]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
27 years agoFixed wrong/incomplete #includes in synopsis.
Bruce Evans [Sun, 13 Apr 1997 13:02:07 +0000 (13:02 +0000)]
Fixed wrong/incomplete #includes in synopsis.

27 years agoFixed wrong function return type in synopsis.
Bruce Evans [Sun, 13 Apr 1997 13:01:05 +0000 (13:01 +0000)]
Fixed wrong function return type in synopsis.

27 years agoDon't use literal semicolons in .Fn macro invocations.
Bruce Evans [Sun, 13 Apr 1997 12:55:36 +0000 (12:55 +0000)]
Don't use literal semicolons in .Fn macro invocations.

27 years agoFixed synopsis. There was a missing arg, 3 args with the wrong type,
Bruce Evans [Sun, 13 Apr 1997 12:36:42 +0000 (12:36 +0000)]
Fixed synopsis.  There was a missing arg, 3 args with the wrong type,
and many apparent dependendcies on <sys/types.h> from use of u_char.
<dialog.h> avoids u_char, so avoid it here too.

27 years agoFixed spelling of __set_ospeed (was _set_ospeed) and improved
Bruce Evans [Sun, 13 Apr 1997 12:16:59 +0000 (12:16 +0000)]
Fixed spelling of __set_ospeed (was _set_ospeed) and improved
nearby English.

27 years agoDeclare the documented (modulo a spelling error) interface
Bruce Evans [Sun, 13 Apr 1997 11:41:59 +0000 (11:41 +0000)]
Declare the documented (modulo a spelling error) interface
`void __set_ospeed(long);' in the appropriate header.

The implementation still uses speed_t instead of long.  This
will break properly when speed_t is fixed (speed_t shall be
unsigned ...).

27 years agoThis file is deprecated not "obsoleted".
Paul Richards [Sun, 13 Apr 1997 11:33:17 +0000 (11:33 +0000)]
This file is deprecated not "obsoleted".

27 years agoRemoved bogus unused #include.
Bruce Evans [Sun, 13 Apr 1997 11:31:20 +0000 (11:31 +0000)]
Removed bogus unused #include.

27 years agoRemoved bogus unused #include.
Bruce Evans [Sun, 13 Apr 1997 11:24:29 +0000 (11:24 +0000)]
Removed bogus unused #include.

27 years agoUpdated to use new MOD_SYSCALL() and MOD_DISPATCH() macros.
Bruce Evans [Sun, 13 Apr 1997 11:10:40 +0000 (11:10 +0000)]
Updated to use new MOD_SYSCALL() and MOD_DISPATCH() macros.

Removed bogus unused #include and unused #includes.

27 years agoAdd more bootstrapping bandaids: copy over some of the host system
Joerg Wunsch [Sun, 13 Apr 1997 08:46:58 +0000 (08:46 +0000)]
Add more bootstrapping bandaids: copy over some of the host system
dependent utilities to /bootstrap inside the chroot tree, so the
hosting and targeted environment are not required the same.  By now,
mount/umount (due to the Lite2 changes) and cpio (due to lchown) make
it there.  This should allow building a 3.0-current release from
inside a 2.2 vintage system, and visa verse.

Isn't lchown also a reason for a shared lib version number bump?

PR:
Reviewed by: jkh

27 years agoDon't create anything on the disk if -O (aka --to-stdout) was given.
Joerg Wunsch [Sun, 13 Apr 1997 08:39:38 +0000 (08:39 +0000)]
Don't create anything on the disk if -O (aka --to-stdout) was given.

PR: gnu/3247: tar -O creates directories...

27 years agoAdd support for GLOBAL style tags.
Jordan K. Hubbard [Sun, 13 Apr 1997 06:47:02 +0000 (06:47 +0000)]
Add support for GLOBAL style tags.

27 years agoSupport GLOBAL style tags.
Jordan K. Hubbard [Sun, 13 Apr 1997 06:44:25 +0000 (06:44 +0000)]
Support GLOBAL style tags.

27 years agoAdd a -b option to make the resulting CDROM image bootable
Jordan K. Hubbard [Sun, 13 Apr 1997 06:43:58 +0000 (06:43 +0000)]
Add a -b option to make the resulting CDROM image bootable
(requires new mkisofs 1.10 port).

