because libmd builds a test program before installation and if
you've used CLOBBER there's no crt.0 to link with. This ensures
that in a make world the csu objects will get installed before
reaching the libmd directory.
Reviewed by:
Submitted by:
Merged changes from 2.0 version (revisions 1.46-1.50) by hand.
Finished conversion to clists: removed flush of tty output buffer
in comflush() (most writes were truncated to 256 bytes) and restored
bypass of ttyinput() in siopoll().
Finished conversion to 2.0 types - more void *'s, less caddr_t's,
less casts, no Dev_t's.
Only these things are seriously broken now compared with 1.1.5:
waiting for output complete is impossible so ttywait() can deadlock;
sioclose() isn't called enough so sioopen() sometimes returns EBUSY
unnecessarily; input flow control is not implemented.
being created and config.h depended on Makefile.inc being in the wrong
directory so real dependencies were not being checked. The "depend"
target was not created either so "make depend" always found work to do.
Submitted by:
Add the 3com 3C507 card to the if_ie.c driver. The files elink.c and
elink.h are helding routines that are shared between the 3C507 and the
3C509/3C579. if_ie507.h are constant declarations unique to the 3C507.
The code is based on the NetBSD driver if_ie.c donated to NetBSD by
Rafal Boni and then modified by Charles Hannum.
Add pre-{build,extract,install,...} targets for Torsten, who apparently
needs them. Can't do effective post-* targets without major work, sorry.
Jordan
Reviewed by:
Submitted by:
Subject: man returns 1
In 1.1.5.1, man returns a status of 1 if the lookup succeeds and 0 if
it fails. Here is a patch for what I believe is a simple oversight:
Submitted by: jkh
FreeBSD system sources installs itself as the standard cc and c++. I've
fixed c++ to call cc instead of gcc and removed all the symlinks
that get created to g** version of the binaries. This means that
you can install a second version of gcc that does use the g prefix
alongside the "system" version of gcc. The only conflict is libgcc
but since we install it as libcc.so.26 and nothing else is likely
to that should be ok.
Reviewed by:
Submitted by:
Changed u_int_inb to just inb and deleted define.
The code generated is identical to that generated with the cast so
the problem was obviously fixed at some point after gcc 1.4
Reviewed by:
Submitted by:
hosing syscons. Doesn anyone know anything about this
or can we just delete it now?
/*
* This roundabout method of returning a u_char helps stop gcc-1.40 from
* generating unnecessary movzbl's.
*/
#ifdef disable_for_gcc-2_6_0
#define inb(port) ((u_char) u_int_inb(port))
#endif
static inline u_int
u_int_inb(u_int port)
{
u_char data;
/*
* We use %%dx and not %1 here because i/o is done at %dx and
not at
* %edx, while gcc-2.2.2 generates inferior code (movw instead
of movl)
* if we tell it to load (u_short) port.
*/
__asm __volatile("inb %%dx,%0" : "=a" (data) : "d" (port));
return data;
}
Reviewed by:
Submitted by:
FDs and fork before redialing) to prevent dialer from hanging with
newer sio.c.
Also moved history from the source to the cvs log (here) per
David G's suggestion.
This compiles but needs testing now... Rich
/*
* Hacks to support "-a|c|n" flags on the command line which enable VJ
* header compresion and disable ICMP.
* If this is good all rights go to B & L Jolitz, otherwise send your
* comments to Reagan (/dev/null).
*
* nerd@percival.rain.com (Michael Galassi) 92.09.03
*
* Hacked to change from sgtty to POSIX termio style serial line control
* and added flag to enable cts/rts style flow control.
*
* blymn@awadi.com.au (Brett Lymn) 93.04.04
*
* Put slattach in it's own process group so it can't be killed
* accidentally. Close the connection on SIGHUP and SIGINT. Write a
* syslog entry upon opening and closing the connection. Rich Murphey
* and Brad Huntting.
*
* Add '-r command' option: runs 'command' upon recieving SIGHUP
* resulting from loss of carrier. Log any errors after forking.
* Rich 8/13/93
*
* This version of slattach includes many changes by David Greenman, Brian
* Smith, Chris Bradley, and me (Michael Galassi). None of them are
* represented as functional anywhere outside of RAINet though they do work
* for us. Documentation is limited to the usage message for now. If you
* make improovments please pass them back.
*
* Added '-u UCMD' which runs 'UCMD <old> <new>' whenever the slip
* unit number changes where <old> and <new> are the old and new unit
* numbers, respectively. Also added the '-z' option which forces
* invocation of the redial command (-r CMD) upon startup regardless
* of whether the com driver claims (sometimes mistakenly) that
* carrier is present. Also added '-e ECMD' which runs ECMD before
* exiting.
*
* marc@escargot.rain.com (Marc Frajola) 93/09/10
*
* Minor fixes to allow passive SLIP links to work (connections with
* modem control that do not have an associated dial command). Added
* code to re-check for carrier after dial command has been executed.
* Added SIGTERM handler to properly handle normal kill signals. Fixed
* bug in logic that caused message about no -u command to be logged
* even when -u was specified and the sl number changes. Tried to get
* rid of redundant syslog()'s to minimize console log output. Improved
* logging of improper command line options or number of command
* arguments. Removed spurious newline characters from syslog() calls.
*
* gjung@gjbsd.franken.de
*
* sighup_handler changed to set CLOCAL before running redial_cmd.
* added flag exiting, so exit_handler is not run twice. Fork
* before reopening tty.
*/
Added a missing #ifdef INET wrapper in lpt.c
Main change:
Removed the timeout_func_t casts from timeout calls and
correctly defined the timeout routines to conform to
the new format.
lpt.c doesn't have this change.
Reviewed by:
Submitted by:
a slight change in how profiled version is selected - may need to adjust
some .mk macros if PROF is foolishly initialized anywhere to a null value.
Submitted by: jkh
