]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoFix logic error resulting in incorrect calculation of login time
jon [Sun, 5 Aug 2001 09:41:20 +0000 (09:41 +0000)]
Fix logic error resulting in incorrect calculation of login time

PR: bin/4204
Submitted by: Anatoly A. Orehovsky <tolik@www.tomsk.su>
MFC after: 4 weeks

22 years agoFix using /usr/bin/ftp as a slave process on the end of a pipe by calling
jon [Sun, 5 Aug 2001 08:39:58 +0000 (08:39 +0000)]
Fix using /usr/bin/ftp as a slave process on the end of a pipe by calling
fflush(stdout).  This is one ancient PR...

PR: bin/1589
Submitted by: imp

22 years agoOnce again throw out the calculation of si_iosize_max and set it to 64K.
scottl [Sun, 5 Aug 2001 07:26:31 +0000 (07:26 +0000)]
Once again throw out the calculation of si_iosize_max and set it to 64K.
Although it can go higher, it is not safe to so do on arrays with many
members.  Compromise by adding a tunable, "hw.aac.iosize_max" that can be
set at boottime.  Also document in the aac(4) manpage.

MFC after: 4 weeks

22 years agoNow that we are setting a bit in the PCIC_INT_GEN (0x3) register, we
imp [Sun, 5 Aug 2001 07:02:42 +0000 (07:02 +0000)]
Now that we are setting a bit in the PCIC_INT_GEN (0x3) register, we
can't blindly write zero into it to disable the card.  We must
preserve this bit.  This changes pcic_disable to only clear the bits
we know we need to clear on card disable, thus preserving the magic
bit for many TI bridges.

This appears to have fixed the problems that people are reporting
about the system failing to recognize cards being inserted or removed
(or both).  Greg: This may fix your problem too :-).

22 years agoOrder PCIC_INT_GEN bits large to small, like all other registers.
imp [Sun, 5 Aug 2001 06:59:32 +0000 (06:59 +0000)]
Order PCIC_INT_GEN bits large to small, like all other registers.

22 years agoBugfixes. Close a race and logic bug in the timeout handling, don't call the
scottl [Sun, 5 Aug 2001 06:20:13 +0000 (06:20 +0000)]
Bugfixes.  Close a race and logic bug in the timeout handling, don't call the
interrupt handler from the upper half, etc.  This fixes some serious stability
problems that we were seeing on our production server.  These patches have
been tested for almost 6 months and are a highly recommended MFC candidate.

Reviewed by: gibbs, merry, msmith
MFC after: 4 days

22 years agoFix a couple of copyright issues and an 80 column nit that I missed before
scottl [Sun, 5 Aug 2001 06:15:00 +0000 (06:15 +0000)]
Fix a couple of copyright issues and an 80 column nit that I missed before

22 years agovarargs -> stdarg
dd [Sun, 5 Aug 2001 05:39:16 +0000 (05:39 +0000)]
varargs -> stdarg

22 years agoAdd a missing semicolon to unbreak the kernel build with INVARIANTS
tmm [Sun, 5 Aug 2001 03:55:02 +0000 (03:55 +0000)]
Add a missing semicolon to unbreak the kernel build with INVARIANTS
(which was unfortunately turned off in the confguration I used for the
last test build).

Spotted by: jake
Pointy hat to: tmm

22 years agoSigh. Add two files needed for the sparc64 fp contect switching code
tmm [Sun, 5 Aug 2001 03:47:02 +0000 (03:47 +0000)]
Sigh. Add two files needed for the sparc64 fp contect switching code
that were forgotten in the last commit.

Pointy hat to:  tmm

22 years agoIt's bad enough to capitalize the name of the executable at the
dd [Sat, 4 Aug 2001 22:11:08 +0000 (22:11 +0000)]
It's bad enough to capitalize the name of the executable at the
beginning of the sentence, but at the very least it shouldn't be done
in the middle of a sentence!

22 years agoFix style/consistency in Makefile and repair static module building.
markm [Sat, 4 Aug 2001 21:51:14 +0000 (21:51 +0000)]
Fix style/consistency in Makefile and repair static module building.

Submitted by: bde(partially)

22 years agoDon't clobber CFLAGS
markm [Sat, 4 Aug 2001 21:49:30 +0000 (21:49 +0000)]
Don't clobber CFLAGS

Submitted by: bde

22 years agoDon't clobber CFLAGS.
markm [Sat, 4 Aug 2001 21:48:18 +0000 (21:48 +0000)]
Don't clobber CFLAGS.

