Bruce Evans [Tue, 18 Feb 1997 14:37:26 +0000 (14:37 +0000)]
Added some ufs #includes so that this compiles with option LOCKF_DEBUG.
Moving this all from ufs wasn't a good move. At least the debugging
routines depend on the file system.
When wrong directory choosed and chdir failed, directory name
was not reset to old name causing any file choosen put error
diagnostic about wrong directory, fix it by resetting back
to old name after chdir failed.
Add \r as alias to \n, some telnets have problem with that.
Should go into 2.2
Submitted by: "Anatoly A. Orehovsky" <tolik@mpeks.tomsk.su> & me
Bruce Evans [Tue, 18 Feb 1997 06:46:41 +0000 (06:46 +0000)]
Fixed namei caching for LOOKUPs. It was broken for lstat() and olstat().
Successful lstat()s purged an existing entry as well as not caching the
result.
This bug was introduced in Lite1 by setting the LOCKPARENT flag for
[o]lstat() in order to support the inherit-attributes-from-parent-
directory misfeature for symlinks. LOCKPARENT was previously only set
for CREATEs and DELETEs. It is now set for LOOKUPs, but only for
[o]lstat(), so the problem wasn't very noticeable.
Bruce Evans [Tue, 18 Feb 1997 04:58:11 +0000 (04:58 +0000)]
Changed `#ifdef COMPAT_PRELITE2' to `#ifndef NO_COMPAT_PRELITE2' so that
the old VFS_VFSCONF sysctl is enabled by default.
Initialize the vfc_vfsops field to non-NULL in sysctl_ovfs_conf()
so that the old VFS_VFSCONF sysctl actually works. The old (still
current) getvfsent.c uses this "kernel-only" field to decide which
vfs's are configured (the old implementation returned null entries
for unconfigured vfs's).
Justin T. Gibbs [Tue, 18 Feb 1997 04:25:31 +0000 (04:25 +0000)]
Don't rely on AAP(Auto Access Pause) when queueing SCBs to the card. This
will increase the overhead of queueing a command, but some recent bug reports
make me believe that AAP isn't really working and that we are losing some
SCBs from the input queue. Hopefully this will cure that problem.
Fix some bugs in the error recovery code. Mainly these could cause us to
inadvertantly forget to untimeout an SCB that was recovered causing later
confusion.
Joerg Wunsch [Mon, 17 Feb 1997 18:56:29 +0000 (18:56 +0000)]
Various sanitizations:
. Don't clobber the TERM setting; it's supposed to be done by /etc/ttys
already.
. Comment out the Interviews stuff, 98 % of all users probably won't
ever use it.
. Install the files with better default permissions in the skeleton
directory; pw(8) retains the permissions when creating a new
acount, and installing them read-only is stupid, yet installing
.rhosts world-readable is dangerous.
David Greenman [Mon, 17 Feb 1997 10:58:46 +0000 (10:58 +0000)]
Pass P_SUGID on to the child of a fork(). It was possible to get rlogin
to coredump previously since it (somewhat uniquely) is setuid and forks
without execing, and thus without passing P_SUGID the child could
coredump and possibly divulge sensitive information (such as encrypted
passwords from the passwd database).
John-Mark Gurney [Mon, 17 Feb 1997 08:30:16 +0000 (08:30 +0000)]
Bring intro(3) a bit closer to reality:
1) comment out xref to non-existant libc(3)
2) comment out reference to Fortran specific section (3f)
3) add libkvm with reference to kvm(3)
4) comment out xref to non-existant pc(1)
5) comment out libplot and libplotf77
6) fix problem with -ltermcap not being parenthesized
7) sort files listing
still missing most of the libraries that exist in /usr/lib.
Always load DIST_COMPAT21 when loading X. This used to be the default
then it got taken out when the COMPAT21 distribution got dangerous.
Now that it's safe (we hope), it comes back.
Don't nuke localtime and timezone files so aggressively now.
This is tzsetup's job (and only if you don't cancel out, which
is why having sysinstall remove them was so bogus).
1. Always write out /etc/hosts.
2. Back out my change to ask about UTC/Localtime here. This *really* needs
to be done in tzsetup instead since putting it here only handles about
1/4 of the places where it needs to be.
