]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
28 years agoThis commit was generated by cvs2svn to compensate for changes in r11884,
peter [Sat, 28 Oct 1995 18:51:33 +0000 (18:51 +0000)]
This commit was generated by cvs2svn to compensate for changes in r11884,
which included commits to RCS files with non-trunk default branches.

28 years agoImport GNU diffutils 2.7
peter [Sat, 28 Oct 1995 18:51:33 +0000 (18:51 +0000)]
Import GNU diffutils 2.7

Note, this is going to be messy.. 2.3 was vendor-branch imported, while
2.6 was done as a delta. Sigh.  I'm importing this on a vendor branch so
that it will be easier to deal with next time..

(cvs-1.6 wants rcs-5.7, and rcs-5.7 suggests diffutils-2.7)

28 years agoAdd extra export symbols for the ATAPI generic interface.
peter [Sat, 28 Oct 1995 17:46:58 +0000 (17:46 +0000)]
Add extra export symbols for the ATAPI generic interface.
And a couple of minor fiddles in the process..

28 years agoReattach atapi and wcd to the lkm Makefile, these compile for me now.
peter [Sat, 28 Oct 1995 17:46:03 +0000 (17:46 +0000)]
Reattach atapi and wcd to the lkm Makefile, these compile for me now.

28 years agoRepair the wcd and atapi code a little, so that it compiles and loads
peter [Sat, 28 Oct 1995 17:45:04 +0000 (17:45 +0000)]
Repair the wcd and atapi code a little, so that it compiles and loads
as an LKM.
I dont have one of these beasts, so I cannot guarantee that I've not
broken it (more).
It does compile with ATAPI_STATIC, and as a LKM.

28 years agoFix incorrect bit definitions for SXFRCTL0 (typo). The affected bits
gibbs [Sat, 28 Oct 1995 17:27:21 +0000 (17:27 +0000)]
Fix incorrect bit definitions for SXFRCTL0 (typo).  The affected bits
aren't referenced by the driver yet, so this error shouldn't have caused
any problems.

Submitted by: Dan Eischen <deischen@iworks.InterWorks.org>

28 years agoAdd a few more heuristics to modload:
peter [Sat, 28 Oct 1995 17:06:05 +0000 (17:06 +0000)]
Add a few more heuristics to modload:
1: generate the outfile in /tmp if it's not specified explicitly.
2: if the outfile was implicitly placed in /tmp, automatically remove it.

This means that you can type: modload /lkm/ipfw_mod.o and it'll work, it
wont try and write to /lkm, and it wont leave the (normally) useless
symbol file.

This should not interfere with things like ibcs2 and atapi, which
export some symbols from one LKM to the other by leaving the symbol file.

28 years agoThis commit was generated by cvs2svn to compensate for changes in r11876,
markm [Sat, 28 Oct 1995 17:04:25 +0000 (17:04 +0000)]
This commit was generated by cvs2svn to compensate for changes in r11876,
which included commits to RCS files with non-trunk default branches.

28 years agoControl utility for the /dev/random device. This allows root to select
markm [Sat, 28 Oct 1995 17:04:25 +0000 (17:04 +0000)]
Control utility for the /dev/random device. This allows root to select
which interrupts are to be used in randomess gathering. Man page for
the random device is included.

28 years agoTheodore Ts'po's random number gernerator for Linux, ported by me.
markm [Sat, 28 Oct 1995 16:58:05 +0000 (16:58 +0000)]
Theodore Ts'po's random number gernerator for Linux, ported by me.
This code will only be included in your kernel if you have
'options DEVRANDOM', but that will fall away in a couple of days.
Obtained from: Theodore Ts'o, Linux

28 years agoAdd #include <sys/user.h>.
phk [Sat, 28 Oct 1995 16:14:18 +0000 (16:14 +0000)]
Add #include <sys/user.h>.

28 years agoClose out PR#723: time -l dumps core
pst [Sat, 28 Oct 1995 16:05:32 +0000 (16:05 +0000)]
Close out PR#723: time -l dumps core

28 years agoRemove unused functions and variables, make things static, and other cleanups.
phk [Sat, 28 Oct 1995 15:39:31 +0000 (15:39 +0000)]
Remove unused functions and variables, make things static, and other cleanups.

