]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoThis commit was generated by cvs2svn to compensate for changes in r77298,
obrien [Mon, 28 May 2001 05:21:37 +0000 (05:21 +0000)]
This commit was generated by cvs2svn to compensate for changes in r77298,
which included commits to RCS files with non-trunk default branches.

23 years agoImport of GNU Binutils version 2.11.0.
obrien [Mon, 28 May 2001 05:21:37 +0000 (05:21 +0000)]
Import of GNU Binutils version 2.11.0.
Believe it or not, this is heavily stripped down.

23 years agogrep -v offending lines from loader.4th until the master version of it
obrien [Mon, 28 May 2001 05:18:13 +0000 (05:18 +0000)]
grep -v offending lines from loader.4th until the master version of it
is fixed.

23 years agoAdd the generated help files to CLEANDIRS.
obrien [Mon, 28 May 2001 05:12:13 +0000 (05:12 +0000)]
Add the generated help files to CLEANDIRS.

Found by: rm -rf /usr/obj/usr/src/sys/boot ; make ; make clean ; cvs -q up

23 years agoFix overflowing on nul character.
mikeh [Mon, 28 May 2001 03:49:23 +0000 (03:49 +0000)]
Fix overflowing on nul character.

PR: bin/16926
MFC after: 1 week

23 years agoTurns out that one bit isn't enough. Introduce two new fields
imp [Mon, 28 May 2001 02:53:02 +0000 (02:53 +0000)]
Turns out that one bit isn't enough.  Introduce two new fields
csc_route and func_route to hold the way that each interrupt is
routed.  csc is Card Status Change in the datasheets and standard, but
is called "Management Interrupt" in FreeBSDese.  There are three types
of interrupt routing:  ISA parallel, PCI parallel and ISA serial (some
chipsets support other types as well, but I don't plan on supporting
them).

When we try to allocate an interrupt, and the type for that interrupt
is pci_parallel, allow it to be shared by oring in RF_SHAREABLE to the
flags argument.  Introduce pcic_alloc_resource to allow this to
happen.

23 years agoImplement snapshots. The new -d option allows the user to find out
dd [Mon, 28 May 2001 01:22:37 +0000 (01:22 +0000)]
Implement snapshots.  The new -d option allows the user to find out
who was logged in at a certain time and date.

Obtained from: OpenBSD

23 years agoHandle NULL struct device *s
brian [Mon, 28 May 2001 01:00:03 +0000 (01:00 +0000)]
Handle NULL struct device *s

23 years agoSilence warnings and compile with WARNS=2 on i386 and alpha.
dd [Mon, 28 May 2001 00:55:50 +0000 (00:55 +0000)]
Silence warnings and compile with WARNS=2 on i386 and alpha.

23 years agoIn sys/boot/common/module.c, near line 105 a request for a raw
jesper [Sun, 27 May 2001 23:26:11 +0000 (23:26 +0000)]
In sys/boot/common/module.c, near line 105 a request for a raw
file is processed by passing its name in argv[1]:
        return(mod_loadobj(typestr, argv[1]));
however, it is not tested to see if argv[1] actually is defined.

At best, mod_loadobj() near line 244 returns an error like
"can't find 'garbage'" but if the "filename" entered is sufficiently
long, some buffer gets overrun. Of course, "load -t filename" is
actually a typo because we meant to type "load -t mfs_root filename";
nevertheless, a hung machine seems like too harsh a punishment for
such a small typo...

PR: i386/27693
Submitted by: Adrian Steinmann <ast@marabu.ch>
MFC after: 1 week

23 years agoAdd two new manual pages related to general firewall and tuning issues
dillon [Sun, 27 May 2001 23:14:27 +0000 (23:14 +0000)]
Add two new manual pages related to general firewall and tuning issues

Reviewed by: hackers

23 years agoDon't confuse the admin with spurious error messages
yar [Sun, 27 May 2001 22:50:22 +0000 (22:50 +0000)]
Don't confuse the admin with spurious error messages
about non-existent mount directories (which would come
into existence after the real mount has occured) when just
testing for if there are any NFS filesystems in /etc/fstab.

