]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
17 years agoMention that BSD gzip was MFC'ed
delphij [Thu, 22 Feb 2007 05:31:39 +0000 (05:31 +0000)]
Mention that BSD gzip was MFC'ed

17 years agoFix a bug in if_findmulti(), whereby it would not find (and thus delete)
bms [Thu, 22 Feb 2007 00:14:02 +0000 (00:14 +0000)]
Fix a bug in if_findmulti(), whereby it would not find (and thus delete)
a link-layer multicast group membership.
Such memberships are needed in order to support protocols such as
IS-IS without putting the interface into PROMISC or ALLMULTI modes.

sa_equal() is not OK for comparing sockaddr_dl as it has deeper structure
than a simple byte array, so add sa_dl_equal() and use that instead.

Reviewed by: rwatson
Verified with: /usr/sbin/mtest
Bug found by: Jouke Witteveen
MFC after: 2 weeks

17 years agoUpdate copyright headers.
sos [Wed, 21 Feb 2007 19:07:19 +0000 (19:07 +0000)]
Update copyright headers.

17 years agoTry again with supporting AHCI chipsets with partly implemented ports.
sos [Wed, 21 Feb 2007 19:03:34 +0000 (19:03 +0000)]
Try again with supporting AHCI chipsets with partly implemented ports.

17 years agoRestore support for the 5706C bce(4) phy that was broken during the
jhb [Wed, 21 Feb 2007 18:17:44 +0000 (18:17 +0000)]
Restore support for the 5706C bce(4) phy that was broken during the
addition of SerDes support.  According to the docs, the 5706C and 5708C
phys are supposed to use the same MII model that is separate from the
SerDes parts, but the 5706C actually uses the MII model of the SerDes
parts.  To fix this, readd the old 5706C entry to miidevs and add a
special check in brgphy_probe() for phys that match the 5706C ID.  If
the phy is supported by the gentbi(4) driver, then it's a SerDes phy, so
we fail the probe and let gentbi(4) grab it.  Otherwise, it's a 5706C phy,
so we let brgphy(4) grab it.

In coordination with: dwhite

17 years agoWork around a firmware bug where broadcast frames would be incorrectly
gallatin [Wed, 21 Feb 2007 17:34:05 +0000 (17:34 +0000)]
Work around a firmware bug where broadcast frames would be incorrectly
treated as multicast frames and filtered, but when only when "adopting"
running firmware.  By "adopting", I mean using pre-existing firmware
loaded from eeprom at PCI reset, rather than firmware loaded by the
driver.

17 years agoBe a little less stringent in getting progress report data.
sos [Wed, 21 Feb 2007 14:06:51 +0000 (14:06 +0000)]
Be a little less stringent in getting progress report data.

17 years agoTemporarily disable the support for some incarnations of the ICH8 chip that has
sos [Wed, 21 Feb 2007 14:05:57 +0000 (14:05 +0000)]
Temporarily disable the support for some incarnations of the ICH8 chip that has
non consecutively numbered ports.
This should fix current SATA problems.
Support AHCI chips where the ports are not consecutively numbered as in
some incarnations of the ICH8 chip.

17 years agoUpdate host-mode multicast group information output.
bms [Wed, 21 Feb 2007 13:59:21 +0000 (13:59 +0000)]
Update host-mode multicast group information output.
 Display IPv4 and IPv6 memberships separately.
  Obey the MK_INET6_SUPPORT flag.
 Display link-layer memberships.
  Use addr2ascii() to correctly print non-IEEE 802 sockaddr_dl instances.
 Eliminate redundant switch..case blocks.
 Update copyright.
 Misc style changes.

MFC after: 3 weeks

17 years agoChange wording of warnings when there is no ip_mroute.ko module
bms [Wed, 21 Feb 2007 13:41:51 +0000 (13:41 +0000)]
Change wording of warnings when there is no ip_mroute.ko module
loaded into the system.
Change wording of comments to reflect the fact we should unconditionally
use KVM if the -M option is used to specify a core file.
Add comments to document the fact that IPv6 multicast forwarding
information display still relies on KVM for gathering information.

17 years agoOptimize set_origin() use in some screensavers to stop them eating
philip [Wed, 21 Feb 2007 12:27:12 +0000 (12:27 +0000)]
Optimize set_origin() use in some screensavers to stop them eating
cpu power when the origin hasn't changed.

