]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoSync usb.h with NetBSD, apart from usb_device_info.speed, which
joe [Wed, 2 Jan 2002 20:16:53 +0000 (20:16 +0000)]
Sync usb.h with NetBSD, apart from usb_device_info.speed, which
requires logic changes.  For now leave it as usb_device_info.lowspeed.
It will get addressed when the usb.c code is sync'd.

22 years agoClarify the "show" ipfw(8) command.
yar [Wed, 2 Jan 2002 20:16:15 +0000 (20:16 +0000)]
Clarify the "show" ipfw(8) command.

PR: docs/31263
Permitted by: luigi

22 years agoCopy the sample `SECURITY CONSIDERATIONS' section from sec-doc.7.
chris [Wed, 2 Jan 2002 19:56:57 +0000 (19:56 +0000)]
Copy the sample `SECURITY CONSIDERATIONS' section from sec-doc.7.
This will be trimmed as the FreeBSD Security Architecture document
is fleshed out and committed.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years agoFix a typo: wierd -> weird
yar [Wed, 2 Jan 2002 19:46:14 +0000 (19:46 +0000)]
Fix a typo: wierd -> weird

22 years agoDocument a new option: -force.
yar [Wed, 2 Jan 2002 19:01:53 +0000 (19:01 +0000)]
Document a new option: -force.

22 years agoAllow weird characters in usernames if an admin persists on that:
yar [Wed, 2 Jan 2002 18:54:40 +0000 (18:54 +0000)]
Allow weird characters in usernames if an admin persists on that:
a) Convert all the remaining older Perl system() calls to the new,
   more secure LIST format so they are robust to whitespace and
   shell metacharacters in their arguments.
b) Add a new option: -force, which allows adding usernames containing
   characters that are otherwise illegal.

PR: bin/22860 bin/31049

22 years ago1. Implement an optimization for pmap_remove() and pmap_protect(): if a
tmm [Wed, 2 Jan 2002 18:49:20 +0000 (18:49 +0000)]
1. Implement an optimization for pmap_remove() and pmap_protect(): if a
   substantial fraction of the number of entries of tte's in the tsb
   would need to be looked up, traverse the tsb instead. This is crucial
   in some places, e.g. when swapping out a process, where a certain
   pmap_remove() call would take very long time to complete without this.
2. Implement pmap_qenter_flags(), which will become used later
3. Reactivate the instruction cache flush done when mapping as executable.
   This is required e.g. when executing files via NFS, but is known to
   cause problems on UltraSPARC-IIe CPU's. If you have such a CPU, you
   will need to comment this call out for now.

Submitted by: jake (3)

22 years agoRevert part of the last commit. UIPROTO_MASS_BBB was renamed to
joe [Wed, 2 Jan 2002 18:36:46 +0000 (18:36 +0000)]
Revert part of the last commit.  UIPROTO_MASS_BBB was renamed to
UIPROTO_MASS_BULK _after_ the previous change was made in NetBSD.

22 years agoSync with NetBSD:
joe [Wed, 2 Jan 2002 18:28:45 +0000 (18:28 +0000)]
Sync with NetBSD:

date: 2000/02/29 21:37:01;  author: augustss;  state: Exp;
Distinguish between device and interface classes.
(I finally found a document that said that they were different.)

22 years agoCorrect the defintion of struct ofw_upa_regs, and use it instead of
tmm [Wed, 2 Jan 2002 18:27:13 +0000 (18:27 +0000)]
Correct the defintion of struct ofw_upa_regs, and use it instead of
struct ofw_nexus_reg. Implement UPA device memory management in the
nexus driver.
Adapt the psycho driver to these changes, and do some minor cleanup work
while being there.

22 years agoClose a window of time during early boot in which an interrupt would
tmm [Wed, 2 Jan 2002 18:21:16 +0000 (18:21 +0000)]
Close a window of time during early boot in which an interrupt would
cause a panic.

Reported and tested (in another version) by: Jamey Wood <Jamey.Wood@Sun.COM>

22 years agoUpdate copyright years.
bmah [Wed, 2 Jan 2002 17:27:58 +0000 (17:27 +0000)]
Update copyright years.

