]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
15 years agoMerge OpenSSL 0.9.8k into head.
simon [Sun, 14 Jun 2009 19:45:16 +0000 (19:45 +0000)]
Merge OpenSSL 0.9.8k into head.

Approved by: re

15 years agoEnable PRINTF_BUFR_SIZE on i386 and amd64 by default.
ed [Sun, 14 Jun 2009 18:01:35 +0000 (18:01 +0000)]
Enable PRINTF_BUFR_SIZE on i386 and amd64 by default.

In the past there have been some reports of PRINTF_BUFR_SIZE not
functioning correctly. Instead of having garbled console messages, we
should just see whether the issues are still there and analyze them.

Approved by: re

15 years agoUnlock process lock when return error from getrobustlist call.
dchagin [Sun, 14 Jun 2009 17:53:55 +0000 (17:53 +0000)]
Unlock process lock when return error from getrobustlist call.

Tested by: Alexander Best <alexbestms at math uni-muenster de>
Approved by: kib (mentor)
MFC after: 3 days

15 years agoWhen the KOBJMETHOD() macro was updated, it resulted in the
rmacklem [Sun, 14 Jun 2009 17:33:46 +0000 (17:33 +0000)]
When the KOBJMETHOD() macro was updated, it resulted in the
compiler doing argument type checking. Since the prototypes in
kgss_if.h used the generic gss_ctx_id_t for the context argument,
but the functions in sys/kgssapi/krb5/krb5_mech.c used the
KerberosV specific context argument, the file would no longer build.
This patch fixes it so it will build by replacing the argument with
a gss_ctx_id_t one and setting a local "struct krb5_context *" variable
to it for use by the function.

Reviewed by: dfr
Approved by: kib (mentor)

15 years agoAdd an optional callback function that will be invoked when a per-CPU
bz [Sun, 14 Jun 2009 17:15:18 +0000 (17:15 +0000)]
Add an optional callback function that will be invoked when a per-CPU
queue was drained.  It will never fire for a directly dispatched packet.

You will most likely never want to use this for any ordinary netisr usage
and you will never blame netisr in case you try to use it and it does
not work as expected.

Reviewed by: rwatson

15 years agoFix the staircase issue properly this time.
ed [Sun, 14 Jun 2009 17:00:35 +0000 (17:00 +0000)]
Fix the staircase issue properly this time.

Even though I thought this bug was somewhere in the TTY layer, it turns
out init(8) doesn't make sure /dev/console is opened initially properly.
I've added revoke() to two pieces of code:

- death(): Apart from killing the gettys on shutdown, this doesn't
  guarantee the TTY to be closed immediately.
- runshutdown(): Just like setctty(), we should revoke /dev/console.
  Applications like syslogd may have file descriptors to the console.

15 years agoAdd myself.
nork [Sun, 14 Jun 2009 16:48:17 +0000 (16:48 +0000)]
Add myself.

Approved by: imp (mentor)

15 years agoAdd tests for r194127 and r194128.
jilles [Sun, 14 Jun 2009 16:21:51 +0000 (16:21 +0000)]
Add tests for r194127 and r194128.

Approved by: ed (mentor)

15 years agoAllow "make update" work with SVN, if you take the time to set up your
markm [Sun, 14 Jun 2009 15:16:24 +0000 (15:16 +0000)]
Allow "make update" work with SVN, if you take the time to set up your
/etc/make.conf properly.

Only one SCM?
I do not think this is right.
Now I have fixed it.

15 years agoImprove the description a bit and add a reference to vlan(4).
marius [Sun, 14 Jun 2009 14:46:20 +0000 (14:46 +0000)]
Improve the description a bit and add a reference to vlan(4).

15 years agoInstead of only being able to compare two terminal definitions, now
edwin [Sun, 14 Jun 2009 13:46:32 +0000 (13:46 +0000)]
Instead of only being able to compare two terminal definitions, now
also be able to print information about...

- length of the terminal capabilties
- dump of one terminal definition
- relationship overview for a terminal definition

15 years agoRename since it has will have more capabilities than just comparing of
edwin [Sun, 14 Jun 2009 13:43:07 +0000 (13:43 +0000)]
Rename since it has will have more capabilities than just comparing of
two termcap entries.

15 years agoInclude <sys/module.h> for modfind().
ed [Sun, 14 Jun 2009 12:47:27 +0000 (12:47 +0000)]
Include <sys/module.h> for modfind().