28 years agoAdd setlocale LC_CTYPE
ache [Sat, 28 Oct 1995 14:30:50 +0000 (14:30 +0000)]
Add setlocale LC_CTYPE

28 years agoAdd ascii links
ache [Sat, 28 Oct 1995 13:39:27 +0000 (13:39 +0000)]
Add ascii links

28 years agoAdd ascii locale dir
ache [Sat, 28 Oct 1995 13:35:32 +0000 (13:35 +0000)]
Add ascii locale dir

28 years agoAdd strict 7bit ASCII locale
ache [Sat, 28 Oct 1995 13:32:30 +0000 (13:32 +0000)]
Add strict 7bit ASCII locale

28 years agoSorry.. I didn't mean to re-enable wcd yet
peter [Sat, 28 Oct 1995 13:28:16 +0000 (13:28 +0000)]
Sorry.. I didn't mean to re-enable wcd yet

28 years agoAdd entries for labpc.4, nca.4, scd.4, and sea.4 to the
dufault [Sat, 28 Oct 1995 13:27:51 +0000 (13:27 +0000)]
Add entries for labpc.4, nca.4, scd.4, and sea.4 to the
Makefile so that they are installed.  The man pages are already
present.

28 years agoSorry, the last commit screwed up for me, this is the right one (I hope!)
phk [Sat, 28 Oct 1995 13:07:28 +0000 (13:07 +0000)]
Sorry, the last commit screwed up for me, this is the right one (I hope!)
Please refer to the previous commit message about sysctl variables.

28 years agoMake modload calculate a reasonable default entry point name, in light
peter [Sat, 28 Oct 1995 13:06:11 +0000 (13:06 +0000)]
Make modload calculate a reasonable default entry point name, in light
of the recent changes to bsd.kmod.mk, which enforces that the entry
point be <modname>_mod, unless overridden.

This means that when modloading "/lkm/foo_mod.o", it'll use "foo_mod"
as the default entry point (instead of xxxinit) unless you specify
another via the -e option.

28 years agoRewamp the way we make sysctl variables to be easier to cope with.
phk [Sat, 28 Oct 1995 12:59:25 +0000 (12:59 +0000)]
Rewamp the way we make sysctl variables to be easier to cope with.

The goal is to make them "user-friendly" :-)

In the end this will allow a SNMP style "getnext" function, sysctl editing
in the boot-editor and/or debugger, LKMs can define sysctl vars when
they get loaded, and remove them when unloaded and other interesting
uses for dynamic sysctl variables.

28 years agoSimplify FILES section in the same manner like groff manpage
ache [Sat, 28 Oct 1995 12:56:05 +0000 (12:56 +0000)]
Simplify FILES section in the same manner like groff manpage
does, add more koi8-r references

28 years agoAdd reference to koi8-r device, fix formatting a little
ache [Sat, 28 Oct 1995 12:49:38 +0000 (12:49 +0000)]
Add reference to koi8-r device, fix formatting a little

28 years agoPart 2 of the linux LKM hook, and use the new screen saver entry
peter [Sat, 28 Oct 1995 12:44:13 +0000 (12:44 +0000)]
Part 2 of the linux LKM hook, and use the new screen saver entry
points in the modload, as enforced by the recent symorder and
bsd.kmod.mk changes..

28 years agoAdd reference to koi8-r device
ache [Sat, 28 Oct 1995 12:41:56 +0000 (12:41 +0000)]
Add reference to koi8-r device

28 years agoIn a moment of weakness, I added the switch for loading the linux LKM,
peter [Sat, 28 Oct 1995 12:41:40 +0000 (12:41 +0000)]
In a moment of weakness, I added the switch for loading the linux LKM,
in the same style as the ibcs2 LKM..

28 years agoDo a pass over the broken LKM's and update them to use the "new"
peter [Sat, 28 Oct 1995 12:35:14 +0000 (12:35 +0000)]
Do a pass over the broken LKM's and update them to use the "new"
convention of having their entry point named "<modname>_mod"".
Symorder is enforcing this when the current bsd.kmod.mk is installed.

I've not tested all these, but at least they all compile now.

Reattach them to the makefile.

Note that the change that I made to symorder needs to be compiled and
installed before any LKM's will work - the last version was corrupting
the relocation tables.  A "make world" will to this, but if you
manually run a make on the lkm's you'll need to take care of it by
hand.