22 years agoRemove the special casing for NetBSD and OpenBSD now that they
joe [Wed, 2 Jan 2002 16:45:15 +0000 (16:45 +0000)]
Remove the special casing for NetBSD and OpenBSD now that they
appear to do what we do.

22 years agoNetBSD have introduced usb_proc_ptr for us (they'll be needing it soon too)
joe [Wed, 2 Jan 2002 16:33:53 +0000 (16:33 +0000)]
NetBSD have introduced usb_proc_ptr for us (they'll be needing it soon too)
to hide the distinction between struct proc and struct thread.

22 years agoBe consistent in what restriction is imposed on a username
yar [Wed, 2 Jan 2002 16:19:14 +0000 (16:19 +0000)]
Be consistent in what restriction is imposed on a username
in the code and what the warning message reads.

22 years agoAdd myself.
okazaki [Wed, 2 Jan 2002 16:02:51 +0000 (16:02 +0000)]
Add myself.

22 years agoForbid adding duplicate users with the name of "0".
yar [Wed, 2 Jan 2002 15:49:49 +0000 (15:49 +0000)]
Forbid adding duplicate users with the name of "0".
Previously, it was possible due to 0 and "0" being
equivalent in Perl.

22 years agoParse atq(1) output correctly under various (but not all perhaps)
yar [Wed, 2 Jan 2002 15:35:33 +0000 (15:35 +0000)]
Parse atq(1) output correctly under various (but not all perhaps)
locale settings: allow any date and time separator characters.

22 years agoIn remove_at_jobs():
yar [Wed, 2 Jan 2002 15:24:10 +0000 (15:24 +0000)]
In remove_at_jobs():
Don't print "Removing at jobs" if there are no jobs to remove.
Add a whitespace before "done." so the output looks better.

22 years agoFix some places where the macro L0IFPIUNIT was used instead of the correct
gj [Wed, 2 Jan 2002 12:55:04 +0000 (12:55 +0000)]
Fix some places where the macro L0IFPIUNIT was used instead of the correct
form L0IFPI2UNIT. This could result in a panic if the user tried to
trace using isdntrace(8). I fixed this locally but forgot to commit it.

Reminded by: "Wittig, Christoph" <wc@medianet-world.de>

22 years agoIf the CLK_VAL register is 0 bits wide, the system does not support
msmith [Wed, 2 Jan 2002 10:53:10 +0000 (10:53 +0000)]
If the CLK_VAL register is 0 bits wide, the system does not support
CPU throttling, so don't do some bogus math to check it.

22 years agoImprove pam_unix/opie related ftpd comment even more
ache [Wed, 2 Jan 2002 09:51:33 +0000 (09:51 +0000)]
Improve pam_unix/opie related ftpd comment even more

22 years agoi4bisppp also needs net/if_spppsubr.c.
gj [Wed, 2 Jan 2002 09:15:18 +0000 (09:15 +0000)]
i4bisppp also needs net/if_spppsubr.c.

MFC after: 1 month

22 years ago o Properly check the file descriptor passed to aio_cancel(2). (Previously,
alc [Wed, 2 Jan 2002 07:04:38 +0000 (07:04 +0000)]
 o Properly check the file descriptor passed to aio_cancel(2).  (Previously,
   no out-of-bounds check was performed on the file descriptor.)
 o Eliminate some excessive white space from aio_cancel(2).

22 years agoAdd support for decoding Buffer objects that contain PnP/ACPI resource
msmith [Wed, 2 Jan 2002 07:01:34 +0000 (07:01 +0000)]
Add support for decoding Buffer objects that contain PnP/ACPI resource
streams.  Since the output is bulky, it's controlled by the '-r'
option.

Document this in the manpage, and clean up some awkward English a
little.

22 years agoAdd ifdef sparc64.
jake [Wed, 2 Jan 2002 06:54:18 +0000 (06:54 +0000)]
Add ifdef sparc64.