PR: bin/26597
Submitted by: Dmitry Morozovsky <marck@rinet.ru>
MFC after: 3 days

23 years agofinally fix intr routing on alphas such as the as500 after months of
gallatin [Sun, 27 May 2001 22:22:03 +0000 (22:22 +0000)]
finally fix intr routing on alphas such as the as500 after months of
breakage:

- call PCIB_ROUTE_INTERRUPT() regardless of how valid the intline looks.
  Some alphas leave garbage in the intline and leave the intr mapping
  to OS platform support routines that map slots/buses to intlines
- Down in the alpha pci code, first try platform.pci_intr_route() and
  if it doesn't exist or returns garbage, just read the intline out of
  config space.

tested on AS500 (garbage in intline) and UP1000 (PC-like, intline is valid)

Note that a nice little hack like the APIC_IO section of pci_cfgregread()
is not workable.  This is because the calling interface for
alpha_pci_route_interrupt() requires us to figure out the bus/slot/etc
from a device_t.  At pci_read_device() time, we don't have a device_t
for the bus/slot/func in question.

23 years agoFix typo: "an socket" -> "a socket".
olgeni [Sun, 27 May 2001 22:17:46 +0000 (22:17 +0000)]
Fix typo: "an socket" -> "a socket".
Place period at the end of AUTHORS section.

MFC after: 1 week

23 years agoo uifree() the cr_ruidinfo in crfree() as well as cr_uidinfo now that the real uid
rwatson [Sun, 27 May 2001 21:43:46 +0000 (21:43 +0000)]
o uifree() the cr_ruidinfo in crfree() as well as cr_uidinfo now that the real uid
  info is in the credential also.

Submitted by: egge

23 years agoMiscellaneous cleanups; sync to OpenBSD as much as feasible.
dd [Sun, 27 May 2001 21:08:19 +0000 (21:08 +0000)]
Miscellaneous cleanups; sync to OpenBSD as much as feasible.
Noteworthy changes include:
 * Use getopt(3).
 * Fix overflows in -b and -w options.
 * Use strlcpy(3) and snprintf(3) in favor of strcpy(3) and
   sprintf(3), respectively.  Also check return values of the former
   two.
 * Fix lots of other gratuitous differences with OpenBSD.

Obtained from: OpenBSD

23 years agoClose the opening comment line on these two files that mysteriouly vanished.
scottl [Sun, 27 May 2001 21:07:39 +0000 (21:07 +0000)]
Close the opening comment line on these two files that mysteriouly vanished.

23 years agoCleanup mail(1)'s varying styles by converting to using style(9).
mikeh [Sun, 27 May 2001 20:26:22 +0000 (20:26 +0000)]
Cleanup mail(1)'s varying styles by converting to using style(9).

Also take a stab at cleaning up BDECFLAGS and convert all uses of
NOSTR, NIL, NONE, NOVAR, NOGRP, NOGE to NULL. Also kill 'register' to
get diffs somewhat closer to OpenBSD/NetBSD.

There are no functional changes however.
Reviewed by: nra (visual inspection)

23 years agoFix reboot buglet when BOOT_BTX_NOHANG is defined.
rnordier [Sun, 27 May 2001 20:15:10 +0000 (20:15 +0000)]
Fix reboot buglet when BOOT_BTX_NOHANG is defined.

Submitted by: Umesh Krishnaswamy <umesh@juniper.net>

23 years agoDrop nested __P(). This trips up Supelec's dcc.
rnordier [Sun, 27 May 2001 19:57:36 +0000 (19:57 +0000)]
Drop nested __P().  This trips up Supelec's dcc.

23 years agobeginnings of virtual playback channel support
cg [Sun, 27 May 2001 17:22:00 +0000 (17:22 +0000)]
beginnings of virtual playback channel support

instead of using two malloced arrays for storing channel lists, use an
slist.  convert the sndstat device to use sbufs and optionally provide more
detail about channel state.

vchans are software mixed playback channels.  they are not enabled by this
commit.  they use the feeder infrastructure to emulate normal playback
channels in a manner transparent to applications, whilst providing as many
channels are desired, especially suitable for devices with only one hardware
playback channel.  in the future they will provide additional features.

those wishing to test this functionality will need to add vchan.c to
sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it.

blocksize and auto-rate selection are not yet supported.

23 years agoAs in with ":", check dictionary space and increase it if necessary before
dcs [Sun, 27 May 2001 16:30:10 +0000 (16:30 +0000)]
As in with ":", check dictionary space and increase it if necessary before
"create".

23 years agoMFS: Add vim-lite and netscape47-navigator.
olgeni [Sun, 27 May 2001 15:45:05 +0000 (15:45 +0000)]
MFS: Add vim-lite and netscape47-navigator.