PR: kern/100683
Submitted by: Gareth McCaughan <gareth -dot- mccaughan -at- pobox.com>
MFC after: 3 days

17 years agoRefine implementation notes for priv(9): clarify ABI comments, mention
rwatson [Wed, 21 Feb 2007 10:32:03 +0000 (10:32 +0000)]
Refine implementation notes for priv(9): clarify ABI comments, mention
updating Jail's list of privileges.

17 years agoDocument vge(4)'s support for altq(4).
brueffer [Wed, 21 Feb 2007 10:00:09 +0000 (10:00 +0000)]
Document vge(4)'s support for altq(4).

17 years agoComplete the support for altq(4).
brueffer [Wed, 21 Feb 2007 09:57:27 +0000 (09:57 +0000)]
Complete the support for altq(4).

Tested by: J.R. Oldroyd
Reviewed by: mlaier
Approved by: rwatson (mentor)
MFC after: 1 week

17 years agoThe functions that set and delete external attributes must check
mckusick [Wed, 21 Feb 2007 08:50:06 +0000 (08:50 +0000)]
The functions that set and delete external attributes must check
that the filesystem is not mounted read-only before proceeding.

Reported by: Ryan Beasley <ryanb@FreeBSD.org>
MFC after: 1 week

17 years agoFix a bug with the release of section's raw data. Both release loops
dumbbell [Wed, 21 Feb 2007 08:14:22 +0000 (08:14 +0000)]
Fix a bug with the release of section's raw data. Both release loops
were using translated data linked list, leading to a memory leak.

Jkoshy's testsuite was used to check for non-regression.

17 years agoInitialise {transport,protocol}{,_version} fields during a PATH_INQ to avoid a
n_hibma [Wed, 21 Feb 2007 07:46:40 +0000 (07:46 +0000)]
Initialise {transport,protocol}{,_version} fields during a PATH_INQ to avoid a
warning message.

17 years agoReduce the noise when plugging in (USB) mass storage devices, like a 4 port
n_hibma [Wed, 21 Feb 2007 07:45:02 +0000 (07:45 +0000)]
Reduce the noise when plugging in (USB) mass storage devices, like a 4 port
flash card reader.
Also remove an 'Opened da0 -> <random number>' which is not needed on a daily
basis (available through bootverbose).

Reviewed by: phk, ken
MFC after: 1 week

17 years agoChange pmap_protect() so that execute access can be removed without
alc [Wed, 21 Feb 2007 06:00:46 +0000 (06:00 +0000)]
Change pmap_protect() so that execute access can be removed without
simultaneously removing write access.

17 years agoIf we already have stat() data, we might be able to
kientzle [Wed, 21 Feb 2007 05:07:43 +0000 (05:07 +0000)]
If we already have stat() data, we might be able to
determine if this is a physical dir without an lstat().
While I'm in here, try to clarify the comments around
the _is_dir() and _is_physical_dir() tests.

17 years agoCheck that the error returned by vfs_getopts() is not ENOENT before assuming
cognet [Wed, 21 Feb 2007 00:30:09 +0000 (00:30 +0000)]
Check that the error returned by vfs_getopts() is not ENOENT before assuming
there's actually an error.
This is just in order to unbreak ntfs on current, before a proper solution is
committed.

17 years agoRemove unnecessary privilege and privilege check for WITNESS sysctl.
rwatson [Tue, 20 Feb 2007 23:49:31 +0000 (23:49 +0000)]
Remove unnecessary privilege and privilege check for WITNESS sysctl.

Head nod: jhb

17 years agoCreate a link from hosts.allow(5) hosts_access(3), to give the user a
n_hibma [Tue, 20 Feb 2007 23:12:04 +0000 (23:12 +0000)]
Create a link from hosts.allow(5) hosts_access(3), to give the user a
starting point for more information on a file in /etc.

MFC after: 1 week

17 years agoKris suggested that swap is a better choice as a default than malloc.
n_hibma [Tue, 20 Feb 2007 22:04:23 +0000 (22:04 +0000)]
Kris suggested that swap is a better choice as a default than malloc.

MFC: 1 week

17 years ago[Found the original diff I made, see previous commit for other part]
n_hibma [Tue, 20 Feb 2007 21:29:30 +0000 (21:29 +0000)]
[Found the original diff I made, see previous commit for other part]
Assume '-a' and '-t malloc' flags for '-s <size>' (malloc ramdisk) if not
specified.