22 years agoRemove broken attempt to compile libc's malloc source directly; this
kbyanc [Wed, 2 Jan 2002 06:42:34 +0000 (06:42 +0000)]
Remove broken attempt to compile libc's malloc source directly; this
allows this tool to compile again. Albeit, now to test a new malloc
implementation one has to install the new libc which may have bad
consequences (i.e. if the new malloc implementation were buggy).

Add logic to workaround malloc's current behaviour of returning an
invalid non-NULL pointer for 0 byte allocation requests; this prevents the
tool from coring during the NOPS loop.

Add $FreeBSD$ tags.

22 years agoInclude definition for struct namecache (it isn't exported via any header)
kbyanc [Wed, 2 Jan 2002 06:25:39 +0000 (06:25 +0000)]
Include definition for struct namecache (it isn't exported via any header)
so this builds again. Replace hardcoded name length magic with NAME_MAX.

22 years agoMerge the changes in the English version (1.242 --> 1.243).
kuriyama [Wed, 2 Jan 2002 04:19:37 +0000 (04:19 +0000)]
Merge the changes in the English version (1.242 --> 1.243).

22 years agoMF4S: Minor wordsmithing and markup fixes.
bmah [Wed, 2 Jan 2002 01:36:07 +0000 (01:36 +0000)]
MF4S:  Minor wordsmithing and markup fixes.

22 years agoPermit NFS filesystems to be forcibly unmounted when the server is
iedowse [Wed, 2 Jan 2002 00:41:26 +0000 (00:41 +0000)]
Permit NFS filesystems to be forcibly unmounted when the server is
down, even if there are hung processes and the mount is non-
interruptible.

This works by having nfs_unmount call a new function nfs_nmcancelreqs()
in the FORCECLOSE case. It scans the list of outstanding requests
and marks as interrupted any requests belonging to the specified
mount. Then it waits up to 30 seconds for all requests to terminate.
A few other changes are necessary to support this:
- Unconditionally set a socket timeout so that even hard mounts
  are guaranteed to occasionally check the R_SOFTTERM flag on
  requests. For hard mounts this flag can only be set by
  nfs_nmcancelreqs().
- Reject requests on a mount that is currently being unmounted.
- Never grant the receive lock to a request that has been cancelled.

This should also avoid an old problem where a forced NFS unmount
could cause a crash; it occurred when a VOP on an unlocked vnode
(usually VOP_GETATTR) was in progress at the time of the forced
unmount.

22 years agoConnect user trap code to the build.
jake [Tue, 1 Jan 2002 21:59:53 +0000 (21:59 +0000)]
Connect user trap code to the build.

22 years agoAdd libc side of user trap handling.
jake [Tue, 1 Jan 2002 21:58:32 +0000 (21:58 +0000)]
Add libc side of user trap handling.
Add support for handling floating point disabled traps mostly in userland
for the simple single threaded case.  Not yet enabled by default.
Implement __sparc_utrap_install as specified by the sparc abi.

22 years agoPrint parm6 too in the !KTR_EXTEND case.
jake [Tue, 1 Jan 2002 21:47:38 +0000 (21:47 +0000)]
Print parm6 too in the !KTR_EXTEND case.

22 years ago1. Lower the poll timeout for the ofw console driver from hz / 50 to hz / 4.
jake [Tue, 1 Jan 2002 21:45:30 +0000 (21:45 +0000)]
1. Lower the poll timeout for the ofw console driver from hz / 50 to hz / 4.
   This gives a bit of a sluggish console, but it prevents the console from
   getting stuck if we poll too fast, as well as other badness on certain
   machines.
2. Fix a test for != 0 that should have been > 0.

Noticed by: Jamey Wood <Jamey.Wood@Sun.COM> and myself
Submitted by: tmm (2)

22 years agoCorrectly identify the cpu in certain ultra 1s.
jake [Tue, 1 Jan 2002 21:26:57 +0000 (21:26 +0000)]
Correctly identify the cpu in certain ultra 1s.

Noticed by: Jamey Wood <Jamey.Wood@Sun.COM>
Submitted by: tmm

22 years agoDefine __ASM__ so that libc will know not to define C things.
jake [Tue, 1 Jan 2002 21:21:05 +0000 (21:21 +0000)]
Define __ASM__ so that libc will know not to define C things.

