]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
15 years agoIntegrate the NDIS USB support code to CURRENT.
Weongyo Jeong [Sat, 27 Dec 2008 08:03:32 +0000 (08:03 +0000)]
Integrate the NDIS USB support code to CURRENT.

Now the NDISulator supports NDIS USB drivers that it've tested with
devices as follows:

  - Anygate XM-142 (Conexant)
  - Netgear WG111v2 (Realtek)
  - U-Khan UW-2054u (Marvell)
  - Shuttle XPC Accessory PN20 (Realtek)
  - ipTIME G054U2 (Ralink)
  - UNiCORN WL-54G (ZyDAS)
  - ZyXEL G-200v2 (ZyDAS)

All of them succeeded to attach and worked though there are still some
problems that it's expected to be solved.

To use NDIS USB support, you should rebuild and install ndiscvt(8) and
if you encounter a problem to attach please set `hw.ndisusb.halt' to
0 then retry.

I expect no changes of the NDIS code for PCI, PCMCIA devices.

Obtained from:  //depot/projects/ndisusb/...

15 years agoFix up after last commit:
Tom Rhodes [Sat, 27 Dec 2008 00:17:41 +0000 (00:17 +0000)]
Fix up after last commit:

Bump doc date;
Kill hard sentence breaks;
Fix commas by moving them off their own line.

15 years agostyle(9)
David E. O'Brien [Fri, 26 Dec 2008 22:55:38 +0000 (22:55 +0000)]
style(9)

15 years agoMake the sub-'argc' static to make it harder to overwrite thru a buffer
David E. O'Brien [Fri, 26 Dec 2008 22:54:53 +0000 (22:54 +0000)]
Make the sub-'argc' static to make it harder to overwrite thru a buffer
overflow.

15 years agoAdd some special handling for AD1986A codec:
Alexander Motin [Fri, 26 Dec 2008 22:47:11 +0000 (22:47 +0000)]
Add some special handling for AD1986A codec:

Disable some unneeded pathes in overcomplicated input mixer to help parser
to handle the rest better. This gives mic input boost control in some
configurations and just more predictable operation in others.

15 years agoClarify the behaviour of conditionals when dealing with comparisons.
Luigi Rizzo [Fri, 26 Dec 2008 22:31:45 +0000 (22:31 +0000)]
Clarify the behaviour of conditionals when dealing with comparisons.
In particular, point out that string comparison can only use != and ==
(how weird, given that the underlying call to strcmp returns more
information), that floating point values are correctly interpreted
as numbers, and that the left-hand side must be a variable expansion.

MFC after: 3 weeks

15 years agoWhen the geometry does not match the label, print out the values.
David E. O'Brien [Fri, 26 Dec 2008 20:27:32 +0000 (20:27 +0000)]
When the geometry does not match the label, print out the values.

15 years agoThis checkin addresses a couple of issues:
Qing Li [Fri, 26 Dec 2008 19:45:24 +0000 (19:45 +0000)]
This checkin addresses a couple of issues:
1. The "route" command allows route insertion through the interface-direct
   option "-iface". During if_attach(), an sockaddr_dl{} entry is created
   for the interface and is part of the interface address list. This
   sockaddr_dl{} entry describes the interface in detail. The "route"
   command selects this entry as the "gateway" object when the "-iface"
   option is present. The "arp" and "ndp" commands also interact with the
   kernel through the routing socket when adding and removing static L2
   entries. The static L2 information is also provided through the
   "gateway" object with an AF_LINK family type, similar to what is
   provided by the "route" command. In order to differentiate between
   these two types of operations, a RTF_LLDATA flag is introduced. This
   flag is set by the "arp" and "ndp" commands when issuing the add and
   delete commands. This flag is also set in each L2 entry returned by the
   kernel. The "arp" and "ndp" command follows a convention where a RTM_GET
   is issued first followed by a RTM_ADD/DELETE. This RTM_GET request fills
   in the fields for a "rtm" object, which is reinjected into the kernel by
   a subsequent RTM_ADD/DELETE command. The entry returend from RTM_GET
   is a prefix route, so the RTF_LLDATA flag must be specified when issuing
   the RTM_ADD/DELETE messages.

2. Enforce the convention that NET_RT_FLAGS with a 0 w_arg is the
   specification for retrieving L2 information. Also optimized the
   code logic.