Reviewed by: phk (some time ago)
MFC: 1 week

17 years agoMake attach the default for -f. That way
n_hibma [Tue, 20 Feb 2007 21:04:12 +0000 (21:04 +0000)]
Make attach the default for -f. That way

mdconfig -f image

works like a charm.

Reviewed by: phk (some time ago)
MFC: 1 week

17 years agoAvoid writing uninitialized stack data into a thread's MMX/SSE state by
emaste [Tue, 20 Feb 2007 18:10:13 +0000 (18:10 +0000)]
Avoid writing uninitialized stack data into a thread's MMX/SSE state by
first getting the current state with td_thr_getxmmregs_p.  Without this,
debugging a threaded app that uses libthr resulted in kernel panics or
spurious SIGFPEs for me.

(As of revision 1.6, sys/i386/i386/ptrace_machdep.c masks off the
reserved bits in the mxcsr register, which prevents the kernel panics.)

Architectures without PT_GETXMMREGS are not affected.

MFC after:      1 week

17 years agoRename IWI_LOCK_ASSERT to IWI_LOCK_CHECK per Sam's suggestion,
luigi [Tue, 20 Feb 2007 17:32:30 +0000 (17:32 +0000)]
Rename IWI_LOCK_ASSERT to IWI_LOCK_CHECK per Sam's suggestion,
and make it print under debug.iwi control same as other debugging stuff.

Remove the device_printf() in iwi_ioctl() and replace with this:

        /*
         * wait until pending iwi_cmd() are completed, to avoid races
         * that could cause problems.
         */
        while (sc->flags & IWI_FLAG_BUSY)
                msleep(sc, &sc->sc_mtx, 0, "iwiioctl", hz);

This at least prevents what has become an almost systematic failure for my
system, presumably due to a previous iwi_cmd() not complete yet by the
time iwi_ioctl() is called.

It has been pointed to my attention that the real problem could be
calling ieee80211_ioctl() with the lock held. If that is true,
there might still be a possibility for a race condition e.g. an
interrupt coming while the ioctl is sleeping.
Need to investigate further on what changes are required to release
the lock before calling ieee80211_ioctl

17 years agoAddress a few issues with the iwi driver, namely:
luigi [Tue, 20 Feb 2007 15:45:59 +0000 (15:45 +0000)]
Address a few issues with the iwi driver, namely:

+ do not release the dma-ble region used for downloading firmware.
  This should fix the problems that some people were seeing, due to
  memory becoming too fragmented which prevented subsequent allocations
  of a suitable contiguous region of memory;

+ document the firmware format and usage in if_iwivar.h

+ use a loop to allocate the four tx rings, instead of replicating
  the body of the loop.

+ add debugging code IWI_LOCK_ASSERT() to detect missing locks.
  These only do a printf, and should go away once we figure out why
  the driver sometimes freezes the system due to a (yet unidentified)
  race condition.

+ add a device_printf() in iwi_ioctl() in certain conditions
  (see comment in the code).  This helps preventing the race condition
  mentioned above, and makes the system survive. This printf will
  also go away once fixing this bug is completed.

+ change iwi_getfw() to return 0 on success, 1 on error, consistently
  with other functions.

+ fix the argument of a sizeof() in iwi_get_firmware()

+ use le32toh() to access little-endian fields

+ simplify error handling in iwi_load_firmware() and iwi_init_locked()

The bugs fixed by this commit (the freezing one especially) are serious
enough to call for a quick MFC

MFC after: 3 days

17 years agoReplace a suser() check with an explicit check for PRIV_NET_SETIFMTU.
rwatson [Tue, 20 Feb 2007 15:20:36 +0000 (15:20 +0000)]
Replace a suser() check with an explicit check for PRIV_NET_SETIFMTU.

17 years agoDocument the endiannes of firmware headers
luigi [Tue, 20 Feb 2007 14:29:09 +0000 (14:29 +0000)]
Document the endiannes of firmware headers
(in preparation for changes in the C code).

17 years agoUpdate auditing of socket information for the inpcb new world order:
rwatson [Tue, 20 Feb 2007 13:38:11 +0000 (13:38 +0000)]
Update auditing of socket information for the inpcb new world order:
so_pcb will always be non-NULL, and lock the inpcb while non-atomically
accessing address data.

17 years agoBreak introductory comment into two paragraphs to separate material on the
rwatson [Tue, 20 Feb 2007 10:50:02 +0000 (10:50 +0000)]
Break introductory comment into two paragraphs to separate material on the
garbage collection complications from general discussion of UNIX domain
sockets.