27 years agoClose PR#3118
Jordan K. Hubbard [Sun, 13 Apr 1997 06:38:56 +0000 (06:38 +0000)]
Close PR#3118

27 years agoThe function union_fsync tries to lock overlaying vnode object when
Poul-Henning Kamp [Sun, 13 Apr 1997 06:29:13 +0000 (06:29 +0000)]
The function union_fsync tries to lock overlaying vnode object when
dolock is not set (that is, targetvp == overlaying vnode object).
Current code use FIXUP macro to do this, and never unlocks overlaying
vnode object in union_fsync.  So, the vnode object will be locked
twice and never unlocked.

PR: 3271
Submitted by: kato

27 years agoThe path name buffer, cn->cn_pnbuf, is FREEed by VOP_MKDIR when
Poul-Henning Kamp [Sun, 13 Apr 1997 06:27:09 +0000 (06:27 +0000)]
The path name buffer, cn->cn_pnbuf, is FREEed by VOP_MKDIR when
relookup() in union_relookup() is succeeded.  However, if relookup()
returns non-zero value, that is relookup fails, VOP_MKDIR is never
called (c.f. union_mkshadow).  Thus, pathname buffer is never FREEed.

Reviewed by: phk
Submitted by: kato
PR: 3262

27 years agoThough malloc allocates only cn.cn_namelen bytes for cn.cn_pnbuf in
Poul-Henning Kamp [Sun, 13 Apr 1997 06:25:03 +0000 (06:25 +0000)]
Though malloc allocates only cn.cn_namelen bytes for cn.cn_pnbuf in
union_vn_create(), following bcopy copies cn.cn_namlen + 1 bytes to
cn.cn_pnbuf

PR: 3255
Reviewed by: phk
Submitted by: kato

27 years agoRemoved libc stuffs. Lite/2 merged libc contains unionfs support, and
KATO Takenori [Sun, 13 Apr 1997 06:12:13 +0000 (06:12 +0000)]
Removed libc stuffs.  Lite/2 merged libc contains unionfs support, and
we don't have to use the libc related files in sys/misc/unionfs
directory.

27 years agoSynchronize with sys/i386/i386/machdep.c revision 1.237.
KATO Takenori [Sun, 13 Apr 1997 06:02:52 +0000 (06:02 +0000)]
Synchronize with sys/i386/i386/machdep.c revision 1.237.

27 years agoDecrease the amount of memory allocated for bouncing. This will
John Dyson [Sun, 13 Apr 1997 04:07:24 +0000 (04:07 +0000)]
Decrease the amount of memory allocated for bouncing.  This will
allow large systems to boot successfully with bounce buffers compiled
in.  We are now limiting bounce space to 512K.  The 8MB allocated for
a 512MB system is very bogus -- and that is now fixed.

27 years agoThe pmap code was too generous in the allocation of kva space for
John Dyson [Sun, 13 Apr 1997 03:35:30 +0000 (03:35 +0000)]
The pmap code was too generous in the allocation of kva space for
the pv entries.  This problem has become obvious due to the increase
in the size of the pv entries.  We need to create a more intelligent
policy for pv entry management eventually.
Submitted by: David Greenman <dg@freebsd.org>

27 years agoImprove the buffer cache memory policy by moving pages over to the
John Dyson [Sun, 13 Apr 1997 03:33:25 +0000 (03:33 +0000)]
Improve the buffer cache memory policy by moving pages over to the
cache queue more often.  The pageout daemon had to be waken up
more often than necessary since pages were not put on the
cache queue, when they should have been.
Submitted by: David Greenman <dg@freebsd.org>

27 years agoCorrect the previous thread-fix commit. I made a clerical error.
John Dyson [Sun, 13 Apr 1997 03:05:31 +0000 (03:05 +0000)]
Correct the previous thread-fix commit.  I made a clerical error.

27 years agoFully implement vfork. Vfork is now much much faster than even our
John Dyson [Sun, 13 Apr 1997 01:48:35 +0000 (01:48 +0000)]
Fully implement vfork.  Vfork is now much much faster than even our
fork. (On my machine, fork is about 240usecs, vfork is 78usecs.)

Implement rfork(!RFPROC !RFMEM), which allows a thread to divorce its memory
from the other threads of a group.

Implement rfork(!RFPROC RFCFDG), which closes all file descriptors, eliminating
possible existing shares with other threads/processes.

