looking at a high resolution clock for each of the following events:
function call, function return, interrupt entry, interrupt exit,
and interesting branches. The differences between the times of
these events are added at appropriate places in a ordinary histogram
(as if very fast statistical profiling sampled the pc at those
places) so that ordinary gprof can be used to analyze the times.
gmon.h:
Histogram counters need to be 4 bytes for microsecond resolutions.
They will need to be larger for the 586 clock.
The comments were vax-centric and wrong even on vaxes. Does anyone
disagree?
gprof4.c:
The standard gprof should support counters of all integral sizes
and the size of the counter should be in the gmon header. This
hack will do until then. (Use gprof4 -u to examine the results
of non-statistical profiling.)
config/*:
Non-statistical profiling is configured with `config -pp'.
`config -p' still gives ordinary profiling.
kgmon/*:
Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b'
still enables ordinary profiling (and distables non-statistical
profiling) if non-statistical profiling is configured.
looking at a high resolution clock for each of the following events:
function call, function return, interrupt entry, interrupt exit,
and interesting branches. The differences between the times of
these events are added at appropriate places in a ordinary histogram
(as if very fast statistical profiling sampled the pc at those
places) so that ordinary gprof can be used to analyze the times.
gmon.h:
Histogram counters need to be 4 bytes for microsecond resolutions.
They will need to be larger for the 586 clock.
The comments were vax-centric and wrong even on vaxes. Does anyone
disagree?
gprof4.c:
The standard gprof should support counters of all integral sizes
and the size of the counter should be in the gmon header. This
hack will do until then. (Use gprof4 -u to examine the results
of non-statistical profiling.)
config/*:
Non-statistical profiling is configured with `config -pp'.
`config -p' still gives ordinary profiling.
kgmon/*:
Non-statistical profiling is enabled with `kgmon -B'. `kgmon -b'
still enables ordinary profiling (and distables non-statistical
profiling) if non-statistical profiling is configured.
Changing a local passwd will now keep the encryption type that
was originally used to encrypt the password, so folks adding DES
to their systems will not be irritated/confused by having MD5'ed
passwords in their master.passwd. Coming later is an option to
allow the user to choose the encryption type.
2) Fix a bunch of compiler warnings announced by turning on -Wall.
I did not get them all, that will come a bit later.
Update rpcgen with the one from the TI-RPC 2.3 distribution.
Note that when built for FreeBSD, this version of rpcgen assumes
backwards compatibility mode by default. This means that it will produce
ONCRPC 4.0 compatible code unless otherwise instructed, instead of the
other way around.
One incompatibility has also been worked around: this rpcgen normally
always emits an '#include <stropts.h>' directive whether you select
backwards compatibility mode or not. We don't have STREAMS, so this
behavior has been changed: now it will only emit this line if run in TI-RPC
mode.
The 'generate output files in current directory instead of the
directory where the protocol definition file lives' hack from the
original rpcgen has been preserved.
Notable new features:
- Can be used to generate RPC servers that can be launched
from port monitors such as inetd(5).
- Can generate ANSI C code.
- Can generate sample client and server top-level programs and
makefiles in addition to the usual client and server stubs.
- Can generate inline XDR routines.
The #ifdef NEWSALT code doesn't NULL terminate the salt string..
We dont appear to use this code anymore, but it shouldn't hurt
Submitted by: Laurence Lopez <lopez@mv.mv.com>
quite right. (Thic causes you to get prompted for an 'Old Password' when
changing someone's NIS password even if your password isn't set yet.)
Do it like local_passwd.c does.
simplest thing is to just calculate the days using curtime - boottime / 86400.
The modification for this is less obtrusive anyway.
Suggested by: Bill Fenner <fenner@parc.xerox.com>
that if you do an rup on a machine that's been running longer than a year,
you get the wrong day count. Now we factor in 365 * (curtime.tm_year -
boottime.tm_year) to get the correct value. (I noticed this while running
rup on a SunOS machine I have that's been up 525 days. My FreeBSD
machines all said it had only been up for 160 (525-365) days. :)
- apply chmod to the targets, not to the sources.
- apply chown to the targets.
It is still bogus to install by building in the target directory. See
mklocale/data/Makefile for a better method.