Staticize unp_addsockcred().

Remove XXX comment regarding Giant and v_socket -- v_socket is protected
by the global UNIX domain socket lock.

17 years agoMove mapping of MBI_APPEND to MBI_WRITE from inside the rule loop in
rwatson [Tue, 20 Feb 2007 10:21:27 +0000 (10:21 +0000)]
Move mapping of MBI_APPEND to MBI_WRITE from inside the rule loop in
mac_bsdextended_check() to before the loop, as it needs to happen only
once.

MFC after: 1 week

17 years agoRename two identically named log_in_vain variables: tcp_input.c's static
rwatson [Tue, 20 Feb 2007 10:20:03 +0000 (10:20 +0000)]
Rename two identically named log_in_vain variables: tcp_input.c's static
log_in_vain to tcp_log_in_vain, and udp_usrreq's global log_in_vain to
udp_log_in_vain.

MFC after: 1 week

17 years agoGratuitous UDP restyling toward style(9) in 7.x.
rwatson [Tue, 20 Feb 2007 10:13:11 +0000 (10:13 +0000)]
Gratuitous UDP restyling toward style(9) in 7.x.

17 years agoRemove discontinuity in network privilege number space.
rwatson [Tue, 20 Feb 2007 00:28:19 +0000 (00:28 +0000)]
Remove discontinuity in network privilege number space.

Spotted by: emaste (ages ago)

17 years agoRemove unused PRIV_IPC_EXEC. Renumbers System V IPC privilege.
rwatson [Tue, 20 Feb 2007 00:12:52 +0000 (00:12 +0000)]
Remove unused PRIV_IPC_EXEC.  Renumbers System V IPC privilege.

17 years agoSync up PRIV_IPC_{ADMIN,READ,WRITE} priv checks in ipcperm() with
rwatson [Tue, 20 Feb 2007 00:06:59 +0000 (00:06 +0000)]
Sync up PRIV_IPC_{ADMIN,READ,WRITE} priv checks in ipcperm() with
kern_jail.c: allow jailed root these privileges.  This only has an
effect if System V IPC is administratively enabled for the jail.

17 years agoRestore sysv_ipc.c:1.30, which was backed out due to interactions with
rwatson [Mon, 19 Feb 2007 22:59:23 +0000 (22:59 +0000)]
Restore sysv_ipc.c:1.30, which was backed out due to interactions with
System V shared memory, now believed fixed in sysv_shm.c:1.109:

  date: 2006/11/06 13:42:01;  author: rwatson;  state: Exp;  lines: +65 -37
  Sweep kernel replacing suser(9) calls with priv(9) calls, assigning
  specific privilege names to a broad range of privileges.  These may
  require some future tweaking.

  Sponsored by:           nCircle Network Security, Inc.
  Obtained from:          TrustedBSD Project
  Discussed on:           arch@
  Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri,
                          Alex Lyashkov <umka at sevcity dot net>,
                          Skip Ford <skip dot ford at verizon dot net>,
                          Antoine Brodin <antoine dot brodin at laposte dot net>

This restores fine-grained privilege support to System V IPC.

PR: 106078

17 years agoRemove call to ipcperm() in shmget_existing(). The flags argument is
rwatson [Mon, 19 Feb 2007 22:56:10 +0000 (22:56 +0000)]
Remove call to ipcperm() in shmget_existing().  The flags argument is
ignored on other systems I investigated when accessing an existing
memory segment rather than creating a new one.  This call to ipcperm()
is the only one to pass in a complete mode flag to the permission
checks rather than a simple access request mask, and caused problems
for the revised ipcperm() based on the priv(9) interface, which can
now be restored.

PR: 106078

17 years agoNote problems I had with bsnmpd while updating from an older -current.
njl [Mon, 19 Feb 2007 22:49:43 +0000 (22:49 +0000)]
Note problems I had with bsnmpd while updating from an older -current.
There may be better ways to fix/work around these issues but this worked
for me.

17 years agoUse privilege PRIV_NET_ADDIFADDR rather than suser() to authorize
rwatson [Mon, 19 Feb 2007 22:40:02 +0000 (22:40 +0000)]
Use privilege PRIV_NET_ADDIFADDR rather than suser() to authorize
adding a netatalk address to an interface.