Reviewed by:   julian

15 years agoRight align the CPU column header.
Rui Paulo [Fri, 26 Dec 2008 11:11:30 +0000 (11:11 +0000)]
Right align the CPU column header.

15 years agoo Fix grammar.
Maxim Konovalov [Fri, 26 Dec 2008 07:16:20 +0000 (07:16 +0000)]
o Fix grammar.

PR: bin/129938
Submitted by: Bruce Cran

15 years agoThe "tun?" dev need not be opened at all. One is allowed to perform
Qing Li [Thu, 25 Dec 2008 22:32:32 +0000 (22:32 +0000)]
The "tun?" dev need not be opened at all. One is allowed to perform
the following operations, e.g.:
1) ifconfig tun0 create
2) ifconfig tun0 10.1.1.1 10.1.1.2
3) route add -net 192.103.54.0/24 -iface tun0
4) ifconfig tun0 destroy
If cv wait on the TUN_CLOSED flag, then the last operation (4) will
block forever.

Revert the previous changes and fix the mtx_unlock() leak.

15 years agoFollow symlinks when deleting directories.
Florent Thoumie [Thu, 25 Dec 2008 16:59:35 +0000 (16:59 +0000)]
Follow symlinks when deleting directories.
Bump PKG_INSTALL_VER to 20081225 (Merry Christmas \o/).

PR: bin/54446
Submitted by: Andrea Barberio <insomniac@slackware.it>
MFC after: 1 month

15 years agoAdd package directory for 7.1-RELEASE in head as well as stable/7.
Florent Thoumie [Thu, 25 Dec 2008 15:44:10 +0000 (15:44 +0000)]
Add package directory for 7.1-RELEASE in head as well as stable/7.

15 years agoFollowing the recent security advisory, add a comment describing our
Robert Watson [Thu, 25 Dec 2008 11:32:38 +0000 (11:32 +0000)]
Following the recent security advisory, add a comment describing our
invariants and approach for protocol switch methods in protsw_init(),
and also some KASSERT's for non-domain init entries in protocol
switch tables: pru_abort and pru_send must both be implemented.

For now, leave those assertions #if 0'd, since there are a few
protocols that violate them in non-harmful ways.  Whether or not we
should enforce pru_abort being implemented for non-stream protocols
is an interesting question: currently abort is only invoked on stream
sockets in situations where un-accepted sockets must be abruptly
closed (i.e., close() on a listen socket with pending connections),
but in principle it is useful for datagram sockets and most datagram
socket types implement it.

MFC after: 3 weeks

15 years agoUpdate for the last API changes.
Alexander Motin [Thu, 25 Dec 2008 10:18:35 +0000 (10:18 +0000)]
Update for the last API changes.

15 years agong_tty(4) module updated to match the new TTY subsystem.
Alexander Motin [Thu, 25 Dec 2008 10:05:00 +0000 (10:05 +0000)]
ng_tty(4) module updated to match the new TTY subsystem.

15 years agoRejoin ng_tty module to the build.
Alexander Motin [Thu, 25 Dec 2008 09:32:20 +0000 (09:32 +0000)]
Rejoin ng_tty module to the build.

15 years agoDocument that kldunloadf can return EINVAL.
Tom Rhodes [Thu, 25 Dec 2008 09:15:31 +0000 (09:15 +0000)]
Document that kldunloadf can return EINVAL.

PR: 125639

15 years agoshave about 7% off the overhead of ng_ether by using per-hook
Julian Elischer [Thu, 25 Dec 2008 09:02:55 +0000 (09:02 +0000)]
shave about 7% off the overhead of ng_ether by using per-hook
receive data methods.

15 years agoRemove reference to unimplemented "-c" option.
Tom Rhodes [Thu, 25 Dec 2008 08:48:08 +0000 (08:48 +0000)]
Remove reference to unimplemented "-c" option.

PR: 119338

15 years agoHook up the ether_echo node and fix the man page
Julian Elischer [Thu, 25 Dec 2008 07:34:14 +0000 (07:34 +0000)]
Hook up the ether_echo node and fix the man page

15 years agoPrint a warning when blackhole and reject are used together.
Tom Rhodes [Thu, 25 Dec 2008 06:44:19 +0000 (06:44 +0000)]
Print a warning when blackhole and reject are used together.
Update arp.8 manual page syntax.