27c27
< 11/29 Thanksgiving Day (Last Thursday in November)
---
> 11/23 Thanksgiving Day (4th Thursday in November)
it's not that the date was wrong for this year (it was the wrong year..
it was that the ALGORYTHM was wrong..
very confusing for non americans wondering why americans were going to be
on holiday on the 23rd..
style of error reporting (i prefer gcc style to be consistent with the
compiler) is left, plus a minor nit he's most likely been overlooking.
There are still problems with bootstrapping, and you should expect
troubles when linting libc...
This is just a vendor import by now. I'll wait until i'll get the
imported files back via CTM before applying the FreeBSD patches.
Don't use it yet.
Submitted by: Jochen Pohl <jpo.drs@sni.de>
Obtained from: (NetBSD -- this version is directly from Jochen)
broke. It's much easier to debug the symbol export lists in lkm makefiles
if you know what your errors are during the build process. :-)
Bleah.. symorder.c is *horrible*. :-(
to the description in the manpage. g flag means "replace every occurence
in each word", and its absence means "replace first occurence in each word".
Previously, absence of the g flag was implemented to mean "replace first
occurence found in all words, and then stop replacing", which was incorrect.
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.
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) :)
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
strings describing the drive status and error bits are so deficient
that the pointer is always NULL.
Reported by: Philippe Charnier <charnier@lirmm.fr>
made other performance improving changes. This improves the performance
of last(1) by as much as 32 times in some cases, and in more typical cases
is about twice as fast.
Added a BUGS section to the manual page to describe the behavior of last(1)
when a login shell terminates abnormally (and thus doesn't write a logout
record to the wtmp file).
finger output from 80 to 79 characters to stop the syscons
every-second-line-is-blank problem.
Also, redo the TTY column mod so that it steals one of the (usually) two
blank spaces from the before the tty column rather than from the office
phone number.
This means the office phone field width in the short finger is back to 15
characters instead of 13.
time of the target if the target file is the same as the source),
-d (debug), and -p (same as -C except for preserving the modification
time of the source if the target doesn't exists or is different from
the source.
Use library err() functions instead of our own and pass them better
exit codes.
Submitted by: wollman (and changed a lot by me)
Our ld is derived from gnu ld which doesn't check the timestamp on
__.SYMDEF. gnu ld is designed to work with gnu ar which doesn't
even have a ranlib option (gnu ar updates __.SYMDEF automatically
if __.SYMDEF already exists, so gnu ld expects __.SYMDEF to be up
to date if it exists).
uncompression by the "proper" kernel. These changes also add a -v option
so you can see how much room you are using, and check to make sure you're
not going past the 4MB boundary.
This depends on the corresponding changes to sys/i386/boot/kzipboot.
Submitted by: Gary Jones(?) <gj@freefall>, and my code merged in.
Kerberos obtains a network address for the local host from the routing
tables and uses it consistently for all Kerberos transactions. This ensures
that packets only leave the *authenticated* interface. Clients who open
and use their own sockets for encrypted or authenticated correspondance
to kerberos services should bind their sockets to the same address as that
used by kerberos. krb_get_local_addr() and krb_bind_local_addr() allow
clients to obtain the local address or bind a socket to the local address
used by Kerberos respectively.
Reviewed by: Mark Murray <markm>, Garrett Wollman <wollman>
Obtained from: concept by Dieter Dworkin Muller <dworkin@village.org>
handling for the tools binaries. E.g., after libc.a is changed, it
previously took two `make' passes and one `make depend' pass following
one of the `make' passes to bring everything up to date. Now one `make'
pass followed by one `make depend' pass is sufficient.
Dependency handling seems to be difficult to handle cleanly when
interdependent things are built in different directories.
The filenames in SRCS must have one of the extensions .s, .S, .c, or .cc
if they are to be handled by bsd.dep.mk. Lex and yacc files must be
converted to C files and kept around for everything to work. This is
handled fairly automatically if the names of the generated C files are
put in SRCS. Unfortunately these names must be put in CLEANFILES too.
pcvt Makefiles:
Fix DPADD. It was missing.
Fix CLEANFILES. Some temporary files were missing.
Fix CFLAGS. There were some `-I dir' options.
There must be no whitespace separating -I and -D options from the
corresponding args if these options are to be handled by bsd.dep.mk.
the way it stores and handles "interface". The previous behavior resulted
in strange output from 'w' and 'ps' when an interface specification was
given to netstat.
[ Find to a file vs. to stdout ] produces different output because find
does not flush stdout when doing a -print.
Submitted by: Jeffrey Hsu <hsu@freefall.freebsd.org>