17 years agoRename three quota privileges from the UFS privilege namespace to the
rwatson [Mon, 19 Feb 2007 13:33:10 +0000 (13:33 +0000)]
Rename three quota privileges from the UFS privilege namespace to the
VFS privilege namespace: exceedquota, getquota, and setquota.  Leave
UFS-specific quota configuration privileges in the UFS name space.

This renumbers VFS and UFS privileges, so requires rebuilding modules
if you are using security policies aware of privilege identifiers.
This is likely no one at this point since none of the committed MAC
policies use the privilege checks.

17 years agoLimit quota privileges in jail to PRIV_UFS_GETQUOTA and
rwatson [Mon, 19 Feb 2007 13:26:39 +0000 (13:26 +0000)]
Limit quota privileges in jail to PRIV_UFS_GETQUOTA and
PRIV_UFS_SETQUOTA.

17 years agoDo allow bypass of mac_seeotheruids in jail in order to be consistent
rwatson [Mon, 19 Feb 2007 13:25:17 +0000 (13:25 +0000)]
Do allow bypass of mac_seeotheruids in jail in order to be consistent
with other uses of PRIV_SEEOTHERUIDS.  This will automatically be
scoped to the jail by the jail policy.

17 years agoDo allow privilege to create over-sized messages on System V IPC
rwatson [Mon, 19 Feb 2007 13:23:45 +0000 (13:23 +0000)]
Do allow privilege to create over-sized messages on System V IPC
message queues in jail.

17 years agoUse priv_check(9) instead of suser(9) for checking the privilege to
rwatson [Mon, 19 Feb 2007 13:22:36 +0000 (13:22 +0000)]
Use priv_check(9) instead of suser(9) for checking the privilege to
set real-time priority on a thread.  It looks like this suser(9)
call was introduced after my first pass through replacing superuser
checks with named privilege checks.

17 years agoFor now, reflect practical reality that Audit system calls aren't
rwatson [Mon, 19 Feb 2007 13:10:29 +0000 (13:10 +0000)]
For now, reflect practical reality that Audit system calls aren't
allowed in Jail: return a privilege error.

17 years agoDo allow PIOCSFL in jail for setguid processes; this is more consistent
rwatson [Mon, 19 Feb 2007 13:04:25 +0000 (13:04 +0000)]
Do allow PIOCSFL in jail for setguid processes; this is more consistent
with other debugging checks elsewhere.  XXX comment on the fact that
p_candebug() is not being used here remains.

17 years agoUnbreak ddb stepping over special frames after the following commit:
kib [Mon, 19 Feb 2007 10:57:47 +0000 (10:57 +0000)]
Unbreak ddb stepping over special frames after the following commit:
 Revision  Changes    Path
  1.113     +4 -2      src/sys/i386/i386/apic_vector.s
  1.117     +7 -1      src/sys/i386/i386/exception.s
  1.36      +7 -7      src/sys/i386/i386/local_apic.c
  1.298     +61 -63    src/sys/i386/i386/trap.c
  1.62      +15 -22    src/sys/i386/i386/vm86.c
  1.32      +4 -2      src/sys/i386/i386/vm86bios.s
  1.21      +2 -2      src/sys/i386/include/apicvar.h
  1.27      +2 -2      src/sys/i386/isa/atpic.c
  1.50      +2 -1      src/sys/i386/isa/atpic_vector.s
  1.35      +1 -1      src/sys/i386/isa/icu.h

Tested by: kris, Peter Holm
No objections from: kmacy

17 years agoRemove union_dircheckp hook, it is not needed by new unionfs code anymore.
kib [Mon, 19 Feb 2007 10:56:09 +0000 (10:56 +0000)]
Remove union_dircheckp hook, it is not needed by new unionfs code anymore.
As consequence, getdirentries() no longer needs to drop/reacquire
directory vnode lock, that would allow it to be reclaimed in between.

Reported and tested by: Peter Holm
Approved by: rodrigc (unionfs)
MFC after: 1 week

17 years agoMFi386 rev. 1.544 of i386/i386/pmap.c:
kib [Mon, 19 Feb 2007 10:55:16 +0000 (10:55 +0000)]
MFi386 rev. 1.544 of i386/i386/pmap.c:
Rounding addr upwards to next 2M boundary in pmap_growkernel() could
cause addr to become 0, resulting in an early return without populating
the last PDE.