PR: 125896
Submitted by: Marc Olzheim <marcolz@stack.nl>
Approved by: sam

15 years agofix a silly bug that I missed a for-loop to initialize AL2230S PHY.
Weongyo Jeong [Thu, 25 Dec 2008 04:29:40 +0000 (04:29 +0000)]
fix a silly bug that I missed a for-loop to initialize AL2230S PHY.

Reported by:    Hans Petter Selasky <hselasky_at_c2i.net>

15 years ago- Close a race during which the open flag could be cleared but the tun_softc would...
Kip Macy [Thu, 25 Dec 2008 02:14:25 +0000 (02:14 +0000)]
- Close a race during which the open flag could be cleared but the tun_softc would still be referenced
  by adding a separate TUN_CLOSED flag that is set after tunclose is done referencing it.

- drop the tun_mtx after the flag check to avoid holding it across if_detach which can recurse in to
  if_tun.c

15 years agoAdd a trivial node to reflect ethernet frames to whence they came.
Julian Elischer [Thu, 25 Dec 2008 00:01:29 +0000 (00:01 +0000)]
Add a trivial node to reflect ethernet frames to whence they came.

MFC after: 1 month

15 years agoIn ugidfw(8), print the rule number and rule contents (as parsed and then
Robert Watson [Wed, 24 Dec 2008 22:40:13 +0000 (22:40 +0000)]
In ugidfw(8), print the rule number and rule contents (as parsed and then
regenerated in libugidfw) rather than simply printing that the rule was
added with only the number.  This makes ugidfw(8) behave a bit more like
ipfw(8), and also means that the administrator sees how the rule was
interpreted once uids/gids/etc were processed.

Obtained from: TrustedBSD Project

15 years agoAdd macro RB_FOREACH_SAFE(), which accepts an additional argument
Bruce M Simpson [Wed, 24 Dec 2008 19:57:22 +0000 (19:57 +0000)]
Add macro RB_FOREACH_SAFE(), which accepts an additional argument
specifying a temporary tree node pointer. It may be used in a
similar way to the *_SAFE() macros in <sys/queue.h>.

15 years agoApply various fixes:
Tom Rhodes [Wed, 24 Dec 2008 11:12:21 +0000 (11:12 +0000)]
Apply various fixes:

Silence mdoc(7) warnings;
Xref correct manual pages;
Point user to the ddb.8 manual page.

PR: 129398
Submitted by: gavin

15 years agoUpdate the description of -k, output files appear to
Tom Rhodes [Wed, 24 Dec 2008 10:08:39 +0000 (10:08 +0000)]
Update the description of -k, output files appear to
be left regardless of an error, signal, or complete
run.

PR: 69861
Submitted by: vlad902@gmail.com

15 years agoFix a typo higter->higher.
Tom Rhodes [Wed, 24 Dec 2008 09:17:30 +0000 (09:17 +0000)]
Fix a typo higter->higher.

Spotted by: ganbold

15 years agomust pack structures for architectures like arm
Sam Leffler [Wed, 24 Dec 2008 06:59:07 +0000 (06:59 +0000)]
must pack structures for architectures like arm

Reviewed by: thompsa

15 years agoFix missed unlock and reference drop of lle
Kip Macy [Wed, 24 Dec 2008 05:31:26 +0000 (05:31 +0000)]
Fix missed unlock and reference drop of lle

Found by: pho

15 years agoGarbage collect 'fflags'.
David E. O'Brien [Wed, 24 Dec 2008 03:07:19 +0000 (03:07 +0000)]
Garbage collect 'fflags'.

15 years agoavoid lock recursion by deferring the link check until after LLE lock is dropped
Kip Macy [Wed, 24 Dec 2008 01:08:18 +0000 (01:08 +0000)]
avoid lock recursion by deferring the link check until after LLE lock is dropped

15 years agoChange message severity level from WARN to INFO. This should reduce
Maksim Yevmenkin [Wed, 24 Dec 2008 00:00:52 +0000 (00:00 +0000)]
Change message severity level from WARN to INFO. This should reduce
amount of messages sent to syslog

MFC after: 1 week

15 years agoOnly set registers if irqs are enabled
Robert Noland [Tue, 23 Dec 2008 22:53:57 +0000 (22:53 +0000)]
Only set registers if irqs are enabled

