phk [Sun, 3 May 1998 08:35:05 +0000 (08:35 +0000)]
Patches are given here for pcisupport.c to recognise most of VIA
Technologies' Socket 7 chipsets. This covers all of the Apollo chipsets
except the Master (82C570) and the MVP3, and it also covers the cheap
VXPro and VXTWO knockoffs of the VP1 and VPX.
PR: 6481
Reviewed by: phk
Submitted by: Lee Cremeans <lcremean@tidalwave.net>
peter [Sun, 3 May 1998 05:25:09 +0000 (05:25 +0000)]
Put named-bootconf (the script to convert named.boot to named.conf) in
/usr/sbin rather than in /etc/namedb (make install isn't supposed to touch
/etc)
peter [Sun, 3 May 1998 04:55:44 +0000 (04:55 +0000)]
Update for some -current quirks, and some other things taken from the
*bsd bind-8 ports.
(our setpwent() was changed to return void, but our setgrent() returns
int still!)
bde [Sun, 3 May 1998 04:17:48 +0000 (04:17 +0000)]
Backed out previous commit. chown(8) doesn't follow symlinks by
default, at least in BSD. This used to be automatic, because chown(2)
didn't follow symlinks. When chown(2) was changed to follow symlinks
in BSD4.4, chown(8) was changed to not follow symlinks by default.
The previous commit broke this. The first victim was bsd.prog.mk,
which uses a plain chown in an attempt to change the ownership of the
symlinks to `dm' in /usr/games. This fails when it is done before
dm is installed, or messes up the ownership of dm if dm is installed.
Unfixed problems:
1. When lchown(2) was implemented, chown(8) wasn't changed to implement
the historical behaviour of changing ownership of symlinks. I'm not
sure if it should have been. The -HLP options give more complete
control, but they unfortunately don't apply unless the -R option is
specified (a problem shared with other commands, e.g., cp; I guess
we're supposed to use -R even for non-recursive traversals).
2. If we implement the historical behaviour, then -h would become a no-op
and should be left undocumented.
3. The man page suggests that without option -h, all symlinks (to files
specified in the command line?) are followed. It's not clear what
"the file" is. These bugs were introduced when -h was documented.
4. The correct interaction of -h with the other flags is not clear.
peter [Sat, 2 May 1998 15:51:54 +0000 (15:51 +0000)]
Resolve some unexpected differences when comparing with the 2.2 version.
One bug was relatively harmless (select's timeout had an uninitialized
tv_usec), the other I'm not so sure.. (neglected to catch select returns
less than zero). Both of these were irrelevant on kernels with poll().
peter [Sat, 2 May 1998 13:11:02 +0000 (13:11 +0000)]
Update libc dns code to 4.9.7-T1B level. This involved chopping out large
chunks of res_comp.c and replacing it with chunks of bind-8.1.1's resolver
code. (There are no interface changes though)
The other parts are better bounds checking related.
dyson [Sat, 2 May 1998 06:36:16 +0000 (06:36 +0000)]
Another minor cleanup of the split code. Make sure that pages are
busied during the entire time, so that the waits for pages being
unbusy don't make the objects inconsistant.
If not interactive, we need the traditional behaviour of the input
not being echoed to the output. So as a _hack_ to get the world building
again, redirect the readline rl_outstream to stderr when not interactive.
The proper way to handle non-interactive mode is to read from stdin
and don't worry about edit mode, but this is GNU so it's not worth the
time thinking about. I'm already pissed off that I even had to look
at this "nice code".
peter [Sat, 2 May 1998 03:02:13 +0000 (03:02 +0000)]
Seatbelts for vm_page_bits() in case a file offset is passed in rather than
the page offset. If a large file offset was passed in, a large negative
array index could be generated which could cause page faults etc at worst
and file corruption at the least. (Pages are allocated within file
space on page alignment boundaries, so a file offset being passed in here
is harmless to DTRT. The case where this was happening has already been
fixed though, this is in case it happens again).
bde [Fri, 1 May 1998 18:10:50 +0000 (18:10 +0000)]
Support compiling with `gcc -ansi'. Just use __inline instead of inline.
[__]inline is only used to bloat the code here. It gives a separate copy
of all the strings for each time this header is included...
bde [Fri, 1 May 1998 16:29:27 +0000 (16:29 +0000)]
Partially fixed write clustering for cases where cluster_wbuild() is
called from vfs_bio_awrite() without going through cluster_write()
or ufs_bmaparray(), in particular for all writes to block disk devices.
