Bring back netns so that Netcon can take over support for it, as agreed.
This was actually easier than I thought - CVS is smart enough to handle
"resurrections" now. :-)
Guido van Rooij [Thu, 17 Oct 1996 18:27:58 +0000 (18:27 +0000)]
When freeing buffers in the db routines, also zeroize them
This should solve the bug where a coredumping ftpd reveals
encrypted passwords.
Obtained from: OpenBSD
I'm not sure why, but Netcon's TFS filesystem code doesn't want to
add free vnodes back to the freelist. They must do their own vnode
management. Anyway, this change is *only* activated with their filesystem
and doesn't affect anyone else. Whoops, forgot the submitted-by lines
in my previous commits too.. :-( Submitted-By: Tony Ardolino <tony@netcon.com>
Don't ever ask for password if it is impossible to confirm it
It happens if 1) regular passwords not allowed, 2) skey database
not activated for given user.
Under some rare circumstanes skey_challenge can return empty
diagnostic or even previous buffer, fix it.
Prevent namespace pollution cause by this library by splitting to
API and non-API functions and giving _sk_ prefix for all non-API
functions. Old names will be available by _SKEY_INTERNAL define
it allows not change anything in keyinit & key, only -D_SKEY_INTERNAL required.
Staticise some things.
If 1) regular Unix passwords not permitted and 2) skey database
for this user not activated, there no correct password exist,
so don't bother ask for it and return "" immediately
John Dyson [Thu, 17 Oct 1996 03:04:43 +0000 (03:04 +0000)]
Fix a problem that could cause msync (or many other things) to deadlock.
The heuristic for managment of memory backing the buffer cache was
nice, but didn't work due to some architectural problems. Simplify
and improve the algorithm.
Alexander Langer [Thu, 17 Oct 1996 01:05:03 +0000 (01:05 +0000)]
Issue a warning if the user specifies an invalid interface in a rule.
The rule is still added to the chain since the interface may get
created later on after loading an LKM.
Peter Wemm [Thu, 17 Oct 1996 00:40:04 +0000 (00:40 +0000)]
Corrently null-terminate the path being passed to the opendir() calls,
malloc() does is not defined to return a zeroed buffer leading to
"strange" problems.
Submitted by: Karl Denninger <karl@mcs.com>, PR#1826
Bruce Evans [Wed, 16 Oct 1996 21:02:49 +0000 (21:02 +0000)]
Use a (signed) int32_t counter instead of an `unsigned int' counter
for the GPROF4 case. This allows a simpler method to be used for
non-statistical profiling (it allows overhead adjustments to be
subtracted from one counter without harm if that counter goes
negative; otherwise the adjustment would have to be distributed).
32 bit counters were already too small for GPROF4 with a 200MHz
clock. int64_t counters should be used.
Bruce Evans [Wed, 16 Oct 1996 18:37:58 +0000 (18:37 +0000)]
Fixed miscounting for non-statistical (GUPROF) profiling:
- use CROSSJUMP() and CROSSJUMP_LABEL() for conditional jumps from idle()
into cpu_switch() and vice versa.
- moved badsw code to after cpu_switch().
Cosmetic changes:
- moved sw0 string to be immediately after its caller (badsw).
- removed unused #include.
Søren Schmidt [Wed, 16 Oct 1996 18:16:22 +0000 (18:16 +0000)]
Utility to "brand" ELF binaries.
Man page to come...
For now use: <brandelf -t Linux linuxbin> to brand, and just
<brandelf> to verify branding on a ELF file. FreeBSD native is
set with <brandelf -t FreeBSD freebsdbin>.
Julian Elischer [Wed, 16 Oct 1996 00:19:40 +0000 (00:19 +0000)]
Add support for embedded operation withou console
The boot.c patch is applied only to teh JULIAN_HACK branch
the muted console is controlable by a sysctl variable kern.consmute
Save half of space in LC_COLLATE and remove unneded code.
This change is not compatible with previous variant, however proper
error code returned in both cases.
Colldef changes will follows.
Bruce Evans [Tue, 15 Oct 1996 18:24:34 +0000 (18:24 +0000)]
Removed nested include of <sys/proc.h> from <vm/vm_object.h> and fixed
the one place that depended on it. wakeup() is now prototyped in
<sys/systm.h> so that it is normally visible.