Submitted by: bde

22 years agoWhitespace fixes.
jhb [Sat, 4 Aug 2001 20:49:29 +0000 (20:49 +0000)]
Whitespace fixes.

22 years agoAxe unused and invalid astpending globaldata member.
jhb [Sat, 4 Aug 2001 20:47:54 +0000 (20:47 +0000)]
Axe unused and invalid astpending globaldata member.

22 years agoAxe unused and invalid GD_ASTPENDING symbol.
jhb [Sat, 4 Aug 2001 20:46:46 +0000 (20:46 +0000)]
Axe unused and invalid GD_ASTPENDING symbol.

22 years ago- Fix a comment.
jhb [Sat, 4 Aug 2001 20:43:48 +0000 (20:43 +0000)]
- Fix a comment.
- Whitespace fixes.

Submitted by: bde

22 years agoAdd a zdestroy() function to the zone allocator. This is needed for the
tmm [Sat, 4 Aug 2001 20:17:05 +0000 (20:17 +0000)]
Add a zdestroy() function to the zone allocator. This is needed for the
unload case of modules that use their own zones.
It has been tested with the nfs module.

22 years agoAdd floating point context switching code for sparc64.
tmm [Sat, 4 Aug 2001 18:55:15 +0000 (18:55 +0000)]
Add floating point context switching code for sparc64.

Reviewed by: jake

22 years agoMake use of the new features of libdevstat to allow iostat to work on
tmm [Sat, 4 Aug 2001 18:30:54 +0000 (18:30 +0000)]
Make use of the new features of libdevstat to allow iostat to work on
crash dumps, and make it use sysctl for all data retrievals in the
"live" case (i.e. when not using iostat on a crash dump).
Remove setgid kmem for the iostat executable, it is not needed any
more after these changes.

Reviewed by: ken

22 years agoAdd some features to libdevstat, and overhaul the interface a bit:
tmm [Sat, 4 Aug 2001 18:25:48 +0000 (18:25 +0000)]
Add some features to libdevstat, and overhaul the interface a bit:

1.) prefix all functions in the library with devstat_ (compatability
    functions are available for all functions that were chaned in an
    incompatible way, but are deprecated).
2.) Add a pointer to a kvm_t as the first argument to functions that
    used to get their information via sysctl; they behave the same
    as before when NULL is passed as this argument, otherwise, the
    information is obtained via libkvm using the supplied handle.
3.) Add a new function, devstat_compute_statistics(), that is intended
    to replace the old compute_stats() function. It offers more
    statistics data, and has a more flexible interface.

libdevstat does now require libkvm; a library depedency is added, so
that libkvm only needs to be explicitely specified for statically linked
programs.
The library major version number is bumped.

Submitted by: Sergey A. Osokin <osa@freebsd.org.ru>, ken (3)
Reviewed by: ken

22 years agoExport the tk_nin and tk_nout variables (number of tty input/output
tmm [Sat, 4 Aug 2001 18:09:24 +0000 (18:09 +0000)]
Export the tk_nin and tk_nout variables (number of tty input/output
characters) as sysctls (kern.tty_nin and kern.tty_nout).

22 years agoExport the head structure for the device statistics STAILQ in
tmm [Sat, 4 Aug 2001 18:02:47 +0000 (18:02 +0000)]
Export the head structure for the device statistics STAILQ in
sys/devicestat.h, so that the queue can be walked in crashdumps using
libkvm.

22 years agoWhen running aplication joined multicast address,
ume [Sat, 4 Aug 2001 17:10:14 +0000 (17:10 +0000)]
When running aplication joined multicast address,
removing network card, and kill aplication.
imo_membership[].inm_ifp refer interface pointer
after removing interface.
When kill aplication, release socket,and imo_membership.
imo_membership use already not exist interface pointer.
Then, kernel panic.

PR: 29345
Submitted by: Inoue Yuichi <inoue@nd.net.fujitsu.co.jp>
Obtained from: KAME
MFC after: 3 days

22 years agoFix the bug where this modulke was not checking the priamry GID, only
markm [Sat, 4 Aug 2001 09:19:31 +0000 (09:19 +0000)]
Fix the bug where this modulke was not checking the priamry GID, only
the GIDS in /etc/group or NIS's group map.