Submitted by: Pawel Worach

15 years agoInclude <camlib.h> for cam_path_string().
ed [Sun, 14 Jun 2009 12:46:34 +0000 (12:46 +0000)]
Include <camlib.h> for cam_path_string().

Submitted by: Pawel Worach

15 years agoInclude <stdio.h> for asprintf().
ed [Sun, 14 Jun 2009 12:45:48 +0000 (12:45 +0000)]
Include <stdio.h> for asprintf().

Submitted by: Pawel Worach

15 years agoInclude <sys/wait.h> and <signal.h> for wait() and kill().
ed [Sun, 14 Jun 2009 12:44:09 +0000 (12:44 +0000)]
Include <sys/wait.h> and <signal.h> for wait() and kill().

15 years agoFix missing includes of <string.h>, to silence some compiler warnings.
ed [Sun, 14 Jun 2009 12:42:06 +0000 (12:42 +0000)]
Fix missing includes of <string.h>, to silence some compiler warnings.

Submitted by: Pawel Worach

15 years agoRemove a line break leaving a function return type attached to the old
bz [Sun, 14 Jun 2009 12:11:15 +0000 (12:11 +0000)]
Remove a line break leaving a function return type attached to the old
function declaration bottom rather than the new function declaration
start.

15 years agoAdd more entries to the Unicode-to-CP437 table.
ed [Sun, 14 Jun 2009 12:04:34 +0000 (12:04 +0000)]
Add more entries to the Unicode-to-CP437 table.

Characters between 0x07 and 0x0d are now also mapped, which means we can
display almost 256 different characters. Also remap certain types of
dashes and quotes, which means we can finally read our manual pages
without red question marks in them.

Submitted by: Christoph Mallon

15 years agoFix type of lowaddr variable.
mav [Sun, 14 Jun 2009 07:34:21 +0000 (07:34 +0000)]
Fix type of lowaddr variable.

15 years agoAdd missing passthrough printing entries for VT100 and workalikes
edwin [Sun, 14 Jun 2009 07:02:50 +0000 (07:02 +0000)]
Add missing passthrough printing entries for VT100 and workalikes
in /etc/termcap:

    VT100 spec indicates that passthrough printing can be enabled
    by sending ESC[5i and disabled by sending ESC[4i These entries
    should be listed as po and pf in /etc/termcap, but are absent.
    See http://www.vt100.net/docs/vt102-ug/chapter5.html#S5.5.2.23

PR: conf/71549
Submitted by: Andrew Webster <andrew@pubnix.net>
MFC after: 1 week

15 years agoTermcap updates for screen and linux console:
edwin [Sun, 14 Jun 2009 06:48:31 +0000 (06:48 +0000)]
Termcap updates for screen and linux console:

screen entry: F11-F20 keys added
linux entry: F10-F20 keys added, ACS line graphics added

PR: kern/108899
Submitted by: Joseph Terner<jtsn@gmx.de>
MFC after: 1 week

15 years agoAdd -m and -t options.
brian [Sun, 14 Jun 2009 06:47:21 +0000 (06:47 +0000)]
Add -m and -t options.

PR: 129554
Submitted by: gavin
MFC after: 3 weeks

15 years agoRemove HOME= - this has surprised me several times in the past.
brian [Sun, 14 Jun 2009 06:37:19 +0000 (06:37 +0000)]
Remove HOME= - this has surprised me several times in the past.

PR: 132135
Submitted by: Craig Leres
MFC after: 3 weeks

15 years agoDon't set $PATH - if the caller doesn't have /bin and /usr/bin in their
brian [Sun, 14 Jun 2009 06:09:33 +0000 (06:09 +0000)]
Don't set $PATH - if the caller doesn't have /bin and /usr/bin in their
PATH then they can expect things to break.

PR: 135417
MFC after: 3 weeks

15 years agoAdd missing terminal definition for Wyse 120 in termcap
edwin [Sun, 14 Jun 2009 04:03:18 +0000 (04:03 +0000)]
Add missing terminal definition for Wyse 120 in termcap

Although the PR contains also the definitions of the Wyse 60, they
are not copied into it since there are already definition for them
in the termcap file since 1997.

Also, the PR didn't use the :tc=xxx: feature, so I've imploded them.

PR: conf/81882
Submitted by: Meister des Chaos <meister@netz00.com>
MFC after: 1 week

15 years agoAdd missing termcap entry for rxvt-unicode.
edwin [Sun, 14 Jun 2009 03:34:54 +0000 (03:34 +0000)]
Add missing termcap entry for rxvt-unicode.