Implement rfork(!RFPROC RFFDG), which divorces the file descriptors for a
thread from the rest of the group.

Fix the case where a thread does an exec.  It is almost nonsense for a thread
to modify the other threads address space by an exec, so we
now automatically divorce the address space before modifying it.

27 years agoUpgrade to the 1997/02/26 version.
Jean-Marc Zucconi [Sun, 13 Apr 1997 01:16:58 +0000 (01:16 +0000)]
Upgrade to the 1997/02/26 version.

27 years agoUpgrade to the 1997/02/19 version.
Jean-Marc Zucconi [Sun, 13 Apr 1997 01:13:52 +0000 (01:13 +0000)]
Upgrade to the 1997/02/19 version.

27 years agoAllow forcing of local_ip:remote_ip in pap-secrets
Brian Somers [Sun, 13 Apr 1997 01:07:00 +0000 (01:07 +0000)]
Allow forcing of local_ip:remote_ip in pap-secrets
and chap-secrets files.  This allows specific users
to have pre-allocated IP numbers while others get
assigned dynamically.

Submitted by: David Nugent <davidn@unique.usn.blaze.net.au>

27 years agoMake -background option attempt each phone number at most
Brian Somers [Sun, 13 Apr 1997 00:54:45 +0000 (00:54 +0000)]
Make -background option attempt each phone number at most
once.  Make -background ignore redial.  Output exit value
to syslog with "PPP Terminated" message.

27 years agoOnly wait for the redial timeout when the last phone number in the
Brian Somers [Sat, 12 Apr 1997 22:58:41 +0000 (22:58 +0000)]
Only wait for the redial timeout when the last phone number in the
list has been dialed.  Alternate number dialing has no "pause".

Suggested by: joerg

Document this behaviour.  Document that the number of dial attempts
applies to the number of phone calls rather than the number of times
each number is dialed.  Add a missing .El.  Give a decent description
of how to connect to an ISP.

27 years agoSync this up with RELENG_2_2. Du-oh, i've totally spammed rev 1.278,
Joerg Wunsch [Sat, 12 Apr 1997 21:55:53 +0000 (21:55 +0000)]
Sync this up with RELENG_2_2.  Du-oh, i've totally spammed rev 1.278,
and quite a number of smaller fixes from 2.2 were missing as well.

27 years agoMinor fixups. Mention that DDB now has a history, and understands
Joerg Wunsch [Sat, 12 Apr 1997 21:54:48 +0000 (21:54 +0000)]
Minor fixups.  Mention that DDB now has a history, and understands
emacs-style editing (which it used to have for a long time already).
Also mention the `gdb' and `help' commands.  Other commands need an
overhaul, too (like the various `show' subcommands), but i don't feel
very competent for these.

27 years agoEveryone's favorite, i think: make DDB understand the arrow keys for the
Joerg Wunsch [Sat, 12 Apr 1997 17:35:02 +0000 (17:35 +0000)]
Everyone's favorite, i think: make DDB understand the arrow keys for the
basic cursor movements.  Assumes ANSI/DEC tty, but you can still resort
to plain emacs ^p/^n etc anyway.

27 years agoFold hardware contributors section into the donor's gallery - no sense
Jordan K. Hubbard [Sat, 12 Apr 1997 16:51:20 +0000 (16:51 +0000)]
Fold hardware contributors section into the donor's gallery - no sense
in having two places for this.

27 years agoOops. Broke golden rule #1: test everything before commit. :-]
Peter Wemm [Sat, 12 Apr 1997 14:34:02 +0000 (14:34 +0000)]
Oops.  Broke golden rule #1: test everything before commit. :-]

27 years agoRespect NOTCL, don't link with libtcl for the scripting support if
Peter Wemm [Sat, 12 Apr 1997 14:28:13 +0000 (14:28 +0000)]
Respect NOTCL, don't link with libtcl for the scripting support if
we're not building it.