Tested by: sheldonh
PR: 29349

22 years agoKill debug output for sn device. It is confusing the natives
imp [Sat, 4 Aug 2001 05:27:52 +0000 (05:27 +0000)]
Kill debug output for sn device.  It is confusing the natives

22 years agoLink to libcipher in the usual way. `bdes' depended on a nonexistent
bde [Fri, 3 Aug 2001 22:28:25 +0000 (22:28 +0000)]
Link to libcipher in the usual way.  `bdes' depended on a nonexistent
library.  This only worked because of the undocmented feature of make(1)
that targets named foo.a are always up to date.

Fixed some style bugs.

22 years agoDon't clobber the default for CFLAGS.
bde [Fri, 3 Aug 2001 21:45:54 +0000 (21:45 +0000)]
Don't clobber the default for CFLAGS.

22 years agoAdd KTR_INTR tracepoints for when clock interrupts are triggered.
jhb [Fri, 3 Aug 2001 20:54:41 +0000 (20:54 +0000)]
Add KTR_INTR tracepoints for when clock interrupts are triggered.

22 years agoWhen global anycast address was assigned to lo0, wrong source
ume [Fri, 3 Aug 2001 19:10:31 +0000 (19:10 +0000)]
When global anycast address was assigned to lo0, wrong source
address was selected.

Reported by: Shingo WATANABE <nabe@nabechan.org>
Submitted by: JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp>
MFC after: 3 days

22 years agoAnton kindly pointed out (and fixed) a bug in the Jail handling of the
rwatson [Fri, 3 Aug 2001 18:21:06 +0000 (18:21 +0000)]
Anton kindly pointed out (and fixed) a bug in the Jail handling of the
bind() call on IPv4 sockets:

  Currently, if one tries to bind a socket using INADDR_LOOPBACK inside a
  jail, it will fail because prison_ip() does not take this possibility
  into account.  On the other hand, when one tries to connect(), for
  example, to localhost, prison_remote_ip() will silently convert
  INADDR_LOOPBACK to the jail's IP address.  Therefore, it is desirable to
  make bind() to do this implicit conversion as well.

  Apart from this, the patch also replaces 0x7f000001 in
  prison_remote_ip() to a more correct INADDR_LOOPBACK.

This is a 4.4-RELEASE "during the freeze, thanks" MFC candidate.

Submitted by: Anton Berezin <tobez@FreeBSD.org>
Discussed with at some point: phk
MFC after: 3 days

22 years agoRemove dangling prototype for the now defunct procfs_kmemaccess()
rwatson [Fri, 3 Aug 2001 17:51:05 +0000 (17:51 +0000)]
Remove dangling prototype for the now defunct procfs_kmemaccess()
call.

Obtained from: TrustedBSD Project

22 years agoMFS: Avoid dropping fragments in the absence of an interface address.
dcs [Fri, 3 Aug 2001 17:36:06 +0000 (17:36 +0000)]
MFS: Avoid dropping fragments in the absence of an interface address.

Noticed by: fenner
Submitted by: iedowse
Not committed to current by: iedowse ;-)

22 years agoCollapse a Pmem case in with the other debugging files case for procfs,
rwatson [Fri, 3 Aug 2001 17:20:34 +0000 (17:20 +0000)]
Collapse a Pmem case in with the other debugging files case for procfs,
as there are now "unusual" protection properties to Pmem that differ
from the other files.  While I'm at it, introduce proc locking for
the other files, which was previously present only in the Pmem case.

Obtained from: TrustedBSD Project

22 years agoRemove read permission for group on the /proc/*/mem file, since kmem
rwatson [Fri, 3 Aug 2001 17:15:40 +0000 (17:15 +0000)]
Remove read permission for group on the /proc/*/mem file, since kmem
no longer requires access.

Reviewed by: tmm
Obtained from: TrustedBSD Project

22 years agoPrior to support for almost all ps activity via sysctl, ps used procfs,
rwatson [Fri, 3 Aug 2001 17:13:23 +0000 (17:13 +0000)]
Prior to support for almost all ps activity via sysctl, ps used procfs,
and so special-casing was introduced to provide extra procfs privilege
to the kmem group.  With the advent of non-setgid kmem ps, this code
is no longer required, and in fact, can is potentially harmful as it
allocates privilege to a gid that is increasingly less meaningful.
Knowledge of specific gid's in kernel is also generally bad precedent,
as the kernel security policy doesn't distinguish gid's specifically,
only uid 0.

This commit removes reference to kmem in procfs, both in terms of
access control decisions, and the applying of gid kmem to the
/proc/*/mem file, simplifying the associated code considerably.
Processes are still permitted to access the mem file based on
the debugging policy, so ps -e still works fine for normal
processes and use.