28 years agosymorder appears to have been designed to run on executable files
peter [Sat, 28 Oct 1995 12:27:20 +0000 (12:27 +0000)]
symorder appears to have been designed to run on executable files
only, as it payes no attention to the relocation table (which
references the symbols).

As a result, running "symorder -c" to clean up the visibility of a LKM
".o" file (as is done in the new bsd.kmod.mk) totally screws up the
relocation table, making the LKM file unloadable. (ld: bogus
relocation record)

This is a pretty crude fix - I've changed symorder so that when
running in "cleanup" mode, it disables the reordering which was
screwing up the relocation table.  I'm sure there is a better fix, but
I didn't have the energy.  Feel free to fix this hack, probably by
renumbering the symbol indexes in the relocation table.

28 years agoFix -h option:
ache [Sat, 28 Oct 1995 11:54:42 +0000 (11:54 +0000)]
Fix -h option:
sense symlink even it is unresolved

28 years agoAt the risk of starting a commit-war, attempt to fix once and for all, the
peter [Sat, 28 Oct 1995 09:39:57 +0000 (09:39 +0000)]
At the risk of starting a commit-war, attempt to fix once and for all, the
"make distrib-dirs" target.
Neither of Andrey's two attempts have worked for me with the [ -h ..]
test both with && and ||.
I've changed it to a full
  if [ -h ...]; then \
    rm ... ; \
  fi ; \
construct.  It's much clearer what's meant to happen, and it works! :-)

28 years agoCall vfs_unbusy() before error returns from sysctl_vnode(). This fixes
bde [Sat, 28 Oct 1995 08:50:08 +0000 (08:50 +0000)]
Call vfs_unbusy() before error returns from sysctl_vnode().  This fixes
PR 795.

Set the size before one error return from sysctl_vnode() the same as before
the other.  The caller might want to know about the amount successfully
read although the current caller doesn't.

28 years agoClose docs/43 manpage for keysu is missing
pst [Sat, 28 Oct 1995 08:07:38 +0000 (08:07 +0000)]
Close docs/43       manpage for keysu is missing

28 years agoTemp. disable following modules:
ache [Sat, 28 Oct 1995 02:49:46 +0000 (02:49 +0000)]
Temp. disable following modules:
atapi coff ibcs2 ipfw linux syscons wcd
(until they will be fixed)
It allows to build/install other modules from top

28 years agoBack out bogus "fix" which was commited over right fix without any check.
ache [Sat, 28 Oct 1995 02:01:34 +0000 (02:01 +0000)]
Back out bogus "fix" which was commited over right fix without any check.

28 years agoSubmitted by: Mike Mitchell
julian [Fri, 27 Oct 1995 20:18:22 +0000 (20:18 +0000)]
Submitted by: Mike Mitchell

oops.
forgot this file then went to bed..

sorry

28 years agoSome last few tweaks - now the sample home page comes up properly by
jkh [Fri, 27 Oct 1995 17:31:03 +0000 (17:31 +0000)]
Some last few tweaks - now the sample home page comes up properly by
default.

28 years agoChange the script-driven custom distribution selection mechanism. This is
jkh [Fri, 27 Oct 1995 17:00:27 +0000 (17:00 +0000)]
Change the script-driven custom distribution selection mechanism.  This is
a bit simpler.

28 years agomisc typos
adam [Fri, 27 Oct 1995 16:56:53 +0000 (16:56 +0000)]
misc typos

28 years ago- #if defined(TF_NEEDSYN) && defined(TF_NEEDSYN)
adam [Fri, 27 Oct 1995 12:32:07 +0000 (12:32 +0000)]
- #if defined(TF_NEEDSYN) && defined(TF_NEEDSYN)
+ #if defined(TF_NEEDSYN) && defined(TF_NEEDFIN)

28 years agoCorrect ip protocol in error message, when binding the tcp port.
adam [Fri, 27 Oct 1995 11:44:07 +0000 (11:44 +0000)]
Correct ip protocol in error message, when binding the tcp port.

28 years agoSubmitted by: john hay
julian [Fri, 27 Oct 1995 10:48:31 +0000 (10:48 +0000)]
Submitted by: john hay

patches to merge the two IPX packages to work with each other and to
not break make-world :)

IPXrouted should be working now, (or at least compiling) :)

28 years agoSubmitted by: john hay
julian [Fri, 27 Oct 1995 06:51:02 +0000 (06:51 +0000)]
Submitted by: john hay
add a link in /usr/include to /sys/netipx

28 years agoRetrofit Apache server setup to use "Powered by FreeBSD" logo at the bottom
jkh [Fri, 27 Oct 1995 05:16:46 +0000 (05:16 +0000)]
Retrofit Apache server setup to use "Powered by FreeBSD" logo at the bottom
of the sample page it sets up.