22 years agoAdd a define for the fp restore soft trap type.
jake [Tue, 1 Jan 2002 21:19:46 +0000 (21:19 +0000)]
Add a define for the fp restore soft trap type.
Only declare C things if __ASM__ is not defined.

22 years agoAdd constants needed by user trap code.
jake [Tue, 1 Jan 2002 21:17:14 +0000 (21:17 +0000)]
Add constants needed by user trap code.

22 years agoA global is used even though the value is
wosch [Tue, 1 Jan 2002 21:16:10 +0000 (21:16 +0000)]
A global is used even though the value is
passed through the function call.

Submitted by: "Crist J . Clark" <cjc@freebsd.org>

22 years agoEnable virtual caching for kernel pages. When we enabled virtual caching
jake [Tue, 1 Jan 2002 21:14:58 +0000 (21:14 +0000)]
Enable virtual caching for kernel pages.  When we enabled virtual caching
for certain user pages, stores to kernel pages would not update the
affected cache lines, which would sometimes cause the wrong data to be
returned for loads from kernel pages.  This was especially fatal when
the addresses affected held the kernel stack pointer, and a random
value was loaded into it.
Fix a harmless off by one error in a dcache_inval_phys call.

22 years agoAdd some more info to traces.
jake [Tue, 1 Jan 2002 21:04:27 +0000 (21:04 +0000)]
Add some more info to traces.
Fix a potential race in setting up the per-cpu pointer if the special
restore fails on return to user mode fails and we need to trap back
into the kernel to fault in more stack.
Remove debug code.

22 years agoEnsure that the syscall trap vector is properly aligned.
jake [Tue, 1 Jan 2002 21:00:29 +0000 (21:00 +0000)]
Ensure that the syscall trap vector is properly aligned.

22 years agoImplement sysarch(SPARC_UTRAP_INSTALL).
jake [Tue, 1 Jan 2002 20:57:51 +0000 (20:57 +0000)]
Implement sysarch(SPARC_UTRAP_INSTALL).

Forgot this file in last commit.

22 years agoImplement user trap delivery as specified by the sparc abi. This provides
jake [Tue, 1 Jan 2002 20:56:28 +0000 (20:56 +0000)]
Implement user trap delivery as specified by the sparc abi.  This provides
an efficient way for the kernel to bounce certain mundane traps back to
userland for handling there.  A user trap handler returns directly to the
trapping user code, rather than going through the kernel again.  Only a
handful of instructions are actually executed in kernel mode.
Implement sysarch(SPARC_UTRAP_INSTALL).
Add code to handle sharing of the user trap table across forks and unsharing
at exec.

This can be used to implement efficient tracking of floating point register
usage in userland, fe by a thread library, and to handle alignment fault
fixups and instruction emulation in userland, for which the code may need
to be different for 32bit and 64bit binaries.

22 years agoAdd a panic stack, which is used as a known good stack when there is
jake [Tue, 1 Jan 2002 20:26:46 +0000 (20:26 +0000)]
Add a panic stack, which is used as a known good stack when there is
something wrong with the kernel stack.
Add code to check the kernel stack pointer in various important places
and try hard not to go down in flames if its wrong.

22 years agoAdd a soft trap for restoring the fpu registers from the pcb.
jake [Tue, 1 Jan 2002 20:15:39 +0000 (20:15 +0000)]
Add a soft trap for restoring the fpu registers from the pcb.

22 years agoFix long lines in the trap table due to the abi specificied trap types
jake [Tue, 1 Jan 2002 20:10:51 +0000 (20:10 +0000)]
Fix long lines in the trap table due to the abi specificied trap types
having overly long names.

22 years agoUpdate of the VIA 82c686b southbridge data corruption fix, also
sos [Tue, 1 Jan 2002 19:58:11 +0000 (19:58 +0000)]
Update of the VIA 82c686b southbridge data corruption fix, also
include more possible chipset candidates.

22 years ago o Remove an errant ';' introduced in the last revision.
alc [Tue, 1 Jan 2002 19:44:01 +0000 (19:44 +0000)]
 o Remove an errant ';' introduced in the last revision.
 o Remove an unused variable.