Approved by: kib
Obtained from: drm git

15 years agoUpdate bmake glue for changes in 9.4.3
Doug Barton [Tue, 23 Dec 2008 22:51:10 +0000 (22:51 +0000)]
Update bmake glue for changes in 9.4.3

15 years agoUpdates for version 9.4.3
Doug Barton [Tue, 23 Dec 2008 22:50:39 +0000 (22:50 +0000)]
Updates for version 9.4.3

15 years agoMerge from vendor/bind9/dist as of the 9.4.3 import
Doug Barton [Tue, 23 Dec 2008 22:47:56 +0000 (22:47 +0000)]
Merge from vendor/bind9/dist as of the 9.4.3 import

15 years agoAdd support for the FPA floating-point format on ARM. The
Marcel Moolenaar [Tue, 23 Dec 2008 22:20:59 +0000 (22:20 +0000)]
Add support for the FPA floating-point format on ARM. The
FPA floating-point format is identical to the VFP format,
but is always stored in big-endian.
Introduce _IEEE_WORD_ORDER to describe the byte-order of
the FP representation.

Obtained from: Juniper Networks, Inc

15 years agoDo not KASSERT when vp->v_dd is NULL. Only directories which have had ".."
Joe Marcus Clarke [Tue, 23 Dec 2008 20:43:42 +0000 (20:43 +0000)]
Do not KASSERT when vp->v_dd is NULL.  Only directories which have had ".."
looked up would have v_dd set to a non-NULL value.  This fixes a panic
seen when running installworld on a diskless system with a separate /usr
file system.

Submitted by: cracauer
Approved by: kib

15 years agoAdd a new program, ether_reflect, which is useful in testing ethernet
George V. Neville-Neil [Tue, 23 Dec 2008 20:25:04 +0000 (20:25 +0000)]
Add a new program, ether_reflect, which is useful in testing ethernet
devices and switches.

15 years agoFix a bug in the man page where we were not showing the correct flags
George V. Neville-Neil [Tue, 23 Dec 2008 20:07:51 +0000 (20:07 +0000)]
Fix a bug in the man page where we were not showing the correct flags
in the explanation. Several of the flags were -i, since it was a copy/paste
operation.

15 years agoKeep the hold on the vnode during VOP_VPTOCNP() call, allowing the vop
Konstantin Belousov [Tue, 23 Dec 2008 20:04:31 +0000 (20:04 +0000)]
Keep the hold on the vnode during VOP_VPTOCNP() call, allowing the vop
implementation to drop vnode lock, if needed.

Reported and tested by: pho

15 years agoRemove CALLOUT_RETURNUNLOCKED from the callouts, there is no reason for them to
Andrew Thompson [Tue, 23 Dec 2008 19:59:21 +0000 (19:59 +0000)]
Remove CALLOUT_RETURNUNLOCKED from the callouts, there is no reason for them to
drop the lock for us.

15 years agoAdd the mctest program and description to the README.
George V. Neville-Neil [Tue, 23 Dec 2008 19:46:12 +0000 (19:46 +0000)]
Add the mctest program and description to the README.

15 years agoThe lib/bind directory has its own configure script. Update accordingly.
Doug Barton [Tue, 23 Dec 2008 19:19:45 +0000 (19:19 +0000)]
The lib/bind directory has its own configure script.  Update accordingly.

15 years agoVendor import of BIND 9.4.3
Doug Barton [Tue, 23 Dec 2008 19:18:41 +0000 (19:18 +0000)]
Vendor import of BIND 9.4.3

15 years agoUpdate copyrights and comments as of 9.4.3 (no functional changes)
Doug Barton [Tue, 23 Dec 2008 19:15:04 +0000 (19:15 +0000)]
Update copyrights and comments as of 9.4.3 (no functional changes)

15 years agoVendor import of BIND 9.4.3
Doug Barton [Tue, 23 Dec 2008 18:35:21 +0000 (18:35 +0000)]
Vendor import of BIND 9.4.3

15 years agoMinor tweak to reflect my actual process.
Doug Barton [Tue, 23 Dec 2008 18:31:38 +0000 (18:31 +0000)]
Minor tweak to reflect my actual process.