The termcap database does not have an entry for rxvt-unicode.
This means that programs that need an entry such as vi fail
to work when connecting via ssh using this terminal emulator.

The added data is not the same as the PR submitted by Richard, it
uses the :tc=xxx: option to inherit everything from rxvt-mono.

PR: conf/117323
Submitted by: Richard Bradshaw <richard.bradshaw@blueyonder.co.uk>
MFC after: 1 week

15 years agoFix typo... bad imp.
imp [Sun, 14 Jun 2009 03:32:52 +0000 (03:32 +0000)]
Fix typo...  bad imp.

15 years agoTool to compare two (roughly similar) termcap entries (like rxvt-mono
edwin [Sun, 14 Jun 2009 03:29:14 +0000 (03:29 +0000)]
Tool to compare two (roughly similar) termcap entries (like rxvt-mono
and rxvt-unicode) so see which fields should don't need to be copied
if the :tc=xxx: option is used.

15 years agoAfter Marcel's change to DEFAULTS, we were bringing in a bogus copy of
imp [Sun, 14 Jun 2009 02:58:56 +0000 (02:58 +0000)]
After Marcel's change to DEFAULTS, we were bringing in a bogus copy of
uart_8250.  Remove it here since the UART on the ADM5120 isn't the
typical 16550: its completely different.

15 years agoFormatting nit.
imp [Sun, 14 Jun 2009 02:55:07 +0000 (02:55 +0000)]
Formatting nit.

15 years agoMissed style fixes.
marius [Sun, 14 Jun 2009 00:15:26 +0000 (00:15 +0000)]
Missed style fixes.

15 years agoFix style.
marius [Sun, 14 Jun 2009 00:05:38 +0000 (00:05 +0000)]
Fix style.

15 years agonote abi change for IEEE80211_IOC_STA_INFO
sam [Sat, 13 Jun 2009 23:44:56 +0000 (23:44 +0000)]
note abi change for IEEE80211_IOC_STA_INFO

15 years agoRev IEEE80211_IOC_STA_INFO abi: ni_flags grew from 16 bits to 32
sam [Sat, 13 Jun 2009 23:43:00 +0000 (23:43 +0000)]
Rev IEEE80211_IOC_STA_INFO abi: ni_flags grew from 16 bits to 32
bits but isi_state did not follow; expand it to 32 bits and pad to
maintain alignment.  Note this is an incompatible change that
requires rebuilding of user applications.

Submitted by: rpaulo, cbzimmer, avatar

15 years agopurge HAL_TXSTAT_ALTRATE; you can figure this out by checking ts_finaltsi
sam [Sat, 13 Jun 2009 23:36:54 +0000 (23:36 +0000)]
purge HAL_TXSTAT_ALTRATE; you can figure this out by checking ts_finaltsi
and it cannot be used with MCS rate codes

15 years ago- Also probe DP83865, which is an is an ultra low power version
marius [Sat, 13 Jun 2009 23:27:04 +0000 (23:27 +0000)]
- Also probe DP83865, which is an is an ultra low power version
  of the DP83861 and DP83891.
- Reset the PHY during attach so it's in a known state.
- Add a comment describing why we hardwire 10baseT support in
  the BMSR.
- Always explicitly set IFM_HDX for half-duplex. [1]

Obtained from: OpenBSD [1]
MFC after: 2 weeks

15 years agoWhen a Solaris10 client does an NFS mount using krb5i or krb5p, the
rmacklem [Sat, 13 Jun 2009 23:16:40 +0000 (23:16 +0000)]
When a Solaris10 client does an NFS mount using krb5i or krb5p, the
server would crash because the Solaris10 client would attempt to use
Sun's NFSACL protocol, which FreeBSD doesn't support. When the server
generated the error reply via svcerr_noprog(), it would cause a crash
because it would try and wrap a NULL reply. According to RFC2203, no
wrapping is required for error cases. This one line change avoids
wrapping of NULL replies.

Reviewed by: dfr
Approved by: kib (mentor)

15 years agoupdate backend_changed to reflect .m prototype
kmacy [Sat, 13 Jun 2009 21:55:08 +0000 (21:55 +0000)]
update backend_changed to reflect .m prototype

15 years agoAvoid leaving unnecessary waiting shells in many forms of sh -c COMMAND.
jilles [Sat, 13 Jun 2009 21:17:45 +0000 (21:17 +0000)]
Avoid leaving unnecessary waiting shells in many forms of sh -c COMMAND.

This change only affects strings passed to -c, when the -s
option is not used.