22 years agoMake `WHERE TO START' and `WHERE TO STOP' subsections within `DESCRIPTION'.
chris [Tue, 1 Jan 2002 19:20:02 +0000 (19:20 +0000)]
Make `WHERE TO START' and `WHERE TO STOP' subsections within `DESCRIPTION'.

Suggested by: sheldonh
Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years ago- Do not uiomove with a mutex locked.
tanimura [Tue, 1 Jan 2002 17:36:26 +0000 (17:36 +0000)]
- Do not uiomove with a mutex locked.

- Move from msleep/wakeup to condvar.

- Return either zero or a positive errno value from a function.
  Return additional result via references.

- Unify the typedef of callback functions.

22 years agoRemove long-obsolete README.
des [Tue, 1 Jan 2002 16:27:26 +0000 (16:27 +0000)]
Remove long-obsolete README.

22 years agoIn verbose mode, display the full error message from the server, stripping
des [Tue, 1 Jan 2002 16:25:29 +0000 (16:25 +0000)]
In verbose mode, display the full error message from the server, stripping
it of HTML tags and comments.

PR: bin/32989
MFC after: 1 week

22 years agoo Remove premature use of nmp->nm_cred, it hasn't been initialized yet.
rwatson [Tue, 1 Jan 2002 16:17:55 +0000 (16:17 +0000)]
o Remove premature use of nmp->nm_cred, it hasn't been initialized yet.

22 years agoRemove VT100 escapes from debugging messages now that they're enabled by
des [Tue, 1 Jan 2002 14:48:09 +0000 (14:48 +0000)]
Remove VT100 escapes from debugging messages now that they're enabled by
default.

PR: 32988
MFC after: 3 days

22 years agoThis file does not need to include procfs.h. This fixes a bunch of warnings.
des [Tue, 1 Jan 2002 14:46:57 +0000 (14:46 +0000)]
This file does not need to include procfs.h.  This fixes a bunch of warnings.

22 years agoClarify comment about pam_unix fallback for ftpd
ache [Tue, 1 Jan 2002 13:38:01 +0000 (13:38 +0000)]
Clarify comment about pam_unix fallback for ftpd

22 years agoAdd support for Nikon Coolpix E775 and E885 cameras.
jedgar [Tue, 1 Jan 2002 13:28:43 +0000 (13:28 +0000)]
Add support for Nikon Coolpix E775 and E885 cameras.

PR: 33407 (E885)
Submitted by: Brian Behlendorf <brian@hyperreal.org> (E885)

22 years agoTurn on pam_opie.so for ftpd by default
ache [Tue, 1 Jan 2002 13:27:11 +0000 (13:27 +0000)]
Turn on pam_opie.so for ftpd by default
It not affect non-OPIE users

22 years agoFix OPIE auth
ache [Tue, 1 Jan 2002 13:14:25 +0000 (13:14 +0000)]
Fix OPIE auth

22 years agoWelcome 2002. Happy new year.
imp [Tue, 1 Jan 2002 09:20:52 +0000 (09:20 +0000)]
Welcome 2002.  Happy new year.

22 years agoFix to conform standard
ache [Tue, 1 Jan 2002 07:43:45 +0000 (07:43 +0000)]
Fix to conform standard

22 years agoAdd 8859-5 keymaps
ache [Tue, 1 Jan 2002 06:51:22 +0000 (06:51 +0000)]
Add 8859-5 keymaps

PR: 32449
Submitted by: Alexey Klimov <klim@unique.kiev.ua>

22 years agoAdd 8859-5 fonts
ache [Tue, 1 Jan 2002 06:44:36 +0000 (06:44 +0000)]
Add 8859-5 fonts

PR: 32448
Submitted by: Alexey Klimov <klim@unique.kiev.ua>

22 years agoBruce thought some of the comments I added were not clear enough. This is
mckay [Tue, 1 Jan 2002 06:14:26 +0000 (06:14 +0000)]
Bruce thought some of the comments I added were not clear enough.  This is
a combination of my words and his.  We will stop fiddling now. :-)

22 years agoAdd the nullmodem device
julian [Tue, 1 Jan 2002 05:16:03 +0000 (05:16 +0000)]
Add the nullmodem device

22 years agoCorrect typo.
grog [Tue, 1 Jan 2002 00:50:30 +0000 (00:50 +0000)]
Correct typo.