15 years agoThis directory was added to src/contrib/bind9 prior to the svn move so
Doug Barton [Tue, 23 Dec 2008 18:30:32 +0000 (18:30 +0000)]
This directory was added to src/contrib/bind9 prior to the svn move so
that BIND 9.4.1 could compile and run on the ARM platform, but is not
part of the vendor source.  It will remain in src/contrib/bind9 since
it is a local modification.

15 years agoNetgroup database can be centralized via NIS, list it.
Tom Rhodes [Tue, 23 Dec 2008 18:00:33 +0000 (18:00 +0000)]
Netgroup database can be centralized via NIS, list it.
Add a missing comma and bump doc date.

PR: 127602
Submitted by: Dmitry Sivachenko <mitya@yandex-team.ru>

15 years agoMake signal handler safer.
David E. O'Brien [Tue, 23 Dec 2008 17:57:17 +0000 (17:57 +0000)]
Make signal handler safer.

Submitted by: Jaakko Heinonen <jh@saunalahti.fi>

15 years agomove IXP4XX EHCI bus shim to the usb directory and rename
Sam Leffler [Tue, 23 Dec 2008 17:40:02 +0000 (17:40 +0000)]
move IXP4XX EHCI bus shim to the usb directory and rename

15 years agoAdd "crit" to the list of keywords.
Tom Rhodes [Tue, 23 Dec 2008 17:39:24 +0000 (17:39 +0000)]
Add "crit" to the list of keywords.

PR: 126934

15 years agoName the bus mutex by the controller name, this allows each bus to be
Andrew Thompson [Tue, 23 Dec 2008 17:36:25 +0000 (17:36 +0000)]
Name the bus mutex by the controller name, this allows each bus to be
distinguished in lock profiling, etc.

15 years agopr_pfh has been gone from protosw since r120386 (more than 5 years ago).
Bjoern A. Zeeb [Tue, 23 Dec 2008 16:54:00 +0000 (16:54 +0000)]
pr_pfh has been gone from protosw since r120386 (more than 5 years ago).

MFC after: 4 weeks

15 years agoRemove long unused netinet/ipprotosw.h (basically since r82884).
Bjoern A. Zeeb [Tue, 23 Dec 2008 16:52:03 +0000 (16:52 +0000)]
Remove long unused netinet/ipprotosw.h (basically since r82884).

Discussed with: rwatson
MFC after: 4 weeks

15 years agoCheck for ipprotosw.h more precisely.
Bjoern A. Zeeb [Tue, 23 Dec 2008 16:49:07 +0000 (16:49 +0000)]
Check for ipprotosw.h more precisely.
It hasn't been needed for more than 5 years, since r120386.

MFC after: 4 weeks

15 years agoAdd missing newlines to flags tags of CPU topology, for prettier
Ivan Voras [Tue, 23 Dec 2008 16:19:59 +0000 (16:19 +0000)]
Add missing newlines to flags tags of CPU topology, for prettier
output.

Reviewed by: jeff (original version)
Approved by: gnn (mentor) (original version)

15 years agoFix up handling of Intel G4X chips some more.
Robert Noland [Tue, 23 Dec 2008 16:16:30 +0000 (16:16 +0000)]
Fix up handling of Intel G4X chips some more.

Note that you need at least xf86-video-intel 2.4.3 for this to work.
The G4X doesn't put the GATT into the same area of stolen memory
as all the other chips and older versions of the driver didn't
handle that properly.

Tested by: ganbold
Approved by: kib
MFC after: 2 weeks

15 years agoClear busy state on the pages which are after the one that failed the bind
Konstantin Belousov [Tue, 23 Dec 2008 16:04:33 +0000 (16:04 +0000)]
Clear busy state on the pages which are after the one that failed the bind
attempt.

Reported and tested by: ganbold
Reviewed by: rnoland
MFC after: 2 weeks

15 years agoMove another block of ASUS events to devd/asus.conf that were missed in r186249
Andrew Thompson [Tue, 23 Dec 2008 15:47:31 +0000 (15:47 +0000)]
Move another block of ASUS events to devd/asus.conf that were missed in r186249

15 years agoNote that when shutdown is run without options, it will place
Tom Rhodes [Tue, 23 Dec 2008 15:00:03 +0000 (15:00 +0000)]
Note that when shutdown is run without options, it will place
the system into single user mode at the time specified.

PR: 129765

15 years agoOrganize Conexant codecs.
Alexander Motin [Tue, 23 Dec 2008 14:58:08 +0000 (14:58 +0000)]
Organize Conexant codecs.
Add CX20561 (Hermosa) codec ID.