28 years agoAdd the power logo to the root floppy.
jkh [Fri, 27 Oct 1995 05:03:26 +0000 (05:03 +0000)]
Add the power logo to the root floppy.

28 years agoBe more aggressive about writing out configuration files.
jkh [Fri, 27 Oct 1995 03:59:38 +0000 (03:59 +0000)]
Be more aggressive about writing out configuration files.

28 years agoLast minute nits.
jkh [Fri, 27 Oct 1995 03:07:14 +0000 (03:07 +0000)]
Last minute nits.

28 years agoClean up some last bogons with the WEB server setup (now that I'm far
jkh [Fri, 27 Oct 1995 02:12:58 +0000 (02:12 +0000)]
Clean up some last bogons with the WEB server setup (now that I'm far
enough along to worry about such things!).

28 years agoFix the bug I introduced with /etc/sysconfig not being updated
jkh [Fri, 27 Oct 1995 01:22:58 +0000 (01:22 +0000)]
Fix the bug I introduced with /etc/sysconfig not being updated
properly.
Make some cosmetic doc changes here and there in response to user
feedback.

28 years agoRecognize the aic7880 and 3940 Ultra.
gibbs [Thu, 26 Oct 1995 23:58:59 +0000 (23:58 +0000)]
Recognize the aic7880 and 3940 Ultra.

If RAMENB is set in devconfig, walk the external SCBs.  Some Intel Xpress
motherboards set this bit.

For external SCBs for the 3940.  It doesn't set RAMPS or RAMENB, but does
have the ram.

28 years agoProperly deal with the Ultra series of adapters. We should now understand
gibbs [Thu, 26 Oct 1995 23:57:18 +0000 (23:57 +0000)]
Properly deal with the Ultra series of adapters.  We should now understand
the new seeprom format and negotiate up to 20MHz sync if set in SCSI-Select.

Reduce the complexity of the timeout code by running it at splhigh().  Fix
a bug that caused rescheduled timeouts at 0 clock ticks in the future causing
an infinite loop.

Obtained from: Timeout bug noticed by David Greenman and wcarchive.

28 years agoSet SCSIGO generically before we determine the phase instead of doing it
gibbs [Thu, 26 Oct 1995 23:54:13 +0000 (23:54 +0000)]
Set SCSIGO generically before we determine the phase instead of doing it
in each phase routine.  Saves a few instructions.

Be more careful in how we deal with SXFRCTL0.  Or in the control bits of
interest instead of using mvi.  The kernel driver will set the ULTRAEN
bit of SXFRCTL0 if we are using Ultra (20MHz) mode and we don't want to
clobber it.

In sdtr_to_rate divide by two if we are in ultra mode to get the correct
setting since its a 20MHz instead of 10MHz scale.

28 years agoAdd setlocale LC_TIME
ache [Thu, 26 Oct 1995 23:15:14 +0000 (23:15 +0000)]
Add setlocale LC_TIME

28 years agoAdd setlocale LC_TIME
ache [Thu, 26 Oct 1995 23:10:10 +0000 (23:10 +0000)]
Add setlocale LC_TIME

28 years agoAdd setlocale LC_TIME
ache [Thu, 26 Oct 1995 22:58:56 +0000 (22:58 +0000)]
Add setlocale LC_TIME

28 years agoRemove unneded ctype.h
ache [Thu, 26 Oct 1995 22:32:00 +0000 (22:32 +0000)]
Remove unneded ctype.h

28 years agoAdd setlocale LC_CTYPE
ache [Thu, 26 Oct 1995 22:26:15 +0000 (22:26 +0000)]
Add setlocale LC_CTYPE

28 years agoAdd setlocale LC_CTYPE
ache [Thu, 26 Oct 1995 22:16:45 +0000 (22:16 +0000)]
Add setlocale LC_CTYPE

28 years agofix a reference in "SEE ALSO" (modunload(8) mistakenly references
torstenb [Thu, 26 Oct 1995 21:44:11 +0000 (21:44 +0000)]
fix a reference in "SEE ALSO" (modunload(8) mistakenly references
itself instead of modload(8))