Reported and tested by: kris
Suggested by: alc
MFC after: 1 week

17 years agoIt'd be kind of nice if we installed the hardware notes to their
bmah [Mon, 19 Feb 2007 02:46:12 +0000 (02:46 +0000)]
It'd be kind of nice if we installed the hardware notes to their
own directory, rather than overwriting the release notes.  :-p
Fix this.

Pointy hat to: bmah

17 years agoAdd two new options, FLASHADDR, which defines the address the flash is
cognet [Mon, 19 Feb 2007 01:03:08 +0000 (01:03 +0000)]
Add two new options, FLASHADDR, which defines the address the flash is
mapped at, and LOADERRAMADDR, the address at which the loader maps the ram at
at the time the kernel is booted.
They are used to detect if the kernel is booted from the onboard flash.
Define those for the IQ31244

17 years agoTeach the kernel and the ELF trampoline how to boot from onboard flash.
cognet [Mon, 19 Feb 2007 00:57:27 +0000 (00:57 +0000)]
Teach the kernel and the ELF trampoline how to boot from onboard flash.

17 years agoThere's no such thing as a GENERIC kernel on arm.
cognet [Mon, 19 Feb 2007 00:37:25 +0000 (00:37 +0000)]
There's no such thing as a GENERIC kernel on arm.

Spotted out by: csjp
MFC After: 3 days

17 years agoBring these files up-to-date.
ceri [Sun, 18 Feb 2007 22:41:41 +0000 (22:41 +0000)]
Bring these files up-to-date.

PR: docs/108536

17 years agoClarify the description of the shutdown command.
ceri [Sun, 18 Feb 2007 22:04:14 +0000 (22:04 +0000)]
Clarify the description of the shutdown command.

17 years agoCorrect typos containing my login name (plus one more in expr.y).
ceri [Sun, 18 Feb 2007 19:48:59 +0000 (19:48 +0000)]
Correct typos containing my login name (plus one more in expr.y).
Found courtesy of a recursive grep in the wrong directory.

17 years agoThe table of known CPU models ends with an entry that has a version
marcel [Sun, 18 Feb 2007 17:40:09 +0000 (17:40 +0000)]
The table of known CPU models ends with an entry that has a version
of 0, not with an entry that has an empty CPU name.

Submitted by: Andrew Turner (andrew@fubar.geek.nz)

17 years ago#ifdef INET6 printing of inpcb IPv6 addresses in DDB. Patch committed
rwatson [Sun, 18 Feb 2007 08:57:23 +0000 (08:57 +0000)]
#ifdef INET6 printing of inpcb IPv6 addresses in DDB.  Patch committed
with minor adjustments.

Submitted by: Florian C. Smeets <flo at kasimir dot com>

17 years agoEliminate some acquisitions and releases of the page queues lock that are
alc [Sun, 18 Feb 2007 06:33:02 +0000 (06:33 +0000)]
Eliminate some acquisitions and releases of the page queues lock that are
no longer necessary.

17 years agoAndrew and Colin each pointed out to me that truncating the backup
kientzle [Sun, 18 Feb 2007 06:23:57 +0000 (06:23 +0000)]
Andrew and Colin each pointed out to me that truncating the backup
of a growing file should not be considered a "bad thing."

PR: bin/108990
MFC after: 7 days
Pointy hat: /me

17 years agoEnable vm_page_free() and vm_page_free_zero() to be called on some pages
alc [Sun, 18 Feb 2007 05:54:42 +0000 (05:54 +0000)]
Enable vm_page_free() and vm_page_free_zero() 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.

17 years agoAdd "show inpcb", "show tcpcb" DDB commands, which should come in handy
rwatson [Sat, 17 Feb 2007 21:02:38 +0000 (21:02 +0000)]
Add "show inpcb", "show tcpcb" DDB commands, which should come in handy
for debugging sblock and other network panics.

17 years agoRemove a stale comment. Add punctuation to a nearby comment.
alc [Sat, 17 Feb 2007 19:37:00 +0000 (19:37 +0000)]
Remove a stale comment.  Add punctuation to a nearby comment.

17 years agocorrect type definition of option routine callback
sam [Sat, 17 Feb 2007 18:22:20 +0000 (18:22 +0000)]
correct type definition of option routine callback

17 years agoremove (now) duplicate definition of IEEE80211_IS_CHAN_PASSIVE
sam [Sat, 17 Feb 2007 18:21:52 +0000 (18:21 +0000)]
remove (now) duplicate definition of IEEE80211_IS_CHAN_PASSIVE

17 years agoAdd support for chipsets that has NULL'd BAR's for legacy ports.
sos [Sat, 17 Feb 2007 16:56:39 +0000 (16:56 +0000)]
Add support for chipsets that has NULL'd BAR's for legacy ports.
This allows DMA to be used on a fine little geode system I got here and
most like on lots of older systems like that.

