13:00 +0000 (18:13 +0000)]
Fix some spelling / markup / grammar.

Pointed out by: ru

17 years agoAdd two new options to quota:
mpp [Sun, 11 Feb 2007 16:25:25 +0000 (16:25 +0000)]
Add two new options to quota:

-f path
    Only print quota information for the file system that path resides on.
-r
    Display the quota information in a raw format.

Reviewed by: freebsd-hackers

17 years agoMakefile changes to reflect moving sys/isofs/cd9660 to sys/fs/cd9660.
rodrigc [Sun, 11 Feb 2007 14:01:32 +0000 (14:01 +0000)]
Makefile changes to reflect moving sys/isofs/cd9660 to sys/fs/cd9660.
Continue to install userland include files in /usr/include/isofs/cd9660
so as not to break userland applications such as libstand.

17 years agoForced commit and #include changes for repo copy from
rodrigc [Sun, 11 Feb 2007 13:54:25 +0000 (13:54 +0000)]
Forced commit and #include changes for repo copy from
sys/isofs/cd9660 to sys/fs/cd9660.

Discussed on freebsd-current.

17 years agoFix PowerPC-targeted cross-builds performed on 64-bit platforms.
marcel [Sun, 11 Feb 2007 07:15:06 +0000 (07:15 +0000)]
Fix PowerPC-targeted cross-builds performed on 64-bit platforms.
Only PowerPC supports both 32-bit and 64-bit targets and the
BFD_DEFAULT_TARGET_SIZE is used by the binutils code to reflect
the preferred ABI. We define BFD_DEFAULT_TARGET_SIZE for all
platforms, but based on the build machine. As such 64-bit build
machines defined BFD_DEFAULT_TARGET_SIZE incorrectly for 32-bit
targets, but since this only affects PowerPC it went unnoticed
for a long time.
The fix is to define BFD_DEFAULT_TARGET_SIZE based on the target
architecture.

PR: amd64/102996
MFC after: 1 month

17 years agoUse the free page queue mutex instead of the page queue mutex to
alc [Sun, 11 Feb 2007 05:18:40 +0000 (05:18 +0000)]
Use the free page queue mutex instead of the page queue mutex to
synchronize sleeping and waking of the zero idle thread.

17 years agoAdd missing MNT_ILOCK around some mnt_kern_flag accesses.
mohans [Sun, 11 Feb 2007 04:01:10 +0000 (04:01 +0000)]
Add missing MNT_ILOCK around some mnt_kern_flag accesses.

17 years agoNow that the free page queue mutex is a sleep mutex, we cannot call
marcel [Sun, 11 Feb 2007 02:52:54 +0000 (02:52 +0000)]
Now that the free page queue mutex is a sleep mutex, we cannot call
vm_page_alloc() from within a critical section in pmap_growkernel().
Since the need for a critical section may never have existed in the
first place, simply get rid of it.

Discussed with: alc@

17 years agoUse MAXTTL.
bms [Sat, 10 Feb 2007 23:15:28 +0000 (23:15 +0000)]
Use MAXTTL.

Obtained from: NetBSD