28 years agoReviewed by: julian and Mike Mitchel
julian [Thu, 26 Oct 1995 21:28:30 +0000 (21:28 +0000)]
Reviewed by: julian  and Mike Mitchel
Submitted by: john Hay (John.Hay@csir.co.za)

John's IPXrouted..
this has not yet been seen to run correctly with Mike's IPX/SPX
code  (he has his own)
bringing them both in is the first step in merging the two packages
for 2.2

28 years agoReviewed by: julian and jhay@mikom.csir.co.za
julian [Thu, 26 Oct 1995 20:31:59 +0000 (20:31 +0000)]
Reviewed by: julian and jhay@mikom.csir.co.za
Submitted by: Mike Mitchell, supervisor@alb.asctmd.com

This is a bulk mport of Mike's IPX/SPX protocol stacks and all the
related gunf that goes with it..
it is not guaranteed to work 100% correctly at this time
but as we had several people trying to work on it
I figured it would be better to get it checked in so
they could all get teh same thing to work on..

Mikes been using it for a year or so
but on 2.0

more changes and stuff will be merged in from other developers now that this is in.

Mike Mitchell, Network Engineer
AMTECH Systems Corporation, Technology and Manufacturing
8600 Jefferson Street, Albuquerque, New Mexico 87113 (505) 856-8000
supervisor@alb.asctmd.com

28 years agoAdd preliminary support for netgroup.byuser and netgroup.byhosts maps.
wpaul [Thu, 26 Oct 1995 18:00:35 +0000 (18:00 +0000)]
Add preliminary support for netgroup.byuser and netgroup.byhosts maps.

(I have to make another pass through here soon; awk doesn't handle
lines broken up with '\', which can sometimes appear in netgroup maps.)

28 years agoAdd revnetgroup.
wpaul [Thu, 26 Oct 1995 16:28:33 +0000 (16:28 +0000)]
Add revnetgroup.

28 years agoThis commit was generated by cvs2svn to compensate for changes in r11814,
wpaul [Thu, 26 Oct 1995 16:25:29 +0000 (16:25 +0000)]
This commit was generated by cvs2svn to compensate for changes in r11814,
which included commits to RCS files with non-trunk default branches.

28 years agoImport the first cut of my (finally finished) revnetgroup program. This
wpaul [Thu, 26 Oct 1995 16:25:29 +0000 (16:25 +0000)]
Import the first cut of my (finally finished) revnetgroup program. This
program parses the /etc/netgroup file into netgroup.byuser and netgroup.byhost
format for NIS.

I used hash tables to store the initial netgroup data in memory and to
construct the 'reverse' netgroup output. It seems just as fast as the
SunOS revnetgroup, which is surprising considering this is my first
attempt at using hash tables in a real application. :)

Note that I canibalized a large chunk of getnetgrent.c to save myself
from having to write my own netgroup parsing functions.

28 years agoAdd setlocale LC_ALL
ache [Thu, 26 Oct 1995 11:12:39 +0000 (11:12 +0000)]
Add setlocale LC_ALL

28 years agoChange locale to LC_ALL, there is no bitmask
ache [Thu, 26 Oct 1995 11:03:08 +0000 (11:03 +0000)]
Change locale to LC_ALL, there is no bitmask

28 years agoAdd setlocale LC_ALL
ache [Thu, 26 Oct 1995 10:59:43 +0000 (10:59 +0000)]
Add setlocale LC_ALL

28 years agoChange local to LC_ALL, there is no bitmask
ache [Thu, 26 Oct 1995 10:57:52 +0000 (10:57 +0000)]
Change local to LC_ALL, there is no bitmask

28 years agoChange locale to LC_ALL, there no bitmask
ache [Thu, 26 Oct 1995 10:56:38 +0000 (10:56 +0000)]
Change locale to LC_ALL, there no bitmask

28 years agoAdd test for symlink presence before remove it
ache [Thu, 26 Oct 1995 10:05:55 +0000 (10:05 +0000)]
Add test for symlink presence before remove it

28 years agoChanged "tools" to "build-tools".
phk [Thu, 26 Oct 1995 09:24:59 +0000 (09:24 +0000)]
Changed "tools" to "build-tools".
Suggested by: Peter Wemm

28 years agoSync up for this snapshot.
jkh [Thu, 26 Oct 1995 08:56:18 +0000 (08:56 +0000)]
Sync up for this snapshot.