HW donated by:  Paul Ghering

17 years agoOops, fix minor braino.
joel [Sat, 17 Feb 2007 11:31:58 +0000 (11:31 +0000)]
Oops, fix minor braino.

Noticed by: ariff

17 years agoLink snd_sb16.4 and snd_sb8.4 to snd_sbc.4.
joel [Sat, 17 Feb 2007 10:30:00 +0000 (10:30 +0000)]
Link snd_sb16.4 and snd_sb8.4 to snd_sbc.4.

Submitted by: ariff

17 years agoThis README file is obsolete. The cited problems were fixed long ago
mckusick [Sat, 17 Feb 2007 08:25:43 +0000 (08:25 +0000)]
This README file is obsolete. The cited problems were fixed long ago
and the code is installed by default so no longer requires action by
the administrator to be included.

17 years agoFix mis-reference of incorrect manual page in ERRORS section.
trhodes [Sat, 17 Feb 2007 01:54:00 +0000 (01:54 +0000)]
Fix mis-reference of incorrect manual page in ERRORS section.

Noticed by: Harlan Stenn <stenn@ntp.isc.org>

17 years agoAdd Sigmatel STAC9271D.
joel [Fri, 16 Feb 2007 18:23:17 +0000 (18:23 +0000)]
Add Sigmatel STAC9271D.

17 years agoAdd codec id for Sigmatel STAC9271D.
ariff [Fri, 16 Feb 2007 17:43:29 +0000 (17:43 +0000)]
Add codec id for Sigmatel STAC9271D.

Submitted by: Michael Fuckner <michael@fuckner.net>

17 years agoRemove VFS_VPTOFH entirely. API is already broken and it is good time to
pjd [Fri, 16 Feb 2007 17:32:41 +0000 (17:32 +0000)]
Remove VFS_VPTOFH entirely. API is already broken and it is good time to
do it.

Suggested by: rwatson

17 years agoVFS_VPTOFH(9) was replaced with VOP_VPTOFH(9).
pjd [Fri, 16 Feb 2007 14:27:59 +0000 (14:27 +0000)]
VFS_VPTOFH(9) was replaced with VOP_VPTOFH(9).
VFS_VPTOFH.9 was repo-copied to VOP_VPTOFH.9.

Repo-copy done by: joe

17 years agoRemove unused inp6_ifindex field from inpcb, as well as unused macro
rwatson [Fri, 16 Feb 2007 14:09:24 +0000 (14:09 +0000)]
Remove unused inp6_ifindex field from inpcb, as well as unused macro
shortcut for it.

17 years agoRemove unused in6p_ip6_hlim macro shortcut for non-present
rwatson [Fri, 16 Feb 2007 13:56:06 +0000 (13:56 +0000)]
Remove unused in6p_ip6_hlim macro shortcut for non-present
inp_depend6.inp6_hlim field in the inpcb.

17 years agoXref altq(4)
brueffer [Fri, 16 Feb 2007 13:53:45 +0000 (13:53 +0000)]
Xref altq(4)

17 years agoCompUSA mutli-format pcmica flash reader/writer
imp [Fri, 16 Feb 2007 07:51:23 +0000 (07:51 +0000)]
CompUSA mutli-format pcmica flash reader/writer

17 years agoUpdated tuple list, kinda
imp [Fri, 16 Feb 2007 07:00:50 +0000 (07:00 +0000)]
Updated tuple list, kinda

17 years agoAdd poitner to JEDEC publication 106
imp [Fri, 16 Feb 2007 06:46:36 +0000 (06:46 +0000)]
Add poitner to JEDEC publication 106

17 years agoTwo changes:
imp [Fri, 16 Feb 2007 05:36:59 +0000 (05:36 +0000)]
Two changes:
(1) change debounce period from 1s to 250ms.  This appears to be fine and
    speeds things up a little.
(2) In the middle of cbb_pcic_power_disable_socket we write 0 to the EXCA_INTR
    register to put the card into reset.  However, this turns off CSC
    interrupts for TI bridges (and maybe others).  So no further card
    insertion events would be noticed.  To compensate, after we've gone
    through the entire power down sequence, turn on EXCA_INTR_ENABLE so
    that CSC events happen.