27 years agoDon't descend into the perl directory if it isn't there (eg: cvsup
Peter Wemm [Sat, 12 Apr 1997 07:19:09 +0000 (07:19 +0000)]
Don't descend into the perl directory if it isn't there (eg: cvsup
refuse file), or if NOPERL is defined (eg: in /etc/make.conf)

27 years agoMake tclsh conditional on the src/contrib/tcl directory existing, and
Peter Wemm [Sat, 12 Apr 1997 07:09:18 +0000 (07:09 +0000)]
Make tclsh conditional on the src/contrib/tcl directory existing, and
allow it to be disabled via NOTCL in /etc/make.conf (similar to other
stuff in src/Makefile, eg: NOGAMES)

27 years agoPut on my flame proof suit and make libtcl build conditional on
Peter Wemm [Sat, 12 Apr 1997 07:06:08 +0000 (07:06 +0000)]
Put on my flame proof suit and make libtcl build conditional on
the src/contrib/tcl directory existing, and also have an /etc/make.conf
override (NOTCL) to stop building libtcl.  This is in similar other
things from src/Makefile, eg: NOGAMES, NOLKM, etc.

This is so that people can put in a refuse entry in their cvsup files and
not fetch the tcl code, and have it not built automatically.  I'll do
something similar for perl.

27 years agoSupport input and output filters with remote printing.
Brian Somers [Sat, 12 Apr 1997 04:23:13 +0000 (04:23 +0000)]
Support input and output filters with remote printing.
Output filters are executed on a per-file basis as it's
necessary to supply the file size to the "other side".

27 years agoEffectively remove the previous commit to fix threads forking. The
John Dyson [Sat, 12 Apr 1997 04:07:50 +0000 (04:07 +0000)]
Effectively remove the previous commit to fix threads forking.  The
change was a false-start, and needs more work.

27 years agoSubmitted by: Philippe Charnier <charnier@xp11.frmug.org>
David Nugent [Sat, 12 Apr 1997 01:04:50 +0000 (01:04 +0000)]
Submitted by: Philippe Charnier <charnier@xp11.frmug.org>
Fixes name size limitation (was hardcoded to 8).
Closes PR#3258

27 years agoAllow a kernel-supported process thread to do an exec without blasting
John Dyson [Fri, 11 Apr 1997 23:37:23 +0000 (23:37 +0000)]
Allow a kernel-supported process thread to do an exec without blasting
away the VM space of all of the other, associated threads.

27 years agoSet wakeup time in pthread_cond_wait() to `forever' (-1) to prevent
John Birrell [Fri, 11 Apr 1997 22:38:22 +0000 (22:38 +0000)]
Set wakeup time in pthread_cond_wait() to `forever' (-1) to prevent
calling thread from being rescheduled based on an unspecified wakeup
time.

Bug/fix pointed out by Alexandre Fenyo <fenyo@email.enst.fr>.

27 years agoAdd -3 option to usage() statement.
David E. O'Brien [Fri, 11 Apr 1997 19:54:09 +0000 (19:54 +0000)]
Add -3 option to usage() statement.

27 years agoFixed missing #include in synopsis.
Bruce Evans [Fri, 11 Apr 1997 18:57:26 +0000 (18:57 +0000)]
Fixed missing #include in synopsis.

27 years agoFixed return type of __syscall() in synopsis.
Bruce Evans [Fri, 11 Apr 1997 18:54:05 +0000 (18:54 +0000)]
Fixed return type of __syscall() in synopsis.

Don't quote "..." in synopsis.  Quoting here is unnecessary and unusual.

27 years agoFixed missing const(s) or #include(s) in synopsis.
Bruce Evans [Fri, 11 Apr 1997 18:47:10 +0000 (18:47 +0000)]
Fixed missing const(s) or #include(s) in synopsis.

27 years agoFixed synopsis (the #include was bogus and the return type for brk() was
Bruce Evans [Fri, 11 Apr 1997 18:39:44 +0000 (18:39 +0000)]
Fixed synopsis (the #include was bogus and the return type for brk() was
wrong).

27 years agoFixed missing header in synopsis.
Bruce Evans [Fri, 11 Apr 1997 18:09:29 +0000 (18:09 +0000)]
Fixed missing header in synopsis.

27 years agoFixed synopsis (prototype was missing a const).
Bruce Evans [Fri, 11 Apr 1997 18:07:45 +0000 (18:07 +0000)]
Fixed synopsis (prototype was missing a const).

27 years agoFixed #include(s) in the synopsis to actually declare the function(s) in the
Bruce Evans [Fri, 11 Apr 1997 17:58:16 +0000 (17:58 +0000)]
Fixed #include(s) in the synopsis to actually declare the function(s) in the
synopsis.