Bugfixes for HTML output:
- use ${MACROS} for macro package and not hard-coded -ms
Note: -me is currently not supported by unroff(1).
- use only ${SRCS} as input files.
Bruce Evans [Sun, 16 Feb 1997 18:26:31 +0000 (18:26 +0000)]
Select between the generic math functions and the i387-specific ones
at runtime.
etc/make.conf:
Nuked HAVE_FPU option.
lib/msun/Makefile:
Always build the i387 objects. Copy the i387 source files at build
time so that the i387 objects have different names. This is simpler
than renaming the files in the cvs repository or repeating half of
bsd.lib.mk to add explicit rules.
lib/msun/src/*.c:
Renamed all functions that have an i387-specific version by adding
`__generic_' to their names.
lib/msun/src/get_hw_float.c:
New file for getting machdep.hw_float from the kernel.
sys/i386/include/asmacros.h:
Abuse the ENTRY() macro to generate jump vectors and associated code.
This works much like PIC PLT dynamic initialization. The PIC case is
messy. The old i387 entry points are renamed. Renaming is easier
here because the names are given by macro expansions.
Mike Pritchard [Sun, 16 Feb 1997 18:24:44 +0000 (18:24 +0000)]
Don't call sendmail with "-f root" when run with "-a". This
eliminates the "X-authentication-warning" header line that
has been coming out since I made it so that sendmail is
run totally as the user whose calendar file is currently
being processed.
Bruce Evans [Sun, 16 Feb 1997 17:54:58 +0000 (17:54 +0000)]
Use the C library version of log10() instead of the inaccurate formula
log10(x) = log10e * log(x). The formula would work if the RHS were
evaluated in extended precision with an extended precision log().
This actually happened with the i387 log() because it returns excess
precision.
Steve Price [Sun, 16 Feb 1997 01:50:25 +0000 (01:50 +0000)]
Implement the -o commandline switch. This does almost exactly
what bison does with -o, except it DTRT with respect to naming
the .code file when the -r switch is also used.
Bruce Evans [Sat, 15 Feb 1997 17:09:00 +0000 (17:09 +0000)]
Restored my fixes for printing the FP registers. Their
history was lost with FreeBSD-1.x and they were blown away
by the gdb-4.16 merge. I needed this to debug e_exp.S ...
Restored even older code (from 386BSD-0.0) for converting the
FP registers to doubles. floatformat_to_double() and/or
valprint() still don't understand NaNs.
Removed unnecessary #include of obsolete <sys/dir.h> again.
Bruce Evans [Sat, 15 Feb 1997 16:14:52 +0000 (16:14 +0000)]
Don't use the private version of vasprintf(). It doesn't understand
"%%" in format strings and tends to dump core for "%%st". I needed
"%%st" to fix the new gdb ...
Don't use the private version of strerror() either.
Use INTERNALLIB and INTERNALSTATICLIB instead of a private install
rules NOPROFILE and NOPIC. This is only slightly cleaner.
INTERNALLIB was previously only used in compatibility libraries
(libgnumalloc etc.) and INTERNALSTATICLIB was previously unused.
INTERNAL*LIB probably should be replaced by something like NOSTATICO
together with NO{STATICO,PROFILE,PIC}INSTALL.
Now that this is becoming (dare I even say it?) more useful for
post-configuration, no longer leave sysinstall.debug files around
by default. Only do this if environment variable SYSINSTALL_DEBUG
is set.
Added an option `-h' which gives you usage info.
(According to the original man page, giving `-' by itself as the argument
is supposed to do this, but it seems that it never worked that way.)
Bruce Evans [Sat, 15 Feb 1997 05:21:16 +0000 (05:21 +0000)]
Disabled the i387 version if log1p(). It just evaluates log(1 + x).
This defeats the point of log1p(). ucbtest reports errors of +-5e+15
ULPs. A correct version would use the i387 fyl2xp1 instruction for
small x and maybe scale to small x. The C version does the scaling
reasonably efficiently, and fyl2px1 is slow (at least on P5s), so not
much is lost by always using the C version (only 25% for small x even
with the broken i387 version; 50% for large x).