The approach is to check if there may be additional data
in the string after parsing each command. If there is none,
use the EV_EXIT flag so that a fork may be omitted in
specific cases.

If there are empty lines after the command, the check will
not see the end and forks will not be omitted. The same
thing seems to happen in bash.

Example:
  sh -c 'ps lT'
No longer shows a shell process waiting for ps to finish.

PR: bin/113860
Reviewed by: stefanf
Approved by: ed (mentor)

15 years agoDon't skip forking for an external command if any traps are active.
jilles [Sat, 13 Jun 2009 21:10:41 +0000 (21:10 +0000)]
Don't skip forking for an external command if any traps are active.

Example:
  sh -c '(trap "echo trapped" EXIT; sleep 3)'
now correctly prints "trapped".

With this check, it is no longer necessary to check for -T
explicitly in that case.

This is a useful bugfix by itself and also important because I plan to
skip forking more often.

PR: bin/113860 (part of)
PR: bin/74404 (part of)
Reviewed by: stefanf
Approved by: ed (mentor)

15 years agoEliminate an unnecessary clearing of a page's dirty bits in
alc [Sat, 13 Jun 2009 20:58:12 +0000 (20:58 +0000)]
Eliminate an unnecessary clearing of a page's dirty bits in
phys_pager_getpages().

15 years agoEliminate unnecessary variables.
alc [Sat, 13 Jun 2009 20:21:08 +0000 (20:21 +0000)]
Eliminate unnecessary variables.

15 years agoCorrect the method of waking the page daemon when the number of allocated
alc [Sat, 13 Jun 2009 18:35:29 +0000 (18:35 +0000)]
Correct the method of waking the page daemon when the number of allocated
pv entries surpasses the high water mark.  The problem was that the page
daemon would only be awakened the first time that the high water mark was
surpassed.  (The variable "pagedaemon_waken" is a non-working vestige of
FreeBSD 4.x, in which it was external and reset by the page daemon whenever
it ran.  This reset allowed subsequent wakeups by the pv entry allocator.)

15 years agoRename the host-related prison fields to be the same as the host.*
jamie [Sat, 13 Jun 2009 15:39:12 +0000 (15:39 +0000)]
Rename the host-related prison fields to be the same as the host.*
parameters they represent, and the variables they replaced, instead of
abbreviated versions of them.

Approved by: bz (mentor)

15 years agoUse getcredhostuuid instead of accessing the prison directly.
jamie [Sat, 13 Jun 2009 15:35:22 +0000 (15:35 +0000)]
Use getcredhostuuid instead of accessing the prison directly.

Approved by: bz (mentor)

15 years agoClobber "cc" instead of using volatile.
ed [Sat, 13 Jun 2009 14:30:08 +0000 (14:30 +0000)]
Clobber "cc" instead of using volatile.

Submitted by: Christoph Mallon

15 years agoRemove "extern" from function prototypes, and fix some (but not all)
des [Sat, 13 Jun 2009 14:12:55 +0000 (14:12 +0000)]
Remove "extern" from function prototypes, and fix some (but not all)
style(9) violations.

MFC after: 1 week

15 years agoClobber "cc" instead of using volatile; remove obsolete register keyword.
ed [Sat, 13 Jun 2009 14:00:10 +0000 (14:00 +0000)]
Clobber "cc" instead of using volatile; remove obsolete register keyword.

Submitted by: Christoph Mallon

15 years agoSimplify the inline assembler (and correct potential error) of pte_load_store().
ed [Sat, 13 Jun 2009 13:56:06 +0000 (13:56 +0000)]
Simplify the inline assembler (and correct potential error) of pte_load_store().

Submitted by: Christoph Mallon

15 years agoWrap some macros that needed wrapping.
des [Sat, 13 Jun 2009 13:54:03 +0000 (13:54 +0000)]
Wrap some macros that needed wrapping.

MFC after: 1 week

15 years agoSync termcap.small with main termcap; add xterm entry for libteken
edwin [Sat, 13 Jun 2009 13:35:18 +0000 (13:35 +0000)]
Sync termcap.small with main termcap; add xterm entry for libteken

PR: conf/135530
Submitted by: Alex Kozlov <spam@rm-rf.kiev.ua>
MFC after: 1 week

15 years agoRemove the static from int hardlink_check_uid.
bz [Sat, 13 Jun 2009 13:09:20 +0000 (13:09 +0000)]
Remove the static from int hardlink_check_uid.
There is an external use in the opensolaris code.