Added nested include of <sys/queue.h> in <vm/vm_object.h>. The queue
macros are a more fundamental prerequisite for <vm/vm_object.h> than
the wakeup prototype and previously happened to be included by
namespace pollution from <sys/proc.h> or elsewhere.
John Dyson [Tue, 15 Oct 1996 03:16:45 +0000 (03:16 +0000)]
Move much of the machine dependent code from vm_glue.c into
pmap.c. Along with the improved organization, small proc fork
performance is now about 5%-10% faster.
Bruce Evans [Mon, 14 Oct 1996 21:44:05 +0000 (21:44 +0000)]
Fixed setting of the keyboard LEDs. For some keyboards and/or keyboard
controllers, it is an error to issue a command before the keyboard
has had time to reply to the previous command. Setting the LEDs
involves issueing 2 commands, so it never worked on these keyboards.
Fixed resetting of keyboard. It is possible for unprocessed
scancodes to be present when the reset routine is called. This
usually occurs after switching from one console driver to another
in userconfig. pcvt and syscons attempt to flush any stale scancodes,
but sometimes fail to do so because keyboard and/or keyboard
controller takes a long time to send the scancodes after reset
(scancodes are apparently not flushed by reset!). syscons handles
this later by not checking for errors at strategic places, but pcvt
was confused.
Use an impossible initial and failure mode setting for the LEDs
so that the LEDs always get set if they are possibly out of sync.
David Greenman [Mon, 14 Oct 1996 16:43:11 +0000 (16:43 +0000)]
Fixed a bug that got introduced when I changed the CY16_RESET and
CY_CLEAR_INTR definitions. This might have affected probing of ISA
versions of the Cyclom 16/32-Y.
Stefan Eßer [Mon, 14 Oct 1996 13:04:34 +0000 (13:04 +0000)]
pci_map_mem() did a too restrictive check on the mapping type:
PCI_MAP_MEMORY_TYPE_32BIT_1M should be accepted as well as
PCI_MAP_MEMORY_TYPE_32BIT (and now is).
(Problem reported by David Greenman.)
Peter Wemm [Mon, 14 Oct 1996 12:58:47 +0000 (12:58 +0000)]
oops, how did this get into the reinstall target? :-]
Must have been a cut/paste foulup... The whole idea of 'make reinstall'
is that it doesn't build anything from the source server mount.
Stefan Eßer [Mon, 14 Oct 1996 10:09:52 +0000 (10:09 +0000)]
Move the initialization of np->ns_sync and ns_async out of ncr_getclock(),
which does no longer get called for non-Ultra cards.
Fix suggested by Gerard Roudier, slightly modified by me.
Joerg Wunsch [Sun, 13 Oct 1996 20:55:45 +0000 (20:55 +0000)]
Yikes! Originally, i intended to apply the patch from PR # 1322,
but Bill has beaten me on this. ;-)
However, he missed the part to compute the kernelsize in kilobytes,
so the minfree consideration was now overcautious. (I've also
changed the return type of dump_size() to void since int was useless.)
Being here, the fact that `vmcore' was written world-readable was just
a plain security hole: everybody who was able to crash a kernel could
later read any confidential information out of it at his will. Create
it with umask 077 instead.
Bruce Evans [Sun, 13 Oct 1996 18:18:50 +0000 (18:18 +0000)]
The dos() function needs a new second argument, containing the size
of the partition. Only if the size is 0 should the
special handling of 0 as first argument be triggered.
[This bug caused offset 0 to give C/H/S = 0/0/0 instead of 0/0/1.]
The init_sector0 function needs to decrease the first argument
to the second call to dos() by one to be consistent with the
calls to dos() in change_part().
[This bug caused fdisk -i to create bogus partition tables with
the ending C/H/S value 1 too high. This usually gives S = 1
instead of S = maximum, so the geometry guessing in the slice
code and perhaps in SCSI BIOSes was defeated.]
Bill Fenner [Sun, 13 Oct 1996 18:12:20 +0000 (18:12 +0000)]
Make the savecore command work like the man page says:
- make minfree work by getting the dump size before checking to see
if the dump will fit on the filesystem
- also fail to dump if no minfree is specified but there are not enough
free blocks.