23 years agoadd a new method for retrieving feeder parameters
cg [Sun, 27 May 2001 14:49:14 +0000 (14:49 +0000)]
add a new method for retrieving feeder parameters

23 years agodon't erase info in sndbuf_setup()
cg [Sun, 27 May 2001 14:39:34 +0000 (14:39 +0000)]
don't erase info in sndbuf_setup()
set free'd pointers to NULL in sndbuf_free()
add a new function

23 years agoAdd an example of a kernel configuration file fragment for ISDN.
hm [Sun, 27 May 2001 11:34:15 +0000 (11:34 +0000)]
Add an example of a kernel configuration file fragment for ISDN.

23 years agoThe PERL_THREADED knob is causing too many people too many problems.
sheldonh [Sun, 27 May 2001 11:23:31 +0000 (11:23 +0000)]
The PERL_THREADED knob is causing too many people too many problems.
Add a dire warning about the experimental nature of threaded Perl.

23 years agoImplement a HTTP_USER_AGENT environment variable.
des [Sun, 27 May 2001 11:00:36 +0000 (11:00 +0000)]
Implement a HTTP_USER_AGENT environment variable.

PR: 27669
Submitted by: Eoin Lawless <eoin@maths.tcd.ie>

23 years agoDamn. Fix typo in previous commit.
des [Sun, 27 May 2001 10:46:19 +0000 (10:46 +0000)]
Damn.  Fix typo in previous commit.

23 years agoupdate the sample isdnd.rc file with a firmware keyword example and a real
hm [Sun, 27 May 2001 08:05:57 +0000 (08:05 +0000)]
update the sample isdnd.rc file with a firmware keyword example and a real
world example of the use of the valid keyword in conjunction with budget
handling.

23 years agoInstead of defining CONS_CLRHIST if it isn't defined, simply disable
dd [Sun, 27 May 2001 06:28:08 +0000 (06:28 +0000)]
Instead of defining CONS_CLRHIST if it isn't defined, simply disable
the -c option [when CONS_CLRHIST isn't defined].  This is okay since
the only time CONS_CLRHIST wouldn't be defined is when kbdcontrol is
being built in bootstrap-tools, and -c isn't needed then.

Submitted by: imp

23 years agoAllow a shareable interrupts. Note, the bridge must set this flag or
imp [Sun, 27 May 2001 05:53:37 +0000 (05:53 +0000)]
Allow a shareable interrupts.  Note, the bridge must set this flag or
the irq will be unshareable, as things are now.

More work likely is needed, but this is a good checkpoint.

# pcic_pci.c is getting closer :-)

23 years agoA la rev. 1.36, define CONS_CLRHIST here if it isn't already since
dd [Sun, 27 May 2001 04:03:52 +0000 (04:03 +0000)]
A la rev. 1.36, define CONS_CLRHIST here if it isn't already since
this is a build tool, so it has to build on 4.x with the old headers.

23 years agoPrint out the asc/ascq and description even when both the asc and ascq
ken [Sun, 27 May 2001 03:22:51 +0000 (03:22 +0000)]
Print out the asc/ascq and description even when both the asc and ascq
are zero.  This is so that users will see the "no addtional sense" printout
and know that they have the full sense information.

23 years agoIn the VIRTUAL_TTY macro, add a check to make sure that the result of
dd [Sun, 27 May 2001 00:57:25 +0000 (00:57 +0000)]
In the VIRTUAL_TTY macro, add a check to make sure that the result of
SC_DEV isn't NULL; if it is, evaluate to NULL and don't dereference
NULL.  Callers of VIRTUAL_TTY must already check for the result being
NULL since si_tty can be NULL, so this should be safe.

This fixes a panic when trying to switch to a different vty in an
environment such as userconfig (-c option to the kernel).

PR: 26508

23 years agoAdd a -c option which clears the history buffer using the new
dd [Sun, 27 May 2001 00:47:38 +0000 (00:47 +0000)]
Add a -c option which clears the history buffer using the new
CONS_CLRHIST ioctl.

PR: 27616
Reviewed by: ru

23 years agoAdd a CONS_CLRHIST ioctl which instructs syscons to clear the history
dd [Sun, 27 May 2001 00:45:59 +0000 (00:45 +0000)]
Add a CONS_CLRHIST ioctl which instructs syscons to clear the history
(scroll-back) buffer.

PR: 27616
Reviewed by: ru

23 years agoAcquire vm_mtx before calling vm_pager_deallocate.
dd [Sun, 27 May 2001 00:42:46 +0000 (00:42 +0000)]
Acquire vm_mtx before calling vm_pager_deallocate.