I am not sure how this ever worked but I have seen two reports of:
link_elf: symbol hardlink_check_uid undefined
lately.

Reported by: Scott Ullrich (sullrich gmail.com), pfsense
Reported by: Mister Olli (mister.olli googlemail.com)

15 years agoRemove casts from {c,m,re}alloc() and simplify sizeof().
des [Sat, 13 Jun 2009 13:07:56 +0000 (13:07 +0000)]
Remove casts from {c,m,re}alloc() and simplify sizeof().

MFC after: 1 week

15 years agoMake the proof-of-concept UTF-8 support in Syscons less useless.
ed [Sat, 13 Jun 2009 12:23:31 +0000 (12:23 +0000)]
Make the proof-of-concept UTF-8 support in Syscons less useless.

Add a small Unicode-to-CP437 remapping table to at least demonstrate
that the terminal emulator is perfectly capable of handling UTF-8. This
will of course break if the user loads a different font map, but it at
least allows people to give it a try.

I can now see the box drawing in dialog(1) and the arrows in mutt(1)
correctly.

15 years agoCorrect entry of vendor ID 0x0d8c. It's C-Media, not ABC.
sanpei [Sat, 13 Jun 2009 11:18:47 +0000 (11:18 +0000)]
Correct entry of vendor ID 0x0d8c.  It's C-Media, not ABC.

Obtained from: NetBSD usbdevs rev.1.418

15 years agoFix Book-E/MPC85XX build. Some prototypes were wrong and got revealed with
raj [Sat, 13 Jun 2009 08:57:04 +0000 (08:57 +0000)]
Fix Book-E/MPC85XX build. Some prototypes were wrong and got revealed with
the recent kobj signature checking.

15 years agoMake variables static where appropriate.
thompsa [Sat, 13 Jun 2009 04:56:45 +0000 (04:56 +0000)]
Make variables static where appropriate.

Found by: cscout

15 years agonscd builds cleanly at WARNS level 3.
des [Sat, 13 Jun 2009 01:23:34 +0000 (01:23 +0000)]
nscd builds cleanly at WARNS level 3.

MFC after: 1 week

15 years agoExplain to the compiler why the aliasing we're doing is OK.
des [Sat, 13 Jun 2009 01:22:56 +0000 (01:22 +0000)]
Explain to the compiler why the aliasing we're doing is OK.

MFC after: 1 week

15 years agoFix a large number of signed - unsigned comparison warnings.
des [Sat, 13 Jun 2009 00:56:17 +0000 (00:56 +0000)]
Fix a large number of signed - unsigned comparison warnings.

MFC after: 1 week

15 years agoChange hashtable_index_t to unsigned.
des [Sat, 13 Jun 2009 00:54:52 +0000 (00:54 +0000)]
Change hashtable_index_t to unsigned.
Generate prototypes for our hash table.

MFC after: 1 week

15 years ago#include the right header so we get prototypes for our own functions.
des [Sat, 13 Jun 2009 00:46:07 +0000 (00:46 +0000)]
#include the right header so we get prototypes for our own functions.

MFC after: 1 week

15 years agoFurther #include cleanup.
des [Sat, 13 Jun 2009 00:43:56 +0000 (00:43 +0000)]
Further #include cleanup.

MFC after: 1 week

15 years agoAdd support for labels derived from GPT metadata.
ivoras [Sat, 13 Jun 2009 00:27:03 +0000 (00:27 +0000)]
Add support for labels derived from GPT metadata.

Approved by: gnn (mentor)
Reviewed by: pjd
PR: 128398
Submitted by: Marius Nuennerich < marius at nuenneri.ch >

15 years agoMake the debugging macros expand to (void)0 instead of simply nothing
des [Sat, 13 Jun 2009 00:13:44 +0000 (00:13 +0000)]
Make the debugging macros expand to (void)0 instead of simply nothing
when debugging is turned off.
Rename debugging functions due to namespace violation.

MFC after: 1 week

15 years agoAdd counterparts to getcredhostname:
jamie [Sat, 13 Jun 2009 00:12:02 +0000 (00:12 +0000)]
Add counterparts to getcredhostname:
getcreddomainname, getcredhostuuid, getcredhostid

Suggested by: rmacklem
Approved by: bz

15 years ago#include cleanup
des [Sat, 13 Jun 2009 00:06:52 +0000 (00:06 +0000)]
#include cleanup

MFC after: 1 week

15 years agoFix typo in cons25l7 definition in etc/termcap.small and share/termcap
edwin [Fri, 12 Jun 2009 23:43:19 +0000 (23:43 +0000)]
Fix typo in cons25l7 definition in etc/termcap.small and share/termcap

    There is a minor typo in the cons25l7 (':' instead of '|') entry
    in src/etc/termcap.small that causes syscons to complain about
    bogus characters in /etc/termcap.db.

