Peter Wemm [Wed, 30 Oct 1996 08:49:33 +0000 (08:49 +0000)]
Regenerate.
No changes other than "instructions" for what other tools that use this
.. "interesting" .. system should name their shared libraries. This was
broken for some tools (eg: expect-5.21) even for the native tcl build (ie,
it would end up using libexpect521.so, not libexpect521.so.1.0)
Peter Wemm [Wed, 30 Oct 1996 08:44:18 +0000 (08:44 +0000)]
Argh, I hate the tcl build/config environment. :-(
It seems that some tools (eg: expect-5.21) use TCL_LIB_SPEC to generate
their shared library suffix. This should be .so.1.0, not .so as ld can't
use it.
Revert part of the previous change here, it did too much. libtcl75.so.1.1
was bumped, but the rules to generate library names for _other_ packages
were not supposed to be. Sigh.
Peter Wemm [Wed, 30 Oct 1996 06:13:10 +0000 (06:13 +0000)]
Fix braino on my part. When we have three different port ranges (default,
"high" and "secure"), we can't use a single variable to track the most
recently used port in all three ranges.. :-] This caused the next
transient port to be allocated from the start of the range more often than
it should.
John Dyson [Wed, 30 Oct 1996 03:45:00 +0000 (03:45 +0000)]
Fix the /proc/???/map file so that it is possible to read an arbitrarily
large process map. Another commit will follow to fix a problem just found
during this one... Sorry!!! :-(.
John Dyson [Tue, 29 Oct 1996 22:07:11 +0000 (22:07 +0000)]
Change mmap to use OBJT_DEFAULT instead of OBJT_SWAP by default
for anonymous objects. The system will automatically change the
type to SWAP if needed (for size or pageout reasons.)
Gary Jennejohn [Tue, 29 Oct 1996 21:52:21 +0000 (21:52 +0000)]
changes to gdb to:
1) add Garrett Wollman's trap frame resolving mods
2) make the `proc' command (kernel debugging) really work
3) allow use of a pid with the `proc' command (previously you had to
provide the address of the proc structure)
Unfortunately, the `proc' command won't work while doing remote debugging.
Bruce Evans [Tue, 29 Oct 1996 13:15:30 +0000 (13:15 +0000)]
Search for labels on all slices whenever any minor for a drive is opened.
If DEVFS is configured, create devfs devices for previously invisible
partitions on the slices.
Fixed an old aliasing bug which caused E=17 errors from DEVFS for
DIOCSDINFO when there were no real slices.
Peter Wemm [Tue, 29 Oct 1996 05:35:24 +0000 (05:35 +0000)]
Convert the DONT_FSYNC compile time option to a runtime option (like
nobiff). The options to turn these on are specified in the
LOCAL_MAILER_ARGS define the the sendmail.mc that you build from.
Peter Wemm [Tue, 29 Oct 1996 05:22:52 +0000 (05:22 +0000)]
Merge in our changes from src/libexec/mail.local. This version doesn't
violate disk quotas, has more robust locking, is still being developed,
etc. The main changes merged in were the ability to not fsync() the
mailbox, not do biff/comsat broadcasts, man page fixes.
Peter Wemm [Tue, 29 Oct 1996 04:31:21 +0000 (04:31 +0000)]
Along the way, the minor version number has been bumped (newer version
of tcl with new functions in the library), but the ports-compatable
link name wasn't bumped as it should have been.
Minor cleanups while here (including preserving the $Id: line)
John Hay [Mon, 28 Oct 1996 19:40:01 +0000 (19:40 +0000)]
Remove some unnecessary code in the interrupt handler.
Check that a received packet isn't longer than MCLBYTES. This will
sometimes happen if a cable is plugged into or removed from a live
system.
Try to cater better for early receive interrupts.
Justin T. Gibbs [Mon, 28 Oct 1996 17:00:24 +0000 (17:00 +0000)]
In data_phase_reinit when I converted the code to use bcopy, I should have
used mvi instead of mov. Luckily this code is most likely never executed
since it is only there for sanity should a target goes into the data phase
twice during a single selection or reselection.
Joerg Wunsch [Mon, 28 Oct 1996 08:28:02 +0000 (08:28 +0000)]
Correct the startup sequence for syslogd:
. crate the symlink for /dev/log if required, then
. remove the old socket (savecore might have already created it
accidentally), then
. start syslogd.
(Btw., our test(1) misses an option to test for a socket.)
Justin T. Gibbs [Mon, 28 Oct 1996 06:10:33 +0000 (06:10 +0000)]
Add basic support for the 398X cards as multi-channel SCSI host adapters.
This involves expanding the support of the SEEPROM routines to deal with
the larger SEEPROMs on these cards and providing a mechanism to share
SCB arrays between multiple controllers.
Most of the 398X support came from Dan Eischer.
ahc_data -> ahc_softc
Clean up some more type bogons I missed from the last pass.
Justin T. Gibbs [Mon, 28 Oct 1996 06:10:02 