Reviewed by: tmm
Obtained from: TrustedBSD Project

22 years agoDon't terminate the uiomove() loop on a zero-length mbuf. It's not
fenner [Fri, 3 Aug 2001 16:51:53 +0000 (16:51 +0000)]
Don't terminate the uiomove() loop on a zero-length mbuf.  It's not
 particularly nice that IPSEC inserts a zero-length mbuf into the
 chain, and that bug should be fixed too, but interfaces should be
 robust to bad input.
Print the interface name when TUNDEBUG()ing about dropping an mbuf.

22 years agoConvert to using the same method of detecting the secure directory as other
gshapiro [Fri, 3 Aug 2001 16:22:47 +0000 (16:22 +0000)]
Convert to using the same method of detecting the secure directory as other
usr.sbin programs.

Noted by: markm

22 years agoRevamp and diff-reduce the various secure telnets. Make sure that
markm [Fri, 3 Aug 2001 16:03:26 +0000 (16:03 +0000)]
Revamp and diff-reduce the various secure telnets. Make sure that
Kerberos5 has _a_ telnet (which is not currently K5 enabled).
Incorporate BDE's static linking fixes.

22 years agoNote MFC of file 3.36.
dd [Fri, 3 Aug 2001 12:31:43 +0000 (12:31 +0000)]
Note MFC of file 3.36.

22 years agoMiscellaneous fixes.
sheldonh [Fri, 3 Aug 2001 11:31:39 +0000 (11:31 +0000)]
Miscellaneous fixes.

In particularly, the -h and -m flags were incorrectly described.

22 years agoFix warnings from `-Wmissing-prototypes'. Staticize.
tg [Fri, 3 Aug 2001 10:47:56 +0000 (10:47 +0000)]
Fix warnings from `-Wmissing-prototypes'. Staticize.

22 years agoSort SRCS.
tg [Fri, 3 Aug 2001 10:47:13 +0000 (10:47 +0000)]
Sort SRCS.

22 years agoReverse the logic here again with regards to "trusted" ACPI timer
msmith [Fri, 3 Aug 2001 09:52:53 +0000 (09:52 +0000)]
Reverse the logic here again with regards to "trusted" ACPI timer
implementations.  More of them seem to be broken, so only "trust"
timers we know work.

22 years agoWarn when a CHAP81 SUCCESS packet is invalid due to an incorrect S= value.
brian [Fri, 3 Aug 2001 09:28:21 +0000 (09:28 +0000)]
Warn when a CHAP81 SUCCESS packet is invalid due to an incorrect S= value.

22 years agoShoud build resources in the _CRS buffer. Oops.
msmith [Fri, 3 Aug 2001 08:38:49 +0000 (08:38 +0000)]
Shoud build resources in the _CRS buffer.  Oops.

Submitted by: "neckpain@nettaxi.com" <neckpain@nettaxi.com>

22 years agoMove the resource pointer when we reallocate the buffer.
msmith [Fri, 3 Aug 2001 08:38:11 +0000 (08:38 +0000)]
Move the resource pointer when we reallocate the buffer.

Submitted by: "neckpain@nettaxi.com" <neckpain@nettaxi.com>

22 years agoRename mb_init() mbuf subsystem initialization routine to mbuf_init(), in
bmilekic [Fri, 3 Aug 2001 05:05:32 +0000 (05:05 +0000)]
Rename mb_init() mbuf subsystem initialization routine to mbuf_init(), in
order to avoid namespace collision with subr_mchain.c's mb_init(). This
wasn't "fatal" as the mbuf initialization routine mb_init() was local to
subr_mbuf.c which in turn didn't pull in subr_mchain.c's mb_init()
declaration, but it should deffinately be changed now before it creates
headache.

22 years agoRemove some code that appears to have endian problems with INVARIANTS.
jake [Fri, 3 Aug 2001 03:31:45 +0000 (03:31 +0000)]
Remove some code that appears to have endian problems with INVARIANTS.
This is #if BIG_ENDIAN, but is only necessary if malloc types are shorts,
not struct malloc_type * like they are now.

22 years agoMove some code related to managing pv entries from the pmap module to
jake [Fri, 3 Aug 2001 01:27:15 +0000 (01:27 +0000)]
Move some code related to managing pv entries from the pmap module to
the pv module.  It works now that vtophys for sttes works.

22 years agoFix a bug translating virtual translation table entry addresses to physical
jake [Fri, 3 Aug 2001 01:21:24 +0000 (01:21 +0000)]
Fix a bug translating virtual translation table entry addresses to physical
addresses.  It helps to use the physical address that the virtual address
actually maps to (doh!).  Comment out some code that crashes.