22 years ago o Some style(9)-motivated changes to white space.
alc [Tue, 1 Jan 2002 00:40:29 +0000 (00:40 +0000)]
 o Some style(9)-motivated changes to white space.

22 years agoCorrect typos.
grog [Mon, 31 Dec 2001 23:59:53 +0000 (23:59 +0000)]
Correct typos.

22 years agoSave stack space by converting areq in sc->areq.
ambrisko [Mon, 31 Dec 2001 22:12:42 +0000 (22:12 +0000)]
Save stack space by converting areq in sc->areq.

Reviewed by: Brooks Davis <brooks@freebsd.org>
Warner Losh <imp@freebsd.org>
Approved by: Brooks Davis <brooks@freebsd.org>
Warner Losh <imp@freebsd.org>

22 years agoFix bugs in the structure for rx_frame by making gap length one byte and
ambrisko [Mon, 31 Dec 2001 22:01:44 +0000 (22:01 +0000)]
Fix bugs in the structure for rx_frame by making gap length one byte and
a packed array so sizeof work.  This broke RFMON mode and passing
up 802.11 packets.

The Linux emulation code was derived from the open source Linux driver to
maintain compatibility.

LEAP support is added, hints from Richard Johnson.  I've verified this
locally with PC350v42510.img firmware.  More bug fixing from Marco to
fix long passwords.

Change DELAYs in flash part of driver to FLASH_DELAY which uses tsleep
so it doesn't look like your system died during a flash update.

Install header files in /usr/include/dev/an

Cleanup some ifmedia bugs add "Home" key mode to ifmedia and ancontrol.
This way you can manage 2 keys a little easier.  Map the home mode into
key 5.  Enhance ifconfig to dump the various configured SSIDs.  I use
a bunch of different ones and roam between them.  Use the syntax similar
to the WEP keys to deal with setting difference SSIDs.

Bump up up the Card capabilities RID since they added 2 bytes to it
in the latest firmware.  Thankfully we changed it from a terminal
failure so the card still worked but the driver whined.

Some cleanup patches from Marco Molteni.

Submitted by: Richard Johnson <raj@cisco.com>
Marco Molteni <molter@tin.it>
and myself
Various checks: David Wolfskill <david@catwhisker.org>
Reviewed by: Brooks Davis <brooks@freebsd.org>
Warner Losh <imp@freebsd.org>
Approved by: Brooks Davis <brooks@freebsd.org>
Warner Losh <imp@freebsd.org>
Obtained from: Linux emulation API's from Aironet driver.

22 years agoo Pass td into nfs_mountroot() to eliminate an XXX'd curthread use.
rwatson [Mon, 31 Dec 2001 21:00:00 +0000 (21:00 +0000)]
o Pass td into nfs_mountroot() to eliminate an XXX'd curthread use.
  Since it's in the parent function anyway, might as well pass it
  another layer down.

Obtained from: TrustedBSD Project

22 years agoo Remove premature leakage of use of td_ucred from base source tree:
rwatson [Mon, 31 Dec 2001 20:56:59 +0000 (20:56 +0000)]
o Remove premature leakage of use of td_ucred from base source tree:
  instead, use td->td_proc->p_ucred.

22 years agoGrrr. The tlb code is strewn over 3 files and I misread it. Revert
dillon [Mon, 31 Dec 2001 20:32:53 +0000 (20:32 +0000)]
Grrr.  The tlb code is strewn over 3 files and I misread it.  Revert
the last change (it was a NOP), and remove the XXX comments that no longer
apply.

22 years agoo Add missing #include's of sys/proc.h, missed in merge, required to
rwatson [Mon, 31 Dec 2001 20:05:26 +0000 (20:05 +0000)]
o Add missing #include's of sys/proc.h, missed in merge, required to
  dereference td->td_proc->p_ucred.