PR: conf/132777
Submitted by: Nikos Ntarmos <ntarmos@cs.uoi.gr>
MFC after: 1 week

15 years agoProvide correct prototypes for functions with no arguments.
des [Fri, 12 Jun 2009 23:39:05 +0000 (23:39 +0000)]
Provide correct prototypes for functions with no arguments.

MFC after: 1 week

15 years ago_nss_cache_cycle_prevention_function doesn't actually need to be a function,
des [Fri, 12 Jun 2009 23:27:31 +0000 (23:27 +0000)]
_nss_cache_cycle_prevention_function doesn't actually need to be a function,
it just needs to have external linkage.

MFC after: 1 week

15 years agoFix printing of some wide-characters by iswprint() on ja_JP.eucJP
edwin [Fri, 12 Jun 2009 23:15:36 +0000 (23:15 +0000)]
Fix printing of some wide-characters by iswprint() on ja_JP.eucJP
and ja_JP.SJIS locale

    The iswprint() function does not return non-zero if used for
    some wide-character that it code was 0x824f-0x8258 on ja_JP.SJIS
    and 0xa3b0-0xa3b9 on ja_JP.eucJP locale.  But those are right
    Japanese wide-character code.

PR: conf/124511
Submitted by: Michihiro NAKAJIMA <ggcueroad@gmail.com>
MFC after: 1 week

15 years agoAdd various new keymaps:
edwin [Fri, 12 Jun 2009 22:53:49 +0000 (22:53 +0000)]
Add various new keymaps:

addition: danish syscons keymap with accents

    There are two danish keymap files shipping with FreeBSD:
    danish.iso.kbd and danish.cp865.kbd.  None of these support
    accents by use of dead keys, which isn't crucial to Danes but
    often used.  This is a new keymap, danish.iso.acc.kbd, that's
    based on danish.iso.kbd with the accent part taken from
    german.iso.acc.kbd.

pl_PL.dvorak keymap for syscons.

    Polish dvorak keymap by Robert Sebastian Gerus (17:05 01-07-2006
    CEST) <arachnist@gmail.com>. Based on pl_PL.ISO-8859-2 and
    us.dvorak keymaps.

British English Dvorak keymap for syscons.

    Attached is a Dvorak keymap for British English, with a
    pound-sterling symbol on on Shift-3 and some other characters
    in different places to the US Dvorak keymap. I've based the
    layout on the X.org GB keymap, Dvorak variant.

PR: conf/72978, kern/99692, conf/117257
Submitted by: Kim Norgaard <jasen@jasen.dk>, Robert Sebastian Gerus <arachnist@gmail.com>, "Benjamin A'Lee" <bma@subvert.org.uk>
MFC after: 1 week

15 years agoFix MALTA build; some prototypes were wrong and blew up when kobj method
jmallett [Fri, 12 Jun 2009 22:49:35 +0000 (22:49 +0000)]
Fix MALTA build; some prototypes were wrong and blew up when kobj method
signature checking was turned on.

15 years agoRevert my previous change, because it reintroduces an old regression.
ed [Fri, 12 Jun 2009 21:21:17 +0000 (21:21 +0000)]
Revert my previous change, because it reintroduces an old regression.

Because our rc scripts also open the /etc/ttyv* nodes, it revokes the
console, preventing startup messages from being displayed.

I really have to think about this. Maybe we should just give the console
its own TTY and let it build on top of other TTYs. I'm still not sure
what to do with input handling there.

15 years agoUpdate the inline version of vn_get_ino() for ".." lookups to match the
jhb [Fri, 12 Jun 2009 21:19:57 +0000 (21:19 +0000)]
Update the inline version of vn_get_ino() for ".." lookups to match the
recentish changes to vn_get_ino().

MFC after: 1 week

15 years agoGarbage collect an extern for a non-existent variable.
bz [Fri, 12 Jun 2009 20:50:28 +0000 (20:50 +0000)]
Garbage collect an extern for a non-existent variable.
While here let the comment end in a '.' and mark the #endif of _KERNEL.