Found independently by: tmm

22 years agoDefine proc0paddr. Call init_param() as early as possible.
jake [Fri, 3 Aug 2001 01:11:49 +0000 (01:11 +0000)]
Define proc0paddr.  Call init_param() as early as possible.

22 years agoAdd a Makefile, ldscript, and config magic for sparc64. This is tailored
jake [Fri, 3 Aug 2001 01:09:10 +0000 (01:09 +0000)]
Add a Makefile, ldscript, and config magic for sparc64.  This is tailored
to build with a cross compiler alongside the standard compiler; it would be
more desirable to build in a chroot.

22 years agoAdd an Elfhashelt type for sparc64.
jake [Fri, 3 Aug 2001 01:00:41 +0000 (01:00 +0000)]
Add an Elfhashelt type for sparc64.

22 years agoReformat for 80 columns. Sorry, but I had to do it.
scottl [Fri, 3 Aug 2001 00:50:30 +0000 (00:50 +0000)]
Reformat for 80 columns.  Sorry, but I had to do it.

22 years agoFixes file descriptor leak in standalone mode.
jon [Fri, 3 Aug 2001 00:23:37 +0000 (00:23 +0000)]
Fixes file descriptor leak in standalone mode.
Prevents simultaneous calls to updatestat() as function is not reentrant.

PR: bin/24857
Submitted by: Martin Butkus <mb@bagheera.thgwf.de>
MFC after: 1 month

22 years agoFurther Makefile.* sync (from Makefile.ia64). The lint target has been
peter [Fri, 3 Aug 2001 00:07:54 +0000 (00:07 +0000)]
Further Makefile.* sync (from Makefile.ia64).  The lint target has been
commented out in the entire life of the 2.x+ branch and given the amount
of gcc-specific code we have and the warning checks that gcc does I'm not
sure that it is going to get us much for some time.

22 years agoMove MKMODULESENV+= *after* MKMODULESENV=
peter [Fri, 3 Aug 2001 00:01:13 +0000 (00:01 +0000)]
Move MKMODULESENV+= *after* MKMODULESENV=

22 years agoAdd a cosmetic comment.
peter [Thu, 2 Aug 2001 23:54:23 +0000 (23:54 +0000)]
Add a cosmetic comment.

22 years agoUse 'p' instead of the potentially more expensive 'curproc' inside of
jhb [Thu, 2 Aug 2001 22:15:31 +0000 (22:15 +0000)]
Use 'p' instead of the potentially more expensive 'curproc' inside of
mi_switch().

22 years agoSet up the via_chinfo structures properly so we write to the correct
greid [Thu, 2 Aug 2001 22:13:10 +0000 (22:13 +0000)]
Set up the via_chinfo structures properly so we write to the correct
registers later on; this fixes the VIA82C686 sound problems recently
reported by a number of people.

22 years agoFix a few bugs, some of which I introduced in recent commits:
iedowse [Thu, 2 Aug 2001 21:46:21 +0000 (21:46 +0000)]
Fix a few bugs, some of which I introduced in recent commits:
- clean_mtab():
    Actually use the strdup'd version of the host that we go to the
    trouble of creating.
- do_umntall/do_umount:
    Don't return success if clnt_create() fails.
    Don't access a client pointer after it has been destroyed.
    Remember to destroy the authentication information we created.

22 years agoIn getclnthandle(), if the address is found in the cache we need
iedowse [Thu, 2 Aug 2001 21:31:21 +0000 (21:31 +0000)]
In getclnthandle(), if the address is found in the cache we need
to strdup() the address string before returning it via *targaddr
because the caller will free the string.

Change the comment at the top of getclnthandle() to clarify that
the caller is responsible for freeing *targaddr.

Noticed by: sobomax

22 years agoMake the fmt arguments to make_dev and make_dev_alias const char *.
imp [Thu, 2 Aug 2001 20:35:35 +0000 (20:35 +0000)]
Make the fmt arguments to make_dev and make_dev_alias const char *.

Approved on IRC as long as it didn't cause a large number of warnings by: phk

MFC After: 700 hours

22 years agoThe wrong-last-byte bug on win98 chap responses is also in winME
brian [Thu, 2 Aug 2001 20:12:48 +0000 (20:12 +0000)]
The wrong-last-byte bug on win98 chap responses is also in winME

22 years agofix memory leak when error during opening of routing socket
jon [Thu, 2 Aug 2001 19:56:29 +0000 (19:56 +0000)]
fix memory leak when error during opening of routing socket

PR: kern/29336
Submitted by: Richard Andrades <richard@xebeo.com>
MFC after: 1 month

22 years agoWith the S/KEY removal, this is no longer buildable or necessary.
markm [Thu, 2 Aug 2001 19:04:20 +0000 (19:04 +0000)]
With the S/KEY removal, this is no longer buildable or necessary.