Reviewed by: phk

23 years agoFix a trivial warning and clamp down with WARNS=2
kris [Sat, 26 May 2001 22:47:58 +0000 (22:47 +0000)]
Fix a trivial warning and clamp down with WARNS=2

MFC after:      1 week

23 years agoFix a trivial warning and clamp down with WARNS=2
kris [Sat, 26 May 2001 22:45:14 +0000 (22:45 +0000)]
Fix a trivial warning and clamp down with WARNS=2

MFC after: 1 week

23 years agoFix some trivial warnings and clamp down with WARNS=2
kris [Sat, 26 May 2001 22:39:33 +0000 (22:39 +0000)]
Fix some trivial warnings and clamp down with WARNS=2

MFC after: 1 week

23 years agoFix WARNS=2 warnings on alpha and i386 and clamp it down
kris [Sat, 26 May 2001 21:49:29 +0000 (21:49 +0000)]
Fix WARNS=2 warnings on alpha and i386 and clamp it down

23 years agoFix warnings to compile with WARNS=2 on i386 and alpha
kris [Sat, 26 May 2001 20:45:25 +0000 (20:45 +0000)]
Fix warnings to compile with WARNS=2 on i386 and alpha

Reviewed by: bde

23 years agoDon't copy the trailing zero in readlink, it confuses namei().
phk [Sat, 26 May 2001 20:07:57 +0000 (20:07 +0000)]
Don't copy the trailing zero in readlink, it confuses namei().

PR: 27656

23 years agoo pcred-removal changes included modifications to optimize the setting of
rwatson [Sat, 26 May 2001 19:59:44 +0000 (19:59 +0000)]
o pcred-removal changes included modifications to optimize the setting of
  the saved uid and gid during execve().  Unfortunately, the optimizations
  were incorrect in the case where the credential was updated, skipping
  the setting of the saved uid and gid when new credentials were generated.
  This change corrects that problem by handling the newcred!=NULL case
  correctly.

Reported/tested by: David Malone <dwmalone@maths.tcd.ie>

Obtained from: TrustedBSD Project

23 years agoWhen running on a tty, install an authentication callback.
des [Sat, 26 May 2001 19:51:05 +0000 (19:51 +0000)]
When running on a tty, install an authentication callback.

23 years agoDocument the authentication callback interface.
des [Sat, 26 May 2001 19:38:34 +0000 (19:38 +0000)]
Document the authentication callback interface.
Update RFC references (should have done that ages ago...)

23 years agoBump major number.
des [Sat, 26 May 2001 19:37:26 +0000 (19:37 +0000)]
Bump major number.

23 years agoAdd rudimentary support for an authentication callback function.
des [Sat, 26 May 2001 19:37:15 +0000 (19:37 +0000)]
Add rudimentary support for an authentication callback function.

23 years agoMAXPATHLEN -> PATH_MAX
des [Sat, 26 May 2001 19:36:49 +0000 (19:36 +0000)]
MAXPATHLEN -> PATH_MAX

23 years agoFrom the submitter:
markm [Sat, 26 May 2001 17:40:09 +0000 (17:40 +0000)]
From the submitter:

$ PERL_DL_NONLAZY=1 perl -MPOSIX -e ''
Can't load '/usr/libdata/perl/5.6.0/mach/auto/POSIX/POSIX.so' for module
POSIX: /usr/libdata/perl/5.6.0/mach/auto/POSIX/POSIX.so: Undefined
symbol "sv_setptrobj" at /usr/libdata/perl/5.6.0/XSLoader.pm line 73.
 at /usr/libdata/perl/5.6.0/mach/POSIX.pm line 24
Compilation failed in require.
BEGIN failed--compilation aborted.

This problem only exists in -CURRENT.  Most often it is reproduced when
compiling some perl extension manually.  Make test uses PERL_DL_NONLAZY,
and if a tested module uses POSIX, boom!

Luckily, we don't see it very often, mostly because the vast majority of
p5 ports do not perform the `make test' step.

Submitted by: Anton Berezin <tobez@tobez.org>
Needs to be a committer already: Anton Berezin <tobez@tobez.org>

23 years agoPlug memory leak.
des [Sat, 26 May 2001 17:23:38 +0000 (17:23 +0000)]
Plug memory leak.

PR: 27506

23 years agoo Include file to provide prototypes for regression testing-specific
rwatson [Sat, 26 May 2001 15:37:25 +0000 (15:37 +0000)]
o Include file to provide prototypes for regression testing-specific
  system calls/et al.  Not much in there just yet, but will most likely
  grow.