22 years agoYou know those 'XXX what about SMP' comments in pmap_kenter()? Well,
dillon [Mon, 31 Dec 2001 20:02:46 +0000 (20:02 +0000)]
You know those 'XXX what about SMP' comments in pmap_kenter()?  Well,
they were right.  Fix both kenter() and kremove() for SMP by ensuring that
the tlb is flushed on other cpu's.  This will directly solve random-corruption
panic issues in -stable when it is MFC'd.  Better to be safe then sorry, we
can optimize this later.

Original Suspicion by: peter
Maybe MFC: immediately on re's permission

22 years agoSpelling fixes.
bp [Mon, 31 Dec 2001 19:29:43 +0000 (19:29 +0000)]
Spelling fixes.

PR: kern/33131
Submitted by: Anders Andersson <anders@hack.org>
MFC after: 1 week

22 years agoo Add a 'showattr' function the extattrctl, allowing a backing file to
rwatson [Mon, 31 Dec 2001 18:21:24 +0000 (18:21 +0000)]
o Add a 'showattr' function the extattrctl, allowing a backing file to
  be inspected to show the maximum attribute size and file.

22 years agoo Make the credential used by socreate() an explicit argument to
rwatson [Mon, 31 Dec 2001 17:45:16 +0000 (17:45 +0000)]
o Make the credential used by socreate() an explicit argument to
  socreate(), rather than getting it implicitly from the thread
  argument.

o Make NFS cache the credential provided at mount-time, and use
  the cached credential (nfsmount->nm_cred) when making calls to
  socreate() on initially connecting, or reconnecting the socket.

This fixes bugs involving NFS over TCP and ipfw uid/gid rules, as well
as bugs involving NFS and mandatory access control implementations.

Reviewed by: freebsd-arch

22 years agoRemove unused includes, remove the duplicated definitions of AUTHNAMELEN
joerg [Mon, 31 Dec 2001 09:32:32 +0000 (09:32 +0000)]
Remove unused includes, remove the duplicated definitions of AUTHNAMELEN
and AUTHKEYLEN and include <net/if_sppp.h> instead.

22 years agoRemove ispppcontrol.
joerg [Mon, 31 Dec 2001 09:31:51 +0000 (09:31 +0000)]
Remove ispppcontrol.

22 years agoRemove unused includes.
joerg [Mon, 31 Dec 2001 09:31:30 +0000 (09:31 +0000)]
Remove unused includes.

22 years agoDisabled the olpt driver temporarily to avoid 'repeat make_dev' panic.
nyan [Mon, 31 Dec 2001 09:02:24 +0000 (09:02 +0000)]
Disabled the olpt driver temporarily to avoid 'repeat make_dev' panic.

22 years agoWhen the old nfsm_adv() macro was moved to nfsm_adv_xx(), a '>='
iedowse [Mon, 31 Dec 2001 06:56:31 +0000 (06:56 +0000)]
When the old nfsm_adv() macro was moved to nfsm_adv_xx(), a '>='
must have been inadvertently changed to '>'. This broke nfsm_adv()
in the case where the advancement count is equal to the amount of
data remaining in the current mbuf. Instead of moving the current
position N bytes forward, nfs_adv() could end up moving it back to
N bytes from the start of the mbuf data.

This should fix the client-side readdirplus problems that have been
reported since September.

22 years ago o Correct an off-by-one error in aio_suspend(2).
alc [Mon, 31 Dec 2001 03:13:24 +0000 (03:13 +0000)]
 o Correct an off-by-one error in aio_suspend(2).

PR: 18350

22 years ago o Use "td->td_proc" instead of "curproc" where possible.
alc [Mon, 31 Dec 2001 02:03:39 +0000 (02:03 +0000)]
 o Use "td->td_proc" instead of "curproc" where possible.
 o Eliminate the unnecessary initialization of several static variables
   to zero.

22 years agoSynchronise with NetBSD.
joe [Mon, 31 Dec 2001 00:52:17 +0000 (00:52 +0000)]
Synchronise with NetBSD.