22 years agoDon't try to make pam_ssh module if NO_OPENSSH is set.
markm [Thu, 2 Aug 2001 19:01:02 +0000 (19:01 +0000)]
Don't try to make pam_ssh module if NO_OPENSSH is set.

22 years agoAdd opieaccess(5) functionality under the INSECURE_OPIE .ifdef.
markm [Thu, 2 Aug 2001 18:58:52 +0000 (18:58 +0000)]
Add opieaccess(5) functionality under the INSECURE_OPIE .ifdef.

Asked for by: ache

22 years agoexists(../../crypto) --> exists(${.CURDIR}/../../crypto)
sobomax [Thu, 2 Aug 2001 18:21:48 +0000 (18:21 +0000)]
exists(../../crypto) --> exists(${.CURDIR}/../../crypto)

MFC after: 1 month

22 years agofix for pkg_add to symlinked prefix directories that are more than 1 link deep.
jon [Thu, 2 Aug 2001 18:20:27 +0000 (18:20 +0000)]
fix for pkg_add to symlinked prefix directories that are more than 1 link deep.

PR: bin/28274
Submitted by: John Hein <jhein@timing.com>
MFC after: 1 month

22 years agoAllow configTtys() be called several times - set VAR_CONSTERM to "NO" after
ache [Thu, 2 Aug 2001 16:36:21 +0000 (16:36 +0000)]
Allow configTtys() be called several times - set VAR_CONSTERM to "NO" after
operation done.

22 years agoFix a cryptoless world by disconnecting libmp from the build when there is no
sobomax [Thu, 2 Aug 2001 15:47:03 +0000 (15:47 +0000)]
Fix a cryptoless world by disconnecting libmp from the build when there is no
crypto bits installed and/or NOCRYPTO/NO_OPENSSL is defined. This unfortunately
meants that usr.bin/chkey, usr.bin/newkey and usr.sbin/keyserv have also to
be disconnected.

IMO it is merely a workaround, the proper solution is to move libmp to
src/crypto where it belongs and use libgmp for the cryptoless builds instead.

Missed by: dd

22 years agoInclude opt_splash.h.
yokota [Thu, 2 Aug 2001 13:23:17 +0000 (13:23 +0000)]
Include opt_splash.h.

22 years agoUse #ifdef DEV_SPLASH (from opt_splash.h) rather than
yokota [Thu, 2 Aug 2001 13:22:33 +0000 (13:22 +0000)]
Use #ifdef DEV_SPLASH (from opt_splash.h) rather than
#if NSPLASH > 0 (from splash.h) to test the presence
of the splash driver.

22 years agoCosmetics: replace dozen instances of "(tmp = getenv(PKG_DBDIR) ? tmp : DEF_LOG_DIR)"
sobomax [Thu, 2 Aug 2001 13:13:06 +0000 (13:13 +0000)]
Cosmetics: replace dozen instances of "(tmp = getenv(PKG_DBDIR) ? tmp : DEF_LOG_DIR)"
with macro.

MFC after: 1 month

22 years agoCosmetics: kill blank lines at the end of file.
sobomax [Thu, 2 Aug 2001 13:08:43 +0000 (13:08 +0000)]
Cosmetics: kill blank lines at the end of file.

MFC after: 1 month

22 years agoRemove duplicate ufsd entry. This seems to go back to the Sun's version
dwmalone [Thu, 2 Aug 2001 12:55:37 +0000 (12:55 +0000)]
Remove duplicate ufsd entry. This seems to go back to the Sun's version
of the file.

PR: 29386
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>

22 years agoUsability tweak:
sobomax [Thu, 2 Aug 2001 12:38:29 +0000 (12:38 +0000)]
Usability tweak:

  Use '' quotes instead of `' to delimit names of files and packages in
  warning and error messages, because it is easier to cut-n-paste name in
  question that way (single click) without confusing the shell. And yes,
  I know that it is less eye-candy...

MFC after: 1 month

22 years agoDocument the deprecated `-w' option in the COMPATIBILITY section.
yar [Thu, 2 Aug 2001 12:38:23 +0000 (12:38 +0000)]
Document the deprecated `-w' option in the COMPATIBILITY section.