Obtained from: TrustedBSD Project

23 years agoMake dg_echo return up to the first 65536 bytes of a datagram.
dwmalone [Sat, 26 May 2001 14:40:39 +0000 (14:40 +0000)]
Make dg_echo return up to the first 65536 bytes of a datagram.
The patch I used isn't quite the one Lars suggested, but the size
of the largest datagram you can recv isn't #defined anywhere, and
probably isn't even bounded for some protocols.

PR: 25050
Submitted by: Lars Eggert <larse@isi.edu>

23 years agoDon't spell requester as requestor.
dwmalone [Sat, 26 May 2001 14:33:47 +0000 (14:33 +0000)]
Don't spell requester as requestor.

23 years agoCURRENT no longer needs -pthread
markm [Sat, 26 May 2001 14:06:45 +0000 (14:06 +0000)]
CURRENT no longer needs -pthread

23 years agoDocument /usr/include/fs/ntfs and /usr/include/fs/nwfs.
ru [Sat, 26 May 2001 12:04:02 +0000 (12:04 +0000)]
Document /usr/include/fs/ntfs and /usr/include/fs/nwfs.

23 years ago- sys/n[tw]fs moved to sys/fs/n[tw]fs
ru [Sat, 26 May 2001 11:57:45 +0000 (11:57 +0000)]
- sys/n[tw]fs moved to sys/fs/n[tw]fs
- /usr/include/n[tw]fs moved to /usr/include/fs/n[tw]fs

23 years agoadd Dualmode Zoom/FaxModem 56K (internal) Model 2919
sanpei [Sat, 26 May 2001 11:17:57 +0000 (11:17 +0000)]
add Dualmode Zoom/FaxModem 56K (internal) Model 2919

PR: kern/27476
Submitted by: Eric Beyer <lnxfrk@earthlink.net>

23 years agoImprove and extend. Use new modules to set policy, and provide another
markm [Sat, 26 May 2001 09:56:17 +0000 (09:56 +0000)]
Improve and extend. Use new modules to set policy, and provide another
example for WHEELSU-type su(1).

23 years agoDeconvolute the authentication mess, and hand total responsiblity
markm [Sat, 26 May 2001 09:52:36 +0000 (09:52 +0000)]
Deconvolute the authentication mess, and hand total responsiblity
for authentication to PAM. This meens that WHEELSU-type logic can
now be effected in the pam.conf "su" configuration stack. While here,
clean up the mess that the code had assumed over years of hacking by
folks using different styles. ANSIfy.

There is more policy in here that can be handed over to PAM. This will
be revisited.

23 years agoDang, I lost the sign in that brucification. DTRT with PCATCH.
phk [Sat, 26 May 2001 09:37:42 +0000 (09:37 +0000)]
Dang, I lost the sign in that brucification.  DTRT with PCATCH.

Cut Out In Cardboard by: bde

23 years agoForgot to cvs add these two files for the previous commit.
phk [Sat, 26 May 2001 09:28:43 +0000 (09:28 +0000)]
Forgot to cvs add these two files for the previous commit.

23 years agoCurrently, each wireless networking driver has it's own control program
phk [Sat, 26 May 2001 09:27:08 +0000 (09:27 +0000)]
Currently, each wireless networking driver has it's own control program
despite the fact that most people want to set exactly the same settings
regardless of which card they have.  It has been repeatidly suggested
that this configuration should be done via ifconfig.  This patch
implements the required functionality in ifconfig and add support to the
wi and an drivers.  It also provides partial, untested support for the
awi driver.

PR: 25577
Submitted by: Brooks Davis <brooks@one-eyed-alien.net>

23 years agoCreate a general facility for making dev_t's depend on another
phk [Sat, 26 May 2001 08:27:58 +0000 (08:27 +0000)]
Create a general facility for making dev_t's depend on another
dev_t.  The dev_depends(dev_t, dev_t) function is for tying them
to each other.

When destroy_dev() is called on a dev_t, all dev_t's depending
on it will also be destroyed (depth first order).

Rewrite the make_dev_alias() to use this dependency facility.

kern/subr_disk.c:
Make the disk mini-layer use dependencies to make sure all
relevant dev_t's are removed when the disk disappears.

Make the disk mini-layer precreate some magic sub devices
which the disk/slice/label code expects to be there.

kern/subr_disklabel.c:
Remove some now unneeded variables.

kern/subr_diskmbr.c:
Remove some ancient, commented out code.

kern/subr_diskslice.c:
Minor cleanup.  Use name from dev_t instead of dsname()

23 years agoRemove pcm hints here now that it's gone from GENERIC.
jkh [Sat, 26 May 2001 08:04:34 +0000 (08:04 +0000)]
Remove pcm hints here now that it's gone from GENERIC.