15 years agoDocument the "-o large" option.
Tom Rhodes [Tue, 23 Dec 2008 13:35:26 +0000 (13:35 +0000)]
Document the "-o large" option.

PR: 129792

15 years agoAdd support for the HP 4470C scanner.
Remko Lodder [Tue, 23 Dec 2008 13:09:17 +0000 (13:09 +0000)]
Add support for the HP 4470C scanner.

Note that there is no working backend (or at least
that is mentioned in the PR ticket) but the device
is now supported on our end.

PR: 117205
Submitted by: Artem Naluzhnyy <tut at nhamon dot com dot ua>
MFC after: 1 week

15 years agoAdd support for the MaxSream USB test carrier.
Remko Lodder [Tue, 23 Dec 2008 12:44:18 +0000 (12:44 +0000)]
Add support for the MaxSream USB test carrier.

PR: 117546
Submitted by: Daniel J. O'Connor <darius at midget dot dons dot net
dot au>
MFC after: 1 week

15 years agoAdd support for 2 EVDO devices.
Remko Lodder [Tue, 23 Dec 2008 12:15:21 +0000 (12:15 +0000)]
Add support for 2 EVDO devices.

PR: 119150
Submitted by: lioux
MFC after: 1 week

15 years agoClose the read side of the pipe to self when exiting.
Joseph Koshy [Tue, 23 Dec 2008 12:08:06 +0000 (12:08 +0000)]
Close the read side of the pipe to self when exiting.

15 years agoAdd quirk for the Storcase InfoStation 12bay
Remko Lodder [Tue, 23 Dec 2008 09:11:05 +0000 (09:11 +0000)]
Add quirk for the Storcase InfoStation 12bay
SATA to FC SAN.

PR: 129858
Submitted by: Nick Triantos <nick-freebsd at triantos dot com>
MFC after: 1 week

15 years agoIf conversion from UTF8 fails, don't mark Unicode text as available.
Tim Kientzle [Tue, 23 Dec 2008 05:01:43 +0000 (05:01 +0000)]
If conversion from UTF8 fails, don't mark Unicode text as available.

Submitted by: Michihiro NAKAJIMA
MFC after: 30 days

15 years agoNPE cleanups needed for ancillary drivers (e.g. crypto acceleration):
Sam Leffler [Tue, 23 Dec 2008 04:51:46 +0000 (04:51 +0000)]
NPE cleanups needed for ancillary drivers (e.g. crypto acceleration):
o check feature bits when probing NPE ethernet support
o move firmware loading logic from if_npe to core npe support
o allow multiple refs to core NPE driver
o while here fix hw.npe.debug tunable path

15 years agokill NPE_PORTS_MAX, it's not used and likely will not be
Sam Leffler [Tue, 23 Dec 2008 04:49:01 +0000 (04:49 +0000)]
kill NPE_PORTS_MAX, it's not used and likely will not be

15 years agoFill in feature control support:
Sam Leffler [Tue, 23 Dec 2008 04:48:27 +0000 (04:48 +0000)]
Fill in feature control support:
o add definitions for more bits, for masking out IXP465-specific bits,
  and %b format string
o add ixp4xx_read_feature_bits to retrieve the mask of valid features
  (aka fuse bits)
o add cpu_is_ixp42x() macro
o print feature bits at boot

15 years agoadd IXP465 and generic IXP425 definition
Sam Leffler [Tue, 23 Dec 2008 04:46:13 +0000 (04:46 +0000)]
add IXP465 and generic IXP425 definition

15 years agoo enable TT and big-endian MMIO
Sam Leffler [Tue, 23 Dec 2008 04:44:23 +0000 (04:44 +0000)]
o enable TT and big-endian MMIO
o force a reset before ehci_init to get byte-select setup

LS, FS, and HS devices now work on the Cambria board

15 years agoo add Transaction Translator support (still missing ISOC xfers)
Sam Leffler [Tue, 23 Dec 2008 04:42:10 +0000 (04:42 +0000)]
o add Transaction Translator support (still missing ISOC xfers)
o add EHCI_SCFLG_BIGEMMIO flag to force big-endian byte-select to be
  set in USBMODE
o split reset work into new public routine ehci_reset so bus shim drivers
  can force big-endian byte-select before ehci_init

15 years agoDon't create a bogus ARP entry for 0.0.0.0.
Qing Li [Tue, 23 Dec 2008 03:33:32 +0000 (03:33 +0000)]
Don't create a bogus ARP entry for 0.0.0.0.