Peter Wemm [Sun, 13 Oct 1996 17:37:03 +0000 (17:37 +0000)]
Build/install c++filt from g++, we already have all the sources and
it's and useful. (Ever tried to read 'nm' and 'ldd -v' output on a c++
object or library? :-) This filter decodes the mangled symbol names.)
Joerg Wunsch [Sun, 13 Oct 1996 16:52:40 +0000 (16:52 +0000)]
Uncomment the "raw" lp entry in /etc/printcap. We start lpd by
default, so there's no use in running it without any printer
definition in printcap. Also added a bunch of hints about the printer
setup, to guide the admin about the printer setup (handbook,
"apsfilter"), and a commented-out sample setup for a remote printer.
In the same line, add /var/spool/lpd/output to BSD.var.dist since it
is referred to by the "lp" entry in printcap.
David Greenman [Sun, 13 Oct 1996 01:12:01 +0000 (01:12 +0000)]
Changed ttyc? and cuac? specifications to support multiple controllers.
A "MAKEDEV ttyc0" now creates 32 units for controller 0, ttyc1 for
controller 1, etc..
Bruce Evans [Sat, 12 Oct 1996 22:12:51 +0000 (22:12 +0000)]
Fixed lblktosize(). It overflowed at 2G. This bug only affected
ufs_read() and ufs_write().
Found by: looking at warnings for comparing the result of lblktosize()
(which is usually daddr_t = long) with file sizes (which are u_quad_t
for ufs). File sizes should probably be off_t's to avoid warnings
when the are compared with file offsets, so the fixed lblktosize()
casts to off_t instead of u_quad_t.
Added definition of smalllblksize(). It is the same as the old
lblksize() and is more efficient for small block numbers on 32-bit
machines.
Use smalllblktosize() instead of its expansion in blksize() and
dblksize(). This keeps the line length short and makes it more
obvious that the shift can't overflow.
John Dyson [Sat, 12 Oct 1996 21:37:59 +0000 (21:37 +0000)]
Undo a mistaken change from splhigh to splvm. Subr_rlist has been
capable of being used for things other than swap space allocation,
and splvm would have been appropriate for only swap space allocation
and other VM things. My commit broke that (and was actually a mistake.)
John Dyson [Sat, 12 Oct 1996 21:35:25 +0000 (21:35 +0000)]
Performance optimizations. One of which was meant to go in before the
previous snap. Specifically, kern_exit and kern_exec now makes a
call into the pmap module to do a very fast removal of pages from the
address space. Additionally, the pmap module now updates the PG_MAPPED
and PG_WRITABLE flags. This is an optional optimization, but helpful
on the X86.
Bruce Evans [Sat, 12 Oct 1996 20:36:15 +0000 (20:36 +0000)]
Cleaned up:
- fixed a sloppy common-style declaration.
- removed an unused macro.
- moved once-used macros to the one file where they are used.
- removed unused forward struct declarations.
- removed __pure.
- declared inline functions as inline in their prototype as well
as in theire definition (gcc unfortunately allows the prototype
to be inconsistent).
- staticized.
Eric L. Hernes [Sat, 12 Oct 1996 20:23:26 +0000 (20:23 +0000)]
Get the pysical address for id_maddr when opening an -incore kernel.
This fixes the kernel panic when propagating userconfig changes to
arbitrary kernels.
Remove obsoleted `#include <tcl.h>' added a few <stdio.h> where
necessary.
Fix getting scsi bus information from an -incore kernel.
Bruce Evans [Sat, 12 Oct 1996 20:09:48 +0000 (20:09 +0000)]
Removed __pure's and __pure2's. __pure is a no-op for recent versions
of gcc by definition, and __pure2 is a no-op in effect (presumably the
compiler can see when an inline function has no side effects).
Bruce Evans [Sat, 12 Oct 1996 19:49:43 +0000 (19:49 +0000)]
Removed nested include if <sys/socket.h> from <net/if.h> and
<net/if_arp.h> and fixed the things that depended on it. The nested
include just allowed unportable programs to compile and made my
simple #include checking program report that networking code doesn't
need to include <sys/socket.h>.
Some cosmetic tweaks, attempt to fix package wait problem by checking
for error return as well (not sure if this is it, but it's one possible
explanation).