Only ufs_bmaparray() sets vp->v_maxio in a correct way, and it doesn't
seem to be called early enough even for regular files.
peter [Fri, 1 May 1998 15:10:59 +0000 (15:10 +0000)]
vm_page_is_valid() wasn't expecting a large offset argument, it's
expecting a sub-page offset. We were passing the file position,
and vm_page_bits() could do some interesting things when base was
larger PAGE_SIZE.
if (size > PAGE_SIZE - base)
size = PAGE_SIZE - base;
is interesting when (PAGE_SIZE - base) is negative. I could imagine that
this could have interesting consequences for memory page -> device block
bit validation.
bde [Fri, 1 May 1998 14:44:00 +0000 (14:44 +0000)]
Fixed apparent bitrot (`=' changed to `?=') in the definition of BINDIR
in the previous commit. Just don't define it here at all. This works
now that the default is inherited properly.
bde [Fri, 1 May 1998 14:37:36 +0000 (14:37 +0000)]
Removed self-inclusion-prevention ifdef. It is unnecessary now that
bsd.man.mk doesn't include ${.CURDIR}/../Makefile.inc.
Removed GDBDIR-redefinition-prevention ifdef. It hasn't done anothing
for a long time, if ever. The directory is defined to the same value in
each subdir and had the same value because all subdirs are at the same
level. Keep defining it in the subdirs since that is more flexible and
no more verbose.
Prepare to inherit BINDIR by including ../Makefile.inc.
bde [Fri, 1 May 1998 12:01:57 +0000 (12:01 +0000)]
Added missing -DHAVE_CONFIG_H to CFLAGS. It happens to be a no-op here
only because dc doesn't use anything in the gnu library that has a
significant dependency on config.h.
jkh [Fri, 1 May 1998 11:36:59 +0000 (11:36 +0000)]
Add -pipe to default CFLAGS. The optimization it provides is cheap
and does not require any special action on the part of the user to
take advantage of it. And no, it probably won't work with c89. Cry me
a river!
andreas [Fri, 1 May 1998 10:01:02 +0000 (10:01 +0000)]
Style:
$Id$ should be preceded by a tab
Don't include ../Makefile.inc when it is not used explicitly
Use the normal amount of horizontal and vertical whitspace (1 tab/none)
Don't override the (correct) default for MAN1
Use the correct order for -I paths
Use config.h generated by `configure', don't use a huge CFLAGS statement
Enable useage of libreadline in config.h, configure didn't enable it itself.
andreas [Fri, 1 May 1998 09:51:31 +0000 (09:51 +0000)]
Style:
- Makefiles shouldn't have copyrights.
- $Id$ should be preceded by a tab.
- Don't include ../Makefile.inc when it is not used explicitly.
- Use the normal amount of horizontal and vertical whitspace (1 tab/none).
- Don't override the (correct) default for MAN1.
- Use the correct order for -I paths.
- Don't use += to initialize SUBDIR.
- use the config.h generated by `configure' and don't use a huge
CFLAGS statement.
I think the other Makefiles under src/gnu needs some polishing as well ;-)
Thanks to Bruce, everythig looks smarter now.
Obtained from: Bruce Evans <bde@zeta.org.au>
peter [Thu, 30 Apr 1998 17:31:22 +0000 (17:31 +0000)]
partially revert rev 1.2 spammage. This file is broken as shipped and
depends on the typo in the #ifdef in order to work.. Since the line has
been touched, leave a note there so that nobody else tries to "fix" it
again.
peter [Thu, 30 Apr 1998 16:48:20 +0000 (16:48 +0000)]
Change euid while reading the user's .login_cap file in case the homedir
is on a NFS partion without root read access. Also, flip euid again for
the duration of the chdir() to the homedir for the same reason.
PR: 5145
Submitted by: Joel.Faedi@esial.u-nancy.fr
Also tested by: A Joseph Koshy <koshy@india.hp.com>
Oops, backout the previous change having confused my underscores.
__thread_create is a syscall that uses the default asm. It is
_thread_create that contains specific asm code, but that lives in
libpthread.
Make errno finally and permanently thread-aware so that it is no longer
necessary to compile with _THREAD_SAFE defined. This means that people
will get thread-aware code whether they like it or not. This change
is required to allow a process to link against libpthread and libc
to use kernel threads (and prevent each thread from clobbering another
thread's errno just be doing a syscall).
This is bound to break some ports, but it is strictly allowed by ANSI C,
so anything that breaks as a result was already broken anyway 8-).
"Sorry".
Make cerror thread aware by calling __error() to get a pointer to the
thread-specific error variable. This change make libc use the same cerror
code that libc_r has been using.
The syscall that creates a kernel thread is coming, but it doesn't use the
default syscall asm, so add it to NOASM. The other syscalls that manipulate
kernel threads use the default asm code, so they just get built
automatically.
Build __error.c into libc, but not libc_r. The weak symbol in the
file works with libpthread, but when built into libc_r which has a non-weak
symbol of the same name, the linker behaves unpredicatably and sometimes
links the wrong symbol. The linker behaviour is a byproduct of what
the program calls from object to object so it is like winning a lottery
if the program actually works. The odds are quite good - 95:1, I think.
We need a sure thing, though, so weak symbols can't be used instead
of renaming things.
Change the name of this source file so that libc_r builds it instead
of the one in libc that contains the weak symbol for __error. FreeBSD's
make accumulates paths to the point that it can find *anything*, possibly
including the car keys.