Reviewed by: rwatson (as part of a larger patch)

15 years agoMove the kernel option FLOWTABLE chacking from the header file to the
bz [Fri, 12 Jun 2009 20:46:36 +0000 (20:46 +0000)]
Move the kernel option FLOWTABLE chacking from the header file to the
actual implementation.
Remove the accessor functions for the compiled out case, just returning
"unavail" values. Remove the kernel conditional from the header file as
it is no longer needed, only leaving the externs.
Hide the improperly virtualized SYSCTL/TUNABLE for the flowtable size
under the kernel option as well.

Reviewed by: rwatson

15 years agoPrevent yet another staircase effect bug in the console device.
ed [Fri, 12 Jun 2009 20:29:55 +0000 (20:29 +0000)]
Prevent yet another staircase effect bug in the console device.

Even though I thought I fixed the staircase issue (and I was no longer
able to reproduce it), I got some reports of the issue still being
there. It turns out the staircase effect still occurred when
/dev/console was kept open while killing the getty on the same TTY
(ttyv0).

For some reason I can't figure out how the old TTY code dealt with that,
so I assume the issue has always been there. I only exposed it more by
merging consolectl with ttyv0, which means that the issue was present,
even on systems without a serial console.

I'm now marking the console device as being closed when closing the
regular TTY device node. This means that when the getty shuts down,
init(8) will open /dev/console, which means the termios attributes will
always be reset in this case.

15 years agoAdd a #include <sys/jail.h> so that it builds when
rmacklem [Fri, 12 Jun 2009 20:18:08 +0000 (20:18 +0000)]
Add a #include <sys/jail.h> so that it builds when
options KGSSAPI is specified in the kernel configuration.

Approved by: kib (mentor)

15 years agoMove the memory layout definitions and logic from mvreg.h to mvwin.h
marcel [Fri, 12 Jun 2009 20:00:38 +0000 (20:00 +0000)]
Move the memory layout definitions and logic from mvreg.h to mvwin.h
so that it isn't exposured unless needed. In particular this means
that it's easier to tune the memory layout based on board details.
While here, remove inclusion of <machine/intr.h> from mvreg.h. This
also contains exposure to SoC specifics in MI drivers, because NIRQ
depends on the SoC.

15 years ago- Note that these interfaces require <sys/user.h> for the structure
jhb [Fri, 12 Jun 2009 18:13:34 +0000 (18:13 +0000)]
- Note that these interfaces require <sys/user.h> for the structure
  definitions.
- Note that these functions return NULL on failure.

MFC after: 3 days

15 years ago- Remove semicolon that should not have been there.
lulf [Fri, 12 Jun 2009 16:37:53 +0000 (16:37 +0000)]
- Remove semicolon that should not have been there.

Submitted by: rdivacky
MFC after: 1 week

15 years agoCleanup claim/release interface code, which is specific to libusb v0.1. Remove
thompsa [Fri, 12 Jun 2009 16:07:06 +0000 (16:07 +0000)]
Cleanup claim/release interface code, which is specific to libusb v0.1. Remove
claim and release interface support from libusb v2.0, because it is not useful.

Submitted by: Hans Petter Selasky

15 years agoChange ums_probe() so it does not need to fetch the usb_interface_descriptor.
thompsa [Fri, 12 Jun 2009 16:04:32 +0000 (16:04 +0000)]
Change ums_probe() so it does not need to fetch the usb_interface_descriptor.

Submitted by: Hans Petter Selasky

15 years agoCheck for a keyboard HID report in addition to the interface class so devices
thompsa [Fri, 12 Jun 2009 16:03:38 +0000 (16:03 +0000)]
Check for a keyboard HID report in addition to the interface class so devices
such as the Yubikey attach.

Submitted by: Hans Petter Selasky
Reported by: Jeremy Faulkner

15 years agoFree the correct memory pointer.
thompsa [Fri, 12 Jun 2009 16:00:12 +0000 (16:00 +0000)]
Free the correct memory pointer.

Submitted by: Tim Borgeaud (via HPS)

15 years agoMinor device side improvement. Make sure a not complete state gets paired with
thompsa [Fri, 12 Jun 2009 15:58:55 +0000 (15:58 +0000)]
Minor device side improvement. Make sure a not complete state gets paired with
a complete state in device side mode for the default control endpoint.

Submitted by: Hans Petter Selasky

15 years agoAdded support for NAT-Traversal (RFC 3948) in IPsec stack.
vanhu [Fri, 12 Jun 2009 15:44:35 +0000 (15:44 +0000)]
Added support for NAT-Traversal (RFC 3948) in IPsec stack.