Reminded-by: bde

23 years agoUpdate reality in the strings comment
dougb [Sat, 26 May 2001 06:01:43 +0000 (06:01 +0000)]
Update reality in the strings comment

23 years agoFix the error buffer passed to kvm_openfiles to have a correct length
tmm [Sat, 26 May 2001 00:02:13 +0000 (00:02 +0000)]
Fix the error buffer passed to kvm_openfiles to have a correct length
of _POSIX2_LINE_MAX.

MFC after: 1 week

23 years agoUpdate comments on RELNOTESng.
bmah [Fri, 25 May 2001 23:52:26 +0000 (23:52 +0000)]
Update comments on RELNOTESng.

23 years agoList new directories created under share/examples/isdn/* in order to
bmah [Fri, 25 May 2001 23:40:58 +0000 (23:40 +0000)]
List new directories created under share/examples/isdn/* in order to
unbreak installworld.

23 years agoFix the error buffer passed to kvm_openfiles to have a correct length
tmm [Fri, 25 May 2001 23:36:09 +0000 (23:36 +0000)]
Fix the error buffer passed to kvm_openfiles to have a correct length
of _POSIX2_LINE_MAX.

MFC after: 3 days

23 years agoMark error() as __printflike() and fix a non-exploitable format string
kris [Fri, 25 May 2001 23:15:05 +0000 (23:15 +0000)]
Mark error() as __printflike() and fix a non-exploitable format string
error.

MFC after: 1 week

23 years agoFix the error buffer passed to kvm_openfiles to have a length of
tmm [Fri, 25 May 2001 23:10:27 +0000 (23:10 +0000)]
Fix the error buffer passed to kvm_openfiles to have a length of
_POSIX2_LINE_MAX as required.
While being there, wrap an overly long line.

MFC after: 3 days

23 years agoInclude elf32.h and elf64.h to be more code compatable with Solaris.
obrien [Fri, 25 May 2001 21:46:22 +0000 (21:46 +0000)]
Include elf32.h and elf64.h to be more code compatable with Solaris.

23 years agocatch up to i386: Don't acquire Giant just to print the trap and panic
gallatin [Fri, 25 May 2001 21:03:40 +0000 (21:03 +0000)]
catch up to i386: Don't acquire Giant just to print the trap and panic

23 years agomake this compile.
gallatin [Fri, 25 May 2001 20:56:28 +0000 (20:56 +0000)]
make this compile.

23 years agoThe standard Xterm geometry is 80x24, not 80x_65_.
obrien [Fri, 25 May 2001 20:51:22 +0000 (20:51 +0000)]
The standard Xterm geometry is 80x24, not 80x_65_.

In the past 2 months or so, after rlogining into another host, the
environment has the geometry wrong.  Peter suggested that this behavior
change was most likely caused by the PAM stuff that changed to run a proper
session with a forked child.  And that for some reason the window size is
no longer being transferred via an OOB message on the socket.

This change fixes my problem and seems to be a good stopgap measure until
someone has time to ktrace/ktrace -i inetd to catch all the child processes
it spawns while doing an rlogin and change window size a few times to see
how far the change messages are getting.

23 years agoAdd warnings about trusting user-supplied data.
eric [Fri, 25 May 2001 20:42:40 +0000 (20:42 +0000)]
Add warnings about trusting user-supplied data.

Reviewed by: ru
Approved by: murray
Obtained from: OpenBSD

23 years agotrack the One True Architecture (i386)
mjacob [Fri, 25 May 2001 20:33:06 +0000 (20:33 +0000)]
track the One True Architecture (i386)

23 years agoFix a minor formatting nit
imp [Fri, 25 May 2001 19:24:58 +0000 (19:24 +0000)]
Fix a minor formatting nit

23 years agoAdd vm locking to sendfile(2) and sf_buf_free().
jhb [Fri, 25 May 2001 19:23:04 +0000 (19:23 +0000)]
Add vm locking to sendfile(2) and sf_buf_free().

Reported by: Tamiji Homma <thomma@BayNetworks.com>
Tested by: Tamiji Homma <thomma@BayNetworks.com>

23 years agoMove to using the common device list.
imp [Fri, 25 May 2001 19:22:36 +0000 (19:22 +0000)]
Move to using the common device list.
Move to table driven probing of these devices since we have such a long list.