#2 should fix the 'dead slot' problem that has been reported after
card ejection (but only 16-bit cards).

17 years agoFix typo in comment
imp [Fri, 16 Feb 2007 05:24:19 +0000 (05:24 +0000)]
Fix typo in comment

17 years agoDown to one MD document in the release documentation set now,
bmah [Fri, 16 Feb 2007 04:07:31 +0000 (04:07 +0000)]
Down to one MD document in the release documentation set now,
reword a note to reflect this fact.

17 years agoBacking out an earlier change. It seems harmless for NFS to miss the "force
mohans [Fri, 16 Feb 2007 03:46:55 +0000 (03:46 +0000)]
Backing out an earlier change. It seems harmless for NFS to miss the "force
unmount" flag, making the acquisition of the MNT_ILOCK in nfs_request() and
nfs_sigintr() unnecessary. Pointed out by tegge@.

17 years agoReflect that the hardware notes is now an MI document.
bmah [Fri, 16 Feb 2007 02:28:15 +0000 (02:28 +0000)]
Reflect that the hardware notes is now an MI document.

17 years agoReflect that the hardware notes article is now a MI document.
bmah [Fri, 16 Feb 2007 02:22:02 +0000 (02:22 +0000)]
Reflect that the hardware notes article is now a MI document.

17 years agoSwitch to unified MI hardware notes. The hardware notes are
bmah [Fri, 16 Feb 2007 02:20:11 +0000 (02:20 +0000)]
Switch to unified MI hardware notes.  The hardware notes are
now contained entirely in a single DocBook article, which has
information pertinent to all architectures.  This will improve
the maintainability of the hardware notes going forward by
reducing complexity of the build process, improving the visibiilty
of the architecture-specific text, and removing the need to generate
an extra version of the document for every architecture.

Remove all of the MD hardware notes stuff; all relevant content
has been folded into hardware/article.sgml.

No objections from: freebsd-doc@

17 years agoAdd support for auto-generating hardware notes entries in a
bmah [Fri, 16 Feb 2007 02:10:12 +0000 (02:10 +0000)]
Add support for auto-generating hardware notes entries in a
manner consistent with the new MI-style hardware notes document.

man2hwnotes.pl now defaults to generating entities for the MI-style
hardware notes (i.e. hardware/article.sgml).  A new -c
option causes it to generate entities for the older MD-style
hardware notes (i.e. hardware/common/dev.sgml).

The Makefile infrastructure supplies the -c option to man2hwnotes.pl
now unless the HWNOTES_MI Makefile variable is defined, so
compatiblity is preserved for hardware notes translations that
aren't converted to the new organization yet.  As translations
convert, they should define HWNOTES_MI in their hardware/Makefile.
When all the relevant translations catch up, the compatibility goop
in share/mk/doc.relnotes.mk and share/sgml/Makefile can be removed.

Thanks go to simon@ for help with the backwards compatiblity
mechanism.

Tested with: en_US.ISO8859-1, zh_CN.GB2312

17 years agoAdd bootverbose printfs to indicate which IDT vectors are assigned to MSI
jhb [Thu, 15 Feb 2007 22:22:57 +0000 (22:22 +0000)]
Add bootverbose printfs to indicate which IDT vectors are assigned to MSI
interrupts.

17 years agoMove vnode-to-file-handle translation from vfs_vptofh to vop_vptofh method.
pjd [Thu, 15 Feb 2007 22:08:35 +0000 (22:08 +0000)]
Move vnode-to-file-handle translation from vfs_vptofh to vop_vptofh method.
This way we may support multiple structures in v_data vnode field within
one file system without using black magic.

Vnode-to-file-handle should be VOP in the first place, but was made VFS
operation to keep interface as compatible as possible with SUN's VFS.
BTW. Now Solaris also implements vnode-to-file-handle as VOP operation.

VFS_VPTOFH() was left for API backward compatibility, but is marked for
removal before 8.0-RELEASE.

Approved by: mckusick
Discussed with: many (on IRC)
Tested with: ufs, msdosfs, cd9660, nullfs and zfs

17 years agoSupport AHCI chips where the ports are not consecutively numbered as in
sos [Thu, 15 Feb 2007 21:51:42 +0000 (21:51 +0000)]
Support AHCI chips where the ports are not consecutively numbered as in
some incarnations of the ICH8 chip.
Also fix the panic introduced by the last commit.

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.)