Thanks to (no special order) Emmanuel Dreyfus (manu@netbsd.org), Larry
Baird (lab@gta.com), gnn, bz, and other FreeBSD devs, Julien Vanherzeele
(julien.vanherzeele@netasq.com, for years of bug reporting), the PFSense
team, and all people who used / tried the NAT-T patch for years and
reported bugs, patches, etc...

X-MFC: never

Reviewed by: bz
Approved by: gnn(mentor)
Obtained from: NETASQ

15 years agogdb: make 'thread apply all bt' always work on all threads
avg [Fri, 12 Jun 2009 14:27:50 +0000 (14:27 +0000)]
gdb: make 'thread apply all bt' always work on all threads

even if some appear to have (partially) corrupted stack traces.
E.g. kernel crashdumps typically have stack weirdness at
userland-kernel boundary.

Obtained from: vendor/upstream (CVS rev 1.118 of stack.c)
Reviewed by: emaste
Approved by: jhb

15 years agoUpdate the content of the nfsv4.4 man page to reflect the final
rmacklem [Fri, 12 Jun 2009 14:18:13 +0000 (14:18 +0000)]
Update the content of the nfsv4.4 man page to reflect the final
choice of variable names for rc.conf and option name for the
experimental server.
Also replace the inaccurate description of the nfsv4 root lines
in /etc/exports, mostly with a reference to exports(5).

Approved by: kib (mentor)

15 years agoFrench translations for keyboards keymaps descriptions
edwin [Fri, 12 Jun 2009 11:16:13 +0000 (11:16 +0000)]
French translations for keyboards keymaps descriptions

    French translations for keyboards keymaps descriptions with
    some corrections on existing translations in the
    /usr/share/syscons/keymaps/INDEX.keymaps

PR: conf/71767
Submitted by: Henri Michelon <hmichelon@e-cml.org>
MFC after: 1 week

15 years agoRemove unnecessary #ifdef lines and code.
zec [Fri, 12 Jun 2009 09:31:14 +0000 (09:31 +0000)]
Remove unnecessary #ifdef lines and code.

Approved by: julian (mentor)

15 years agoAssign ng_eiface nodes a netgraph name on instantiation, in a way which
zec [Fri, 12 Jun 2009 09:20:31 +0000 (09:20 +0000)]
Assign ng_eiface nodes a netgraph name on instantiation, in a way which
is consistent with the current behavior of ng_iface, i.e. borrow the
same naming code from ng_iface.c.

Approved by: julian (mentor)

15 years agoKOBJMETHOD: throw the switch on kobj method signature checking
avg [Fri, 12 Jun 2009 09:05:23 +0000 (09:05 +0000)]
KOBJMETHOD: throw the switch on kobj method signature checking

It seems that current code should pass the check.
This commit should not lead to any changes in compiled code.
From now on a warning shall be produced if kobj method implementation
function has a mismatching signature.

Verified by: md5
Reviewed by: imp
Approved by: jhb (mentor)

15 years agoLast (hopefully) attempt to fix both parallel build and
ariff [Fri, 12 Jun 2009 07:39:10 +0000 (07:39 +0000)]
Last (hopefully) attempt to fix both parallel build and
cyclic dependencies, in a simmilar fashion to kmod.mk.

Discussed with: bde, bz

15 years agodevclass_find_free_unit(xxx, 0) is identical to -1 for most
imp [Fri, 12 Jun 2009 00:07:09 +0000 (00:07 +0000)]
devclass_find_free_unit(xxx, 0) is identical to -1 for most
applications (including this one).  Switch to it since the former is a
racy interface.

15 years agopjd has requested that I keep the tunable as zfs_prefetch_disable to minimize gratuitous
kmacy [Thu, 11 Jun 2009 22:24:08 +0000 (22:24 +0000)]
pjd has requested that I keep the tunable as zfs_prefetch_disable to minimize gratuitous
differences with Opensolaris' ZFS

Sorry for the churn

15 years agoRe-add the send queue tunable for people who do not use buffering.
gnn [Thu, 11 Jun 2009 21:32:26 +0000 (21:32 +0000)]
Re-add the send queue tunable for people who do not use buffering.

Reviewed by: jhb
MFC after: 3 days

15 years agosyslog.conf(5): correct example
avg [Thu, 11 Jun 2009 18:42:31 +0000 (18:42 +0000)]
syslog.conf(5): correct example

security.* and console.* are moved out of ftpd program block

Approved by: jhb (mentor)
MFC after: 2 weeks