15 years agoPrevent cross-site forgery attacks on ftpd(8) due to splitting
Colin Percival [Tue, 23 Dec 2008 01:23:09 +0000 (01:23 +0000)]
Prevent cross-site forgery attacks on ftpd(8) due to splitting
long commands into multiple requests. [08:12]

Avoid calling uninitialized function pointers in protocol switch
code. [08:13]

Merry Christmas everybody...

Approved by: so (cperciva)
Approved by: re (kensmith)
Security: FreeBSD-SA-08:12.ftpd, FreeBSD-SA-08:13.protosw

15 years agoAdd a basic manual page for elf2aout.
Tom Rhodes [Tue, 23 Dec 2008 01:22:57 +0000 (01:22 +0000)]
Add a basic manual page for elf2aout.
Alter Makefile and remove NO_MAN.

PR: 109975

15 years agoAdd 9 more Analog Devices codec names.
Alexander Motin [Mon, 22 Dec 2008 22:40:17 +0000 (22:40 +0000)]
Add 9 more Analog Devices codec names.

15 years agoFix typo for udav include.
Alfred Perlstein [Mon, 22 Dec 2008 21:37:06 +0000 (21:37 +0000)]
Fix typo for udav include.
Add link to the u3g2 driver.

15 years agoImplement a new feature for the "-m" option: if the month number is
Garrett Wollman [Mon, 22 Dec 2008 21:22:42 +0000 (21:22 +0000)]
Implement a new feature for the "-m" option: if the month number is
followed by 'f' or 'p', use the following or preceding month of that
number, respectively.  Document this.  Also includes other minor
grammatical and punctuation fixes to the manual page (capitalize
Easter, etc.).

MFC after: 1 month

15 years agoOnly show the pathname in verbose output as intended in the last commit (r172862).
Andrew Thompson [Mon, 22 Dec 2008 19:52:21 +0000 (19:52 +0000)]
Only show the pathname in verbose output as intended in the last commit (r172862).

15 years agoMake preparations for resurrecting shared/read locks on vm maps:
Alan Cox [Mon, 22 Dec 2008 17:32:52 +0000 (17:32 +0000)]
Make preparations for resurrecting shared/read locks on vm maps:

mac_proc_vm_revoke_recurse() requests a read lock on the vm map at the start
but does not handle failure by vm_map_lock_upgrade() when it seeks to modify
the vm map.  At present, this works because all lock request on a vm map are
implemented as exclusive locks.  Thus, vm_map_lock_upgrade() is a no-op that
always reports success.  However, that is about to change, and
proc_vm_revoke_recurse() will require substantial modifications to handle
vm_map_lock_upgrade() failures.  For the time being, I am changing
mac_proc_vm_revoke_recurse() to request a write lock on the vm map at the
start.

Approved by: rwatson
MFC after: 3 months

15 years agoFix a comment to reflect what really happens.
Scott Long [Mon, 22 Dec 2008 16:58:47 +0000 (16:58 +0000)]
Fix a comment to reflect what really happens.

15 years ago- According to comments in OpenBSD, E{2,4}50 tend to have fragile
Marius Strobl [Mon, 22 Dec 2008 15:34:06 +0000 (15:34 +0000)]
- According to comments in OpenBSD, E{2,4}50 tend to have fragile
  firmware versions which wedge when using the OFW test service,
  so given that we don't really depend on SUNW,stop-self just nuke
  it altogether instead of risking problems.
- At least Fire V880 have a small hardware glitch which causes the
  reception of IDR_NACKs for CPUs we actually haven't tried to send
  an IPI to, even not as part of the initial try. According to tests
  this apparently can be safely ignored though, so just return if
  checking for the individual IDR_NACKs indicates no outstanding
  dispatch. Serializing the sending of IPIs between MD and MI code
  by the combined usage of smp_ipi_mtx makes no difference to this
  phenomenon. [1]
- Provide relevant debugging bits already with the initial panic
  in case of problems with the IPI dispatch, which would have
  allowed to diagnose the above problem without a specially built
  kernel.
- In case of cheetah_ipi_selected() base the delay we wait for
  other CPUs which also might want to dispatch IPIs on the total
  amount of CPUs instead of just the number of CPUs we let this
  CPU send IPIs to because in the worst case all CPUs also want
  to IPI us at the same time.