28 years agoImport my laptop's installation configuration file, just as a working example.
jkh [Thu, 26 Oct 1995 08:21:48 +0000 (08:21 +0000)]
Import my laptop's installation configuration file, just as a working example.

28 years agoo Fix installUpgrade to start the holographic shell as well.
jkh [Thu, 26 Oct 1995 08:11:24 +0000 (08:11 +0000)]
o  Fix installUpgrade to start the holographic shell as well.
o  Lots of documentation fixes.
o  Rename FTP active to "FTP" and explain passive mode better.
o  Make tcpip screen a bit more friendly.
o  Literally dozens of nits.

28 years agoFrom Matt Thomas:
dg [Thu, 26 Oct 1995 07:40:11 +0000 (07:40 +0000)]
From Matt Thomas:

"I screwed the initialization of the burstsize.  Right now it will default
to 0 (which can cause corruption problems on high latency PCI buses).  It
should be set to 8 longwords to avoid problems with certain PCI chipsets."

Submitted by: Matt Thomas <matt@lkg.dec.com>

28 years agoConvert manpage to -mandoc macros.
nate [Thu, 26 Oct 1995 05:36:24 +0000 (05:36 +0000)]
Convert manpage to -mandoc macros.

Submitted by: Gary Palmer <gary@palmer.demon.co.uk>

Minor cleanup by me in the English.

28 years agoConvert to proper English in the same manner as the sliplogin manpage.
nate [Thu, 26 Oct 1995 05:13:54 +0000 (05:13 +0000)]
Convert to proper English in the same manner as the sliplogin manpage.

28 years agoMinor cleanup of English once more.
nate [Thu, 26 Oct 1995 05:12:59 +0000 (05:12 +0000)]
Minor cleanup of English once more.

28 years agoRe-write new stuff in English.
nate [Wed, 25 Oct 1995 18:58:40 +0000 (18:58 +0000)]
Re-write new stuff in English.

28 years agoFixed data loss in writes to pty masters. Data was almost always lost
bde [Wed, 25 Oct 1995 18:23:58 +0000 (18:23 +0000)]
Fixed data loss in writes to pty masters.  Data was almost always lost
at the end of each write for writes of more than 1K.

Fixed handling of residual count for early returns in writes to pty masters.
It was only adjusted in 2 out of 6 cases.

Added prototypes.

28 years agoFix clist limits. They were usually one too low. E.g., for a limit of
bde [Wed, 25 Oct 1995 17:59:58 +0000 (17:59 +0000)]
Fix clist limits.  They were usually one too low.  E.g., for a limit of
TTYHOG = 1024 bytes, 10 cblocks were reserved.  This was thought to
provide 10 * CBSIZE = 1080 bytes of buffering, but if the head pointer
is at the end of a cblock, then it only provides 1 + 9 * CBSIZE = 973
bytes of buffering.  This caused serious data loss for ptys because the
flow control is deterministic and requires at least TTYHOG bytes of
buffering.  For ttys, if input flow control is used then there is
usually enough slop in the high watermark to avoid problems, and if
input flow control isn't used then a limit of 973 is not much different
from a limit of 1024.

Add prototypes.

Continue cleaning up new init stuff.

28 years agoSync with version 1(26).
jkh [Wed, 25 Oct 1995 16:47:54 +0000 (16:47 +0000)]
Sync with version 1(26).
Submitted by: Frank Durda IV <uhclem@fw.ast.com>

28 years agoStable matcd port to 0x230, as per request by Bruce and Frank.
jkh [Wed, 25 Oct 1995 16:43:01 +0000 (16:43 +0000)]
Stable matcd port to 0x230, as per request by Bruce and Frank.
Submitted by: Frank Durda IV <uhclem@fw.ast.com>

28 years agoBring in version 1(26) - considerably less memory usage (and may bring us
jkh [Wed, 25 Oct 1995 16:41:22 +0000 (16:41 +0000)]
Bring in version 1(26) - considerably less memory usage (and may bring us
down to 4MB bootability again).
Submitted by: Frank Durda IV <uhclem@fw.ast.com>

28 years agoRun-time linker speedups - Round One
nate [Wed, 25 Oct 1995 16:16:35 +0000 (16:16 +0000)]
Run-time linker speedups - Round One

Implemented symbol memorizing to reduce the number of calls to lookup(),
making relocation go faster.  While relocating a given shared object,
the dynamic linker maintains a memorizing vector that is directly
indexed by the symbol number in the relocation entry.  The first time a
given symbol is looked up, the memorizing vector is filled in with a
pointer to the symbol table entry, and a pointer to the so_map of the
shared object in which the symbol was defined.  On subsequent uses of
the same symbol, that information is retrieved directly from the
memorizing vector, without calling lookup() again.