22 years agoWhen there is a file that can't be deleted due to checksum mismatch print name
sobomax [Thu, 2 Aug 2001 12:19:32 +0000 (12:19 +0000)]
When there is a file that can't be deleted due to checksum mismatch print name
of that file to stdout to simplify debugging. IMO it was a mistake to print
this warning only when `verbose' mode is on.

MFC after: 1 month

22 years agoWhen talking about new versions, use the word "updated" instead of
dd [Thu, 2 Aug 2001 12:16:17 +0000 (12:16 +0000)]
When talking about new versions, use the word "updated" instead of
"upgraded" for consistency.  Prior to this commit, 9 of the entires
used the latter, and 51 used the former.

22 years agoNote MFC or Binutils 2.11.2 (what we have is close enough to that
dd [Thu, 2 Aug 2001 12:08:10 +0000 (12:08 +0000)]
Note MFC or Binutils 2.11.2 (what we have is close enough to that
version that there's no need to talk about the snapshot business; all
it would do is serve to confuse).

22 years agoAdd FBIO_BLANK ioctl support. Return ENODEV for yet-to-be-
yokota [Thu, 2 Aug 2001 11:26:30 +0000 (11:26 +0000)]
Add FBIO_BLANK ioctl support.  Return ENODEV for yet-to-be-
supported ioctls for now.

22 years agoAdd some definitions. Their actual support will be added
yokota [Thu, 2 Aug 2001 11:17:38 +0000 (11:17 +0000)]
Add some definitions. Their actual support will be added
to video drivers later.

22 years agoWhen building a debugging kernel with modules, build modules with
sheldonh [Thu, 2 Aug 2001 10:56:59 +0000 (10:56 +0000)]
When building a debugging kernel with modules, build modules with
debugging support as well.

This relies on support added in rev 1.105 to kmod.mk.

Requested by: peter

22 years agoRepair the get/set UID() stuff so this works in both su(1) and login(1)
markm [Thu, 2 Aug 2001 10:35:41 +0000 (10:35 +0000)]
Repair the get/set UID() stuff so this works in both su(1) and login(1)
modes.

22 years ago- Deny detaching requests until device is still open, otherwise it is possible
sobomax [Thu, 2 Aug 2001 10:19:13 +0000 (10:19 +0000)]
- Deny detaching requests until device is still open, otherwise it is possible
  to hang or panic kernel by detaching disk from which fs is mounted;
- replace "md" with MD_NAME in yet another place.

Reviewed by: phk
Approved by: phk

22 years agoAdd a ``nat punch_fw'' command for punching FTP and IRC DCC holes through
brian [Thu, 2 Aug 2001 10:16:32 +0000 (10:16 +0000)]
Add a ``nat punch_fw'' command for punching FTP and IRC DCC holes through
the firewall.

22 years agoPack struct uniqtag declarations to stop our data field from being pushed
brian [Thu, 2 Aug 2001 09:28:31 +0000 (09:28 +0000)]
Pack struct uniqtag declarations to stop our data field from being pushed
4 bytes to the right on the alpha.

Tested by: Thomas Pornin <Thomas.Pornin@ens.fr>
MFC after: 1 week

22 years agoWhen building a debugging kernel with modules, build modules with
sheldonh [Thu, 2 Aug 2001 09:22:18 +0000 (09:22 +0000)]
When building a debugging kernel with modules, build modules with
debugging support as well.  Debugging module support is handled
identically to kernel debugging support, right down to poor
choice of make variable names.

22 years agoRefine cursor type/shape control escape sequences and
yokota [Thu, 2 Aug 2001 08:30:40 +0000 (08:30 +0000)]
Refine cursor type/shape control escape sequences and
ioctls. We can now add ve, vi and vs capabilities to
cons25 in termcap.

Discussed with and tested by: ache

22 years agoFixups for the initial allocation by dillon:
alfred [Thu, 2 Aug 2001 07:54:58 +0000 (07:54 +0000)]
Fixups for the initial allocation by dillon:
  1) allocate fewer buckets
  2) when failing to allocate swap zone, keep reducing the zone by
     a third rather than a half in order to reduce the chance of
     allocating way too little.

I also moved around some code for readability.

Suggested by: dillon
Reviewed by: dillon

22 years agoOnly try to allocated properly aligned I/O segments. This should stop
imp [Thu, 2 Aug 2001 07:06:32 +0000 (07:06 +0000)]
Only try to allocated properly aligned I/O segments.  This should stop
some of the config problems that we've been seeing (where wi0 tries to
allocate 0x138-0x198, for example).