23 years agoSeparate out pci cardbus bridge definitions to a common file for use
imp [Fri, 25 May 2001 19:12:58 +0000 (19:12 +0000)]
Separate out pci cardbus bridge definitions to a common file for use
by both OLDCARD and NEWCARD.

# didn't make the tables the same because oldcard supports more devices than
# newcard and newcard's 16-bit stuff needs some work.

23 years agoTeach gcc about '%+' so it will not warn when using '%+' in a format
obrien [Fri, 25 May 2001 19:00:07 +0000 (19:00 +0000)]
Teach gcc about '%+' so it will not warn when using '%+' in a format
string when using strftime(), since our libc supports it.

PR:             26827
Submitted by:   Marc Olzheim <marcolz@ilse.nl>

23 years agoJHB would prefer the KTR examples not be in here.
obrien [Fri, 25 May 2001 18:49:08 +0000 (18:49 +0000)]
JHB would prefer the KTR examples not be in here.

23 years agoMigrate from unit based to dev base. Don't save unit number, but do save
imp [Fri, 25 May 2001 18:28:49 +0000 (18:28 +0000)]
Migrate from unit based to dev base.  Don't save unit number, but do save
dev.  Convert all uses of unit to dev as appropriate.  Minor comment fixes
to pcic_softc definition.

23 years agoAdd a check to determine whether extended attributes have been
tmm [Fri, 25 May 2001 18:24:52 +0000 (18:24 +0000)]
Add a check to determine whether extended attributes have been
initialized on the file system before trying to grab the lock of the
per-mount extattr structure, as this lock is unitialized in that case.
This is needed because ufs_extattr_vnode_inactive is called from
ufs_inactive, which is also used by EA-unaware file systems such as
ext2fs.

Reviewed by: rwatson

23 years agoUpdate copyright info
imp [Fri, 25 May 2001 18:03:07 +0000 (18:03 +0000)]
Update copyright info

23 years agoRELNOTESng is now the default for -CURRENT release builds. (Floppy
bmah [Fri, 25 May 2001 18:01:34 +0000 (18:01 +0000)]
RELNOTESng is now the default for -CURRENT release builds.  (Floppy
images get TXT renderings only; CDROM and FTP areas get TXT and HTML.)

Remove the old *.TXT release documentation files, as they
have been subsumed into RELNOTESng.  The new layout will greatly
facilitate their maintenence and help keep them internally consistent.

23 years agoTake pcm (audio) back out of GENERIC; there appears to be some
jkh [Fri, 25 May 2001 17:55:39 +0000 (17:55 +0000)]
Take pcm (audio) back out of GENERIC; there appears to be some
concensus, most notably among the maintainers, that it's better
loaded as a module.

Finally-pushed-over-the-edge-by-the-anguished-cries-of: rwatson

23 years agoMF*.TXT: isdn4bsd updates.
bmah [Fri, 25 May 2001 17:25:47 +0000 (17:25 +0000)]
MF*.TXT:  isdn4bsd updates.

23 years agoDavid kindly added the 'zs0' to /dev/MAKEDEV to make TurboLaser users
wilko [Fri, 25 May 2001 17:16:07 +0000 (17:16 +0000)]
David kindly added the 'zs0' to /dev/MAKEDEV to make TurboLaser users
happy.

Submitted by: obrien@freebsd.org

23 years agoNew release notes: O_DIRECT flag for open(2) and fcntl(2), RFC 1323
bmah [Fri, 25 May 2001 17:08:22 +0000 (17:08 +0000)]
New release notes:  O_DIRECT flag for open(2) and fcntl(2), RFC 1323
enabled by default for TCP, ipfw(8) -d, Makefile WARNS, boot-time
enable of background_fsck, ping(8) -A, vidcontrol(1) -p, telnet(1) -u,
default vinum(8) stripe size change to 279KB, OpenSSH 0.9.6.

MFCs noted:  smbfs kernel part.