Reported and access for extensive tests provided by: Beat Gaetzi [1]

15 years agoo Correct a database name: "group" not "groups".
Maxim Konovalov [Mon, 22 Dec 2008 13:36:15 +0000 (13:36 +0000)]
o Correct a database name: "group" not "groups".

PR: docs/129853
Submitted by: Denis Barov
MFC after: 1 week

15 years agoCorrect variable name in comment.
Bjoern A. Zeeb [Mon, 22 Dec 2008 12:54:52 +0000 (12:54 +0000)]
Correct variable name in comment.

MFC after: 4 weeks

15 years agoSimilar to the INET case, do not destroy the nd6 entries for
Qing Li [Mon, 22 Dec 2008 07:11:15 +0000 (07:11 +0000)]
Similar to the INET case, do not destroy the nd6 entries for
interface addresses until those addresses are removed. I already
made the patch in INET but forgot to bring the code over for
INET6.

15 years agoProvide a condition variable to delay the cloned interface
Qing Li [Mon, 22 Dec 2008 01:56:56 +0000 (01:56 +0000)]
Provide a condition variable to delay the cloned interface
destroy operation until the referenced clone device has
been closed by the process properly. The behavior is now
consistently with the previous release.

Reviewed by:    Kip Macy

15 years agoSince we don't request reset for rlphy(4), the link state 'UP'
Pyun YongHyeon [Mon, 22 Dec 2008 00:53:47 +0000 (00:53 +0000)]
Since we don't request reset for rlphy(4), the link state 'UP'
event from mii(4) may not be delivered if valid link was already
established. To address the issue, check current link state after
driving MII_TICK. This should fix a regression introduced in
r184245.

PR: kern/129647

15 years agoSince we don't request reset for rlphy(4), the link state 'UP'
Pyun YongHyeon [Mon, 22 Dec 2008 00:46:22 +0000 (00:46 +0000)]
Since we don't request reset for rlphy(4), the link state 'UP'
event from mii(4) may not be delivered if valid link was already
established. To address the issue, check current link state after
driving MII_TICK. This should fix a regression introduced in
r185753 on fast ethernet controllers.

Reported by: csjp, Bruce Cran < bruce <> cran DOT org DOT uk >
Tested by: csjp, Bruce Cran (initial version)

15 years agoConvert DRM_[DEBUG,ERROR,INFO] macros to c99 __VA_ARGS__.
Robert Noland [Sun, 21 Dec 2008 22:32:01 +0000 (22:32 +0000)]
Convert DRM_[DEBUG,ERROR,INFO] macros to c99 __VA_ARGS__.

Approved by: kib

15 years agoDeal with 0 length args...
Robert Noland [Sun, 21 Dec 2008 22:30:37 +0000 (22:30 +0000)]
Deal with 0 length args...

Approved by: kib

15 years agoFix AGP_DEBUG macro to use c99 __VA_ARGS__ and build if enabled.
Robert Noland [Sun, 21 Dec 2008 22:00:39 +0000 (22:00 +0000)]
Fix AGP_DEBUG macro to use c99 __VA_ARGS__ and build if enabled.

Approved by: kib

15 years agoRevert r185891.
Ed Schouten [Sun, 21 Dec 2008 21:54:01 +0000 (21:54 +0000)]
Revert r185891.

In r185891 I removed the newlines from messages written to /dev/console,
because it made startup messages from rc-scripts harder to read. This,
unfortunately, causes the kernel message that is printed after a
non-terminated log message to be concatenated.

This could be fixed, but on short term it's better to just revert the
change.

Reported by: Jaakko Heinonen <jh saunalahti fi>

15 years agoSet PTS_FINISHED before waking up any threads.
Ed Schouten [Sun, 21 Dec 2008 21:16:57 +0000 (21:16 +0000)]
Set PTS_FINISHED before waking up any threads.

Inside ptsdrv_{in,out}wakeup() we call KNOTE_LOCKED() to wake up any
kevent(2) users. Because the kqueue handlers are executed synchronously,
we must set PTS_FINISHED before calling ptsdrv_{in,out}wakeup().

Discovered by: nork