Use err(1,"foo") rather than perror + exit while I'm here.

22 years agoFix thinko: FORCE_PKG_RESIDENT -> FORCE_PKG_REGISTER
greid [Thu, 2 Aug 2001 04:25:06 +0000 (04:25 +0000)]
Fix thinko: FORCE_PKG_RESIDENT -> FORCE_PKG_REGISTER

Submitted by: Raymond Kohler <ray.kohler@mail.com>

22 years agoAdd the ability to modify /etc/ttys before first reboot during the
rwatson [Thu, 2 Aug 2001 03:53:36 +0000 (03:53 +0000)]
Add the ability to modify /etc/ttys before first reboot during the
system installation process.  This allows users installing via serial
console to enable serial console login during the installation
process using an un-customized install.  The user is not prompted to
modify /etc/ttys during a normal install, but is offered the
opportunity during post-install configuration.

- Introduce configTTYs(), which describes the benefits of editing
  /etc/ttys, and asks for confirmation before spawning the editor.
- add configTTYs to the post-install configuration, as well as to
  the global configuration index.

22 years agoCompensate for default disabling of network services in inetd.conf(5)
rwatson [Thu, 2 Aug 2001 03:25:16 +0000 (03:25 +0000)]
Compensate for default disabling of network services in inetd.conf(5)
by providing the opportunity to edit inetd.conf during the system
installation process.  The following modifications were made:

(1) Expand the Anonymous FTP description dialog to indicate that inetd
    and ftpd must be enabled before it can be used.

(2) Introduce a new configInetd() pair of dialogs, the first describing
    inetd, giving a couple of examples of services that require it, and
    hinting at potential risk, then asking the user if they wish to
    enable it.  The second indicates that inetd.conf must be configured
    to enabled specific services, and asks if the user would like to
    load inetd.conf into the editor to modify it.  Add this
    configuration action to the index.

There are some further improvements that might be considered:

(1) Provide a more inetd.conf-specific configuration tool that speaks
    inetd.conf(5).  However, this is made difficult by the "yet another
    configuration format" nature of inetd.conf, as well as its use of
    commenting to disable services, rather than an in-syntax way to
    disable a service without commenting it out.  Submissions here
    would probably be welcome.

(2) There's some overlap between settings in the somewhat obtuse
    Security Profile mechanism and other settings, including the inetd
    setting, and NFS server configuration.  As features become
    individually tunable, they should probably be removed from the
    security profile mechanism.  Otherwise, somewhat counter-intuitively,
    sysinstall (in practice) queries multiple times whether inetd, nfsd,
    etc, should be enabled/disabled.  A possible future direction might
    be to drive profiles not by degree of paranoia, rather, the set
    of services desired.  Or simply to remove the Security Profile
    mechanism and resort to feature-driven configuration.

Reviewed by: imp, chris, jake, nate, -arch, -stable

22 years agoDefault to disabling all inetd.conf entries, in particular, telnetd
rwatson [Thu, 2 Aug 2001 02:19:56 +0000 (02:19 +0000)]
Default to disabling all inetd.conf entries, in particular, telnetd
and ftpd.  This more conservative default reduces the exposure of
freshly installed machines, which is especially valuable for machines
that receive minimal further configuration before being put into
production.  Generally speaking, SSH has superseded the use of both
telnet and ftp in many environments.  In light of recent remotely
exploitable security holes in both telnetd and ftpd, this choice
retains flexibility (both telnetd and ftpd daemons remain installed
and easily enableable) while protecting users who don't need the
additional risk.  This change brings our configuration into line with
the majority of other UNIX vendors, including OpenBSD and NetBSD.

To address the concerns of those requiring remote access via telnet
from first install, changes will shortly be committed to sysinstall
to provide the ability to edit inetd.conf during the installation
process, allowing telnetd and ftp to be re-enabled during the
installation process.

While I'm at it, slightly improve commenting for inetd.conf so that
it's more clear to users how to enable and disable services.
Further commenting to indicate the functions of various columns would
probably also be useful.

Reviewed by: imp, chris, jake, nate, -arch, -stable

22 years agoOops- don't set 'goal' twice when you mean to set 'nvrm' as well.
mjacob [Thu, 2 Aug 2001 00:34:56 +0000 (00:34 +0000)]
Oops- don't set 'goal' twice when you mean to set 'nvrm' as well.
This breaks bogus NVRAM boards.

MFC after: 1 day