27 years agoCleaned up synopsis:
Bruce Evans [Fri, 11 Apr 1997 17:55:19 +0000 (17:55 +0000)]
Cleaned up synopsis:
- don't quote function names, since quoting them is unnecessary and
  unusual and confuses my synopsis checker.
- include <sys/types.h> instead of <sys/param.h>.  It is normal to
  (have to) include <sys/param.h> instead of <sys/types.h>, but it
  is more useful for man pages to document minimal prerequisites.

27 years agoFixed synopsis:
Bruce Evans [Fri, 11 Apr 1997 17:49:56 +0000 (17:49 +0000)]
Fixed synopsis:
- don't declare nonexistent function sleep().
- don't include <sys/errno.h> explicitly.

sleep() should be nuked some more, e.g., this man page should not be
named after a nonexistent function.

27 years agoDon't print nonexistent library names for `ld -f [-Lfoo] -lbar'. A
Bruce Evans [Fri, 11 Apr 1997 17:08:56 +0000 (17:08 +0000)]
Don't print nonexistent library names for `ld -f [-Lfoo] -lbar'.  A
dependency on `bar' is very unlikely to be correct.

This is a quick fix for broken dependencies in gdb and many other
places.  The dependencies on internal libraries are now missing
instead of wrong when `make depend' is run before the libraries
are created.

27 years agoDeleted ddb_inb and ddb_outb functions that provide I/O access
KATO Takenori [Fri, 11 Apr 1997 12:29:51 +0000 (12:29 +0000)]
Deleted ddb_inb and ddb_outb functions that provide I/O access
routines as function for DDB.  The inb and outb are provided as
functions in machdep.c when DDB is defined.

27 years agoAdd ``%{'' and ``%}'' to one of the examples so that you can actually
David E. O'Brien [Fri, 11 Apr 1997 10:33:12 +0000 (10:33 +0000)]
Add ``%{'' and ``%}'' to one of the examples so that you can actually
cut-n-paste the example and have it actually lex and compile.

27 years agoUlf has now also donated the case for "the new freefall" - amend his
Jordan K. Hubbard [Fri, 11 Apr 1997 10:04:46 +0000 (10:04 +0000)]
Ulf has now also donated the case for "the new freefall" - amend his
entry accordingly.

27 years agoSynchronize with sys/i386/isa/syscons.c revision 1.209.
KATO Takenori [Fri, 11 Apr 1997 07:41:22 +0000 (07:41 +0000)]
Synchronize with sys/i386/isa/syscons.c revision 1.209.

27 years agoMake depend rely on config.h
Jordan K. Hubbard [Fri, 11 Apr 1997 05:17:25 +0000 (05:17 +0000)]
Make depend rely on config.h

27 years agoAdd Joe McGuckin for the Kingston ethernet card he donated.
Jordan K. Hubbard [Fri, 11 Apr 1997 04:32:44 +0000 (04:32 +0000)]
Add Joe McGuckin for the Kingston ethernet card he donated.

27 years agoDon't even think about processing bogus domain names here.
Bill Paul [Thu, 10 Apr 1997 20:26:04 +0000 (20:26 +0000)]
Don't even think about processing bogus domain names here.

27 years agoDrop the number of allowed tags back down to 8. Pluto uses a higher value
Justin T. Gibbs [Thu, 10 Apr 1997 19:14:58 +0000 (19:14 +0000)]
Drop the number of allowed tags back down to 8.  Pluto uses a higher value
which mistakenly got committed.

Fix two bugs in the ahc_reset_device code:
Limit search for SCBs to process to those that are active and
are not queued for done processing.

It's okay for an SCB to not have a waiting next SCB.

27 years agoaic7xxx.seq:
Justin T. Gibbs [Thu, 10 Apr 1997 19:13:07 +0000 (19:13 +0000)]
aic7xxx.seq:
Be consistant about testing for parity errors after waiting for a
REQ on the bus.

Don't ack the last byte in a transaction until after we've cleared
all target state.

aic7xxx_asm.c:
Test the return value of getopt against -1 not EOF. (Yet another
shameless victum of the style guide being wrong).