In order of importance:
* Make ugen use updated frlengths.
* More tests for NULL pipes.
* Generate better error codes on bulk write.
* Error messages in general.

22 years agoRevert previous definition of cpu_throw(). Non-MP configurations
marcel [Sun, 30 Dec 2001 23:39:15 +0000 (23:39 +0000)]
Revert previous definition of cpu_throw(). Non-MP configurations
were broken as well.

22 years agoPrint open(2) flags in hex instread of decimal for clarity.
alfred [Sun, 30 Dec 2001 23:11:52 +0000 (23:11 +0000)]
Print open(2) flags in hex instread of decimal for clarity.

22 years agoPartial sync with NetBSD:
joe [Sun, 30 Dec 2001 23:02:31 +0000 (23:02 +0000)]
Partial sync with NetBSD:
Replace LE() macro with htole32()/le32toh().

22 years agoImplement the option to disable IPv6 on an sppp interface (counterpart
joerg [Sun, 30 Dec 2001 20:43:09 +0000 (20:43 +0000)]
Implement the option to disable IPv6 on an sppp interface (counterpart
to the kernel code just committed).

MFC after: 1 month

22 years agoImplement an option to administratively disable the negotiation of
joerg [Sun, 30 Dec 2001 20:42:29 +0000 (20:42 +0000)]
Implement an option to administratively disable the negotiation of
IPv6 on an sppp interface.  In an IPv6-enabled kernel, every IPv6
interface automatically gets an IPv6 address assigned (and IPv6
multicast packets sent at initialization time).  For sppp links where
we know our remote peer wouldn't support IPv6 at all, there's no point
in attempting to negotiate IPV6CP (or to even dial out for an IPv6
packet at all for dial-on-demand interfaces).

I wish there were a more generic way to administratively disable IPv6
on an interface instead.  ume told me there isn't.

While i was at it, converted both, enable_vj and enable_ipv6 into flag
bits in struct sppp (enable_vj used to be an int of its own).

MFC after: 1 month

22 years agoo Logically split up sentences, removing hard sentence breaks.
chris [Sun, 30 Dec 2001 20:34:53 +0000 (20:34 +0000)]
o Logically split up sentences, removing hard sentence breaks.
o Use .Sx where a section cross-reference is intended.

Obtained from: TrustedBSD Project
Sponsored by: DARPA, NAI Labs

22 years agoGC an alternate trap_pfault() which has rotted away behind an "#ifdef notyet"
phk [Sun, 30 Dec 2001 19:43:59 +0000 (19:43 +0000)]
GC an alternate trap_pfault() which has rotted away behind an "#ifdef notyet"
since 21-Mar-95 .

22 years agoFix two typos in previous commit.
joerg [Sun, 30 Dec 2001 19:37:41 +0000 (19:37 +0000)]
Fix two typos in previous commit.

22 years agoUse LC_ALL to pick LC_COLLATE too (for strcoll())
ache [Sun, 30 Dec 2001 19:30:10 +0000 (19:30 +0000)]
Use LC_ALL to pick LC_COLLATE too (for strcoll())

22 years agoMerge last-minute fix from the i4b file made by gj:
joerg [Sun, 30 Dec 2001 19:02:27 +0000 (19:02 +0000)]
Merge last-minute fix from the i4b file made by gj:

Protect mtx_init() invocations with mtx_intialized() checks to avoid a
reported panic.

MFC after: 1 month

22 years agoEliminate semexit_hook using at_exit(9) and rm_at_exit(9).
alc [Sun, 30 Dec 2001 18:55:09 +0000 (18:55 +0000)]
Eliminate semexit_hook using at_exit(9) and rm_at_exit(9).

Reviewed by: alfred

22 years agoOops, the arguments to a bcopy() were reversed, which broke zeroing
iedowse [Sun, 30 Dec 2001 18:51:51 +0000 (18:51 +0000)]
Oops, the arguments to a bcopy() were reversed, which broke zeroing
of unused partition entries and later detection of unused entries.

Use memcpy to be consistent with the rest of the code, and fix a
minor style nit.

Submitted by: bde