23 years agoo Merge contents of struct pcred into struct ucred. Specifically, add the
rwatson [Fri, 25 May 2001 16:59:11 +0000 (16:59 +0000)]
o Merge contents of struct pcred into struct ucred.  Specifically, add the
  real uid, saved uid, real gid, and saved gid to ucred, as well as the
  pcred->pc_uidinfo, which was associated with the real uid, only rename
  it to cr_ruidinfo so as not to conflict with cr_uidinfo, which
  corresponds to the effective uid.
o Remove p_cred from struct proc; add p_ucred to struct proc, replacing
  original macro that pointed.
  p->p_ucred to p->p_cred->pc_ucred.
o Universally update code so that it makes use of ucred instead of pcred,
  p->p_ucred instead of p->p_pcred, cr_ruidinfo instead of p_uidinfo,
  cr_{r,sv}{u,g}id instead of p_*, etc.
o Remove pcred0 and its initialization from init_main.c; initialize
  cr_ruidinfo there.
o Restruction many credential modification chunks to always crdup while
  we figure out locking and optimizations; generally speaking, this
  means moving to a structure like this:
        newcred = crdup(oldcred);
        ...
        p->p_ucred = newcred;
        crfree(oldcred);
  It's not race-free, but better than nothing.  There are also races
  in sys_process.c, all inter-process authorization, fork, exec, and
  exit.
o Remove sigio->sio_ruid since sigio->sio_ucred now contains the ruid;
  remove comments indicating that the old arrangement was a problem.
o Restructure exec1() a little to use newcred/oldcred arrangement, and
  use improved uid management primitives.
o Clean up exit1() so as to do less work in credential cleanup due to
  pcred removal.
o Clean up fork1() so as to do less work in credential cleanup and
  allocation.
o Clean up ktrcanset() to take into account changes, and move to using
  suser_xxx() instead of performing a direct uid==0 comparision.
o Improve commenting in various kern_prot.c credential modification
  calls to better document current behavior.  In a couple of places,
  current behavior is a little questionable and we need to check
  POSIX.1 to make sure it's "right".  More commenting work still
  remains to be done.
o Update credential management calls, such as crfree(), to take into
  account new ruidinfo reference.
o Modify or add the following uid and gid helper routines:
      change_euid()
      change_egid()
      change_ruid()
      change_rgid()
      change_svuid()
      change_svgid()
  In each case, the call now acts on a credential not a process, and as
  such no longer requires more complicated process locking/etc.  They
  now assume the caller will do any necessary allocation of an
  exclusive credential reference.  Each is commented to document its
  reference requirements.
o CANSIGIO() is simplified to require only credentials, not processes
  and pcreds.
o Remove lots of (p_pcred==NULL) checks.
o Add an XXX to authorization code in nfs_lock.c, since it's
  questionable, and needs to be considered carefully.
o Simplify posix4 authorization code to require only credentials, not
  processes and pcreds.  Note that this authorization, as well as
  CANSIGIO(), needs to be updated to use the p_cansignal() and
  p_cansched() centralized authorization routines, as they currently
  do not take into account some desirable restrictions that are handled
  by the centralized routines, as well as being inconsistent with other
  similar authorization instances.
o Update libkvm to take these changes into account.

Obtained from: TrustedBSD Project
Reviewed by: green, bde, jhb, freebsd-arch, freebsd-audit

23 years agoFix DMA on machines with more than 128MB.
scottl [Fri, 25 May 2001 16:05:22 +0000 (16:05 +0000)]
Fix DMA on machines with more than 128MB.

Obtained from: cg
MFC after: 5 days

23 years agoMake if_tun's clone create SI_CHEAPCLONE devices.
phk [Fri, 25 May 2001 13:32:53 +0000 (13:32 +0000)]
Make if_tun's clone create SI_CHEAPCLONE devices.

23 years agoAdd the contributed directory from the i4b tarball and the files from it
hm [Fri, 25 May 2001 13:25:59 +0000 (13:25 +0000)]
Add the contributed directory from the i4b tarball and the files from it
which would be nice to have.
Move isdnd_acct and isdnctl to the contrib directory where they belong to.

23 years agoMake the PTY drivers cloning algorithm create "CHEAPCLONE" dev_t,
phk [Fri, 25 May 2001 13:23:42 +0000 (13:23 +0000)]
Make the PTY drivers cloning algorithm create "CHEAPCLONE" dev_t,
so that some twit cannot allocate all 256 PTY's with "ls -l".

23 years agoSubmitted by: phk
hm [Fri, 25 May 2001 13:04:58 +0000 (13:04 +0000)]
Submitted by: phk
Add the V.21 modem emulator code written by phk from the old i4b tarball.

23 years agoThis script can control the state of your ISDN line. It counts
alex [Fri, 25 May 2001 12:27:00 +0000 (12:27 +0000)]
This script can control the state of your ISDN line.  It counts
how many scripts/users currently use the ISDN line and uses
"ifconfig down" if noone uses it any more.

Not objected by: hm