27 years agoFixed bogus existence test related to searching for a nearby obj
Bruce Evans [Thu, 10 Apr 1997 16:33:32 +0000 (16:33 +0000)]
Fixed bogus existence test related to searching for a nearby obj
directory.  The object (= current) directory always exists, so
don't search for it.

Use ${.TARGET} instead of deprecated $@.

27 years agoFixed bogus existence test related to searching for a nearby obj
Bruce Evans [Thu, 10 Apr 1997 16:14:09 +0000 (16:14 +0000)]
Fixed bogus existence test related to searching for a nearby obj
directory.  manpath.1 is always in the current (= object) directory,
so don't search for it.

27 years agoFixed bogus existence test related to searching for a nearby obj
Bruce Evans [Thu, 10 Apr 1997 16:13:17 +0000 (16:13 +0000)]
Fixed bogus existence test related to searching for a nearby obj
directory.  man.1 is always in the current (= object) directory,
so don't search for it.

27 years agoFixed bogus existence test related to searching for a nearby obj
Bruce Evans [Thu, 10 Apr 1997 16:04:43 +0000 (16:04 +0000)]
Fixed bogus existence test related to searching for a nearby obj
directory.  config.h is always in the current (= object) directory,
so don't search for it.

config.h is not a source for the library, so don't put it in SRCS and
don't make the library depend on it.

Don't put unused flags in CFLAGS.

Simplify using INTERNALLIB*.

27 years agoFixed bogus existence check related to searching for a nearby obj directory.
Bruce Evans [Thu, 10 Apr 1997 15:15:43 +0000 (15:15 +0000)]
Fixed bogus existence check related to searching for a nearby obj directory.
Now the search is as incomplete as elsewhere instead of broken.

27 years agoGet the declaration of `struct dirent' from <sys/dirent.h>, not from
Bruce Evans [Thu, 10 Apr 1997 15:05:38 +0000 (15:05 +0000)]
Get the declaration of `struct dirent' from <sys/dirent.h>, not from
<sys/dir.h>, and use the new macro GENERIC_DIRSIZ() instead of DIRSIZ().

Removed unused #includes.

27 years agoGet the declaration of `struct dirent' from <sys/dirent.h>, not from
Bruce Evans [Thu, 10 Apr 1997 14:56:49 +0000 (14:56 +0000)]
Get the declaration of `struct dirent' from <sys/dirent.h>, not from
<sys/dir.h>.

Removed unused #include.

Fixed type and order of struct members in pseudo-declaration of `struct
vop_readdir_args'.

27 years agoRemoved unused or apparently-unused #includes, especially of the
Bruce Evans [Thu, 10 Apr 1997 14:35:32 +0000 (14:35 +0000)]
Removed unused or apparently-unused #includes, especially of the
deprecated header <sys/dir.h>.

27 years agoAvoid processing obviously bogus domain names in the YPBINDPROC_DOMAIN
Bill Paul [Thu, 10 Apr 1997 14:18:03 +0000 (14:18 +0000)]
Avoid processing obviously bogus domain names in the YPBINDPROC_DOMAIN
and YPBINDPROC_SETDOM procedures.

27 years agoAdd a sanity test to ypproc_xfr_2_svc(): check that the map and domain
Bill Paul [Thu, 10 Apr 1997 14:12:51 +0000 (14:12 +0000)]
Add a sanity test to ypproc_xfr_2_svc(): check that the map and domain
in the transfer request actually exist. Technically ypxfr can do this too,
but why waste the cycles getting ypxfr off the ground for a transfer we
already know is going to fail.

Also apply stricter access control rules; ypproc_xfr_2_svc() is in a
different class than the normal map access procedures procedures.

27 years agoDeprecated <sys/dir.h> some more - changed the comment saying that
Bruce Evans [Thu, 10 Apr 1997 13:54:19 +0000 (13:54 +0000)]
Deprecated <sys/dir.h> some more - changed the comment saying that
<dirent.h> should be used instead to a warning.  If this causes too
many warnings in ports then it should be changed back after checking
some ports for related configuration errors.

Moved the definition of DIRSIZ() from <sys/dir.h> to <sys/dirent.h>
so that it can be used in the kernel without including <sys/dir.h>.
Renamed it in some cases to avoid new namespace pollution.