A symbol that is referenced in a relocation entry is typically
referenced in many relocation entries, so this memorizing reduces the
number of calls to lookup() dramatically.  The overall improvement in
the speed of dynamic linking is also dramatic -- as much as a factor of
three for programs that use many shared libaries.

Submitted by: jdp@polstra.com "John Polstra"

28 years agoBring forward my changes from 2.1
jkh [Wed, 25 Oct 1995 15:38:37 +0000 (15:38 +0000)]
Bring forward my changes from 2.1

28 years agofix bad dependencies (LIBMATH -> LIBM)
torstenb [Wed, 25 Oct 1995 15:08:21 +0000 (15:08 +0000)]
fix bad dependencies (LIBMATH -> LIBM)

28 years agoFix a typo ("-It Fl" to ".It Fl")
torstenb [Wed, 25 Oct 1995 02:17:44 +0000 (02:17 +0000)]
Fix a typo ("-It Fl" to ".It Fl")
document the "-P" switch

Obtained from: the NetBSD bugs mailinglist

28 years agochange cross reference from binmail(1) to mail.local(8)
torstenb [Wed, 25 Oct 1995 02:15:40 +0000 (02:15 +0000)]
change cross reference from binmail(1) to mail.local(8)

Obtained from: the NetBSD bugs mailinglist

28 years agoBuild all of the libraries: targets libs in the same manner. Append the
nate [Tue, 24 Oct 1995 18:51:08 +0000 (18:51 +0000)]
Build all of the libraries: targets libs in the same manner.  Append the
optional ${CLEANDIR} and ${OBJDIR} targets to those missng them.

Also, check for the existance of all library targets before building
them to follow convention.

28 years agoFix a silly bug where MAXPATHLEN was subtracted from the string length rather
jkh [Tue, 24 Oct 1995 13:46:35 +0000 (13:46 +0000)]
Fix a silly bug where MAXPATHLEN was subtracted from the string length rather
than the other way around!
Submitted by: Elmar Bartel <bartel@informatik.tu-muenchen.de>

28 years agoAdded symorder to tools target.
phk [Tue, 24 Oct 1995 08:37:28 +0000 (08:37 +0000)]
Added symorder to tools target.
Suggested by: rgrimes.

28 years agoRemove LD_NOSTD_PATH unsetenv, it isn't exist anymore
ache [Tue, 24 Oct 1995 06:52:36 +0000 (06:52 +0000)]
Remove LD_NOSTD_PATH unsetenv, it isn't exist anymore

28 years agoRemove LD_NOSTD_PATH unsetenv, isn't exist anymore
ache [Tue, 24 Oct 1995 06:50:45 +0000 (06:50 +0000)]
Remove LD_NOSTD_PATH unsetenv, isn't exist anymore

28 years agoRemove LD_NOSTD_PATH implementation, it isn't works and
ache [Tue, 24 Oct 1995 06:48:16 +0000 (06:48 +0000)]
Remove LD_NOSTD_PATH implementation, it isn't works and
can cause some problems.
Suggested-by: davidg
28 years agoAdd setlocale LC_CTYPE
ache [Tue, 24 Oct 1995 06:15:05 +0000 (06:15 +0000)]
Add setlocale LC_CTYPE

28 years agoAdd setlocale LC_CTYPE
ache [Tue, 24 Oct 1995 06:02:24 +0000 (06:02 +0000)]
Add setlocale LC_CTYPE

28 years agoRemove wrong LC_TIME env. usage
ache [Tue, 24 Oct 1995 05:56:45 +0000 (05:56 +0000)]
Remove wrong LC_TIME env. usage

28 years agoAdd setlocale LC_TIME
ache [Tue, 24 Oct 1995 05:37:15 +0000 (05:37 +0000)]
Add setlocale LC_TIME