27 years agoAdd extra paranoia test. In theory this doesn't really gain you anything,
Bill Paul [Thu, 10 Apr 1997 13:48:59 +0000 (13:48 +0000)]
Add extra paranoia test. In theory this doesn't really gain you anything,
but the extra warning message could be useful.

27 years agoUse smalllblktosize() instead of multiplying small block numbers
Bruce Evans [Thu, 10 Apr 1997 13:17:09 +0000 (13:17 +0000)]
Use smalllblktosize() instead of multiplying small block numbers
by fs->fs_bsize.  The macro is usually faster and makes it clearer
that the multiplication can't overflow.

27 years agoPreserve some bits in the keyboard controller command byte when
Kazutaka YOKOTA [Thu, 10 Apr 1997 12:26:50 +0000 (12:26 +0000)]
Preserve some bits in the keyboard controller command byte when
resetting the keyboard.

Well, sorry, this bug is totally my fault. I DID intend to preserve
them, but somehow I failed.

The bug puts some old keyboard controllers in a strange state,
resulting in keyboard freeze or random key input.

The fix closes PR kern/3067.

27 years agoAdd a short blurb about the ports page in the "applications" question.
Satoshi Asami [Thu, 10 Apr 1997 12:06:53 +0000 (12:06 +0000)]
Add a short blurb about the ports page in the "applications" question.
List all three packages directories in the ensuing packages description.

Add a short (in fact, very short) section about libc.so.3.0 and wrong
package versions.

27 years agoAdd a section on sysmouse configuration.
Brian Somers [Thu, 10 Apr 1997 10:20:46 +0000 (10:20 +0000)]
Add a section on sysmouse configuration.
Suggested by: jkh

27 years agoMake the use of mousedtype/sysmouse clearer.
Brian Somers [Thu, 10 Apr 1997 10:20:22 +0000 (10:20 +0000)]
Make the use of mousedtype/sysmouse clearer.
Suggested by: jkh

27 years agoCatch up with the original.
Hiroyuki Hanai [Thu, 10 Apr 1997 06:06:22 +0000 (06:06 +0000)]
Catch up with the original.

27 years agoNate Johnson is now the Gallery/Commercial editor since Charles Wimmer
Jordan K. Hubbard [Thu, 10 Apr 1997 05:23:25 +0000 (05:23 +0000)]
Nate Johnson is now the Gallery/Commercial editor since Charles Wimmer
seems to have disappeared.

27 years agoAdd Jack O'Neill to donor's list.
Jordan K. Hubbard [Thu, 10 Apr 1997 05:20:19 +0000 (05:20 +0000)]
Add Jack O'Neill to donor's list.

27 years agoCatch up with the original.
Hiroyuki Hanai [Thu, 10 Apr 1997 04:41:55 +0000 (04:41 +0000)]
Catch up with the original.

27 years agoExpand donors list. Note: Anyone I missed, *please* contact me!
Jordan K. Hubbard [Thu, 10 Apr 1997 02:15:59 +0000 (02:15 +0000)]
Expand donors list.  Note:  Anyone I missed, *please* contact me!
I'm only now starting to get this a little more well organized.

27 years agoTypo police. Part of PR# 3242.
Mike Pritchard [Wed, 9 Apr 1997 23:05:04 +0000 (23:05 +0000)]
Typo police.  Part of PR# 3242.

Submitted by: Philippe Charnier <charnier@xp11.frmug.org>

27 years agoUse "-pg" instead of "-p" for generating profiled object files.
John Polstra [Wed, 9 Apr 1997 20:31:15 +0000 (20:31 +0000)]
Use "-pg" instead of "-p" for generating profiled object files.
The two options differ only when linking, but "-p" is incorrect
because it is associated with the old-style "prof" command, which we
do not support.

27 years agoOops...now really commit the doc change.
Guido van Rooij [Wed, 9 Apr 1997 20:25:43 +0000 (20:25 +0000)]
Oops...now really commit the doc change.

27 years agoMake a sysconfig variable controlling if teh kernel should accept
Guido van Rooij [Wed, 9 Apr 1997 20:17:15 +0000 (20:17 +0000)]
Make a sysconfig variable controlling if teh kernel should accept
nfs requests from non-privileged ports.

Change mountd such that it does never set this variable, but only clears
it when run with -n. Also document this in the man page.

27 years agoMake "gcc -shared" work properly for building shared libraries. An
John Polstra [Wed, 9 Apr 1997 20:13:41 +0000 (20:13 +0000)]
Make "gcc -shared" work properly for building shared libraries.  An
up-to-date version of c++rt0.o is necessary for this to work right.

Closes PR gnu/3035: gcc -shared flag is broken.

27 years agoAdd Kent Vander Velden for his various contributions via problem reports.
Mike Pritchard [Wed, 9 Apr 1997 19:44:58 +0000 (19:44 +0000)]
Add Kent Vander Velden for his various contributions via problem reports.

Also updated Keith Bostic's mail address while I was in the area.

27 years agoDeclare the constructor/destructor linker sets as extern rather than
John Polstra [Wed, 9 Apr 1997 19:14:31 +0000 (19:14 +0000)]
Declare the constructor/destructor linker sets as extern rather than
common.  Add one do-nothing element to each set.  This ensures that
the linker realizes that they are linker sets rather than simple
commons, and makes it possible to link c++rt0.o into every shared
library regardless of whether it is a C++ library or not.  Without
this change, the constructors and destructors in the main program
could be executed multiple times.

This change is going to make it possible to get rid of the
CPLUSPLUSLIB makefile variable once and for all.  It is a piece of
the solution to PR gnu/3505 (gcc -shared).  Finally, it fixes a
heretofore unreported bug:  If CPLUSPLUSLIB was set in a makefile
for a C++ shared library that had no static constructors or
destructors in it, then the main program's constructors and
destructors would be executed multiple times.

27 years agoCheck for -1 instead of EOF in the getopt() example.
Mike Pritchard [Wed, 9 Apr 1997 19:03:14 +0000 (19:03 +0000)]
Check for -1 instead of EOF in the getopt() example.

Submitted by: Kent Vander Velden

27 years agoProcess "lib/csu/i386" at the beginning of the "libraries" target,
John Polstra [Wed, 9 Apr 1997 18:59:36 +0000 (18:59 +0000)]
Process "lib/csu/i386" at the beginning of the "libraries" target,
instead of in the middle of it.  Soon, "c++rt0.o" will be linked
into all shared libraries.  An up-to-date version of it must be in
place before any shared libraries are built.

This is one piece of the solution for PR gnu/3035 (gcc -shared).

27 years agoIgnore SIGPIPE or any user can quit ppp just telnetting to it
Andrey A. Chernov [Wed, 9 Apr 1997 17:35:54 +0000 (17:35 +0000)]
Ignore SIGPIPE or any user can quit ppp just telnetting to it
and quitting telnet immediately (while phone number dialed)
Log client connection/disconnection with PHASE_BIT now.
Add more error recovery on client disconnection

27 years agoRemoved support for OLD_PIPE. <sys/stat.h> is now missing the hack that
Bruce Evans [Wed, 9 Apr 1997 16:53:45 +0000 (16:53 +0000)]
Removed support for OLD_PIPE.  <sys/stat.h> is now missing the hack that
supported nameless pipes being indistinguishable from fifos.  We're not
going back.

27 years agoDistinguish fifos from sockets in the S_IS* macros. This is possible now
Bruce Evans [Wed, 9 Apr 1997 16:32:23 +0000 (16:32 +0000)]
Distinguish fifos from sockets in the S_IS* macros.  This is possible now
that nameless pipes are not implemented as sockets.

Don't include <sys/time.h> if KERNEL is defined.  It should already have
been included by including <sys/param.h>.  Fixed a nearby typo.

27 years agoUse `ld -f' (actually `cc -Wl,-f') at `make depend' time to generate
Bruce Evans [Wed, 9 Apr 1997 16:10:27 +0000 (16:10 +0000)]
Use `ld -f' (actually `cc -Wl,-f') at `make depend' time to generate
almost perfect dependencies on crt0's and libraries.  DPADD and
bsd.libnames.mk should go away soon.  Use a new _EXTRADEPEND target
to implement this and to avoid editing of .depend when .depend isn;t
being rebuilt.  The afterdepend target doesn't seem to be good for
anything and is now unused.

Fixed LDDESTDIR for the DESTDIR case when ${SHLIBDIR} != /usr/lib.

Added commented-out -nostdlib to LDDESTDIR for the DESTDIR case.
The wrong libraries may be used without this; however it breaks
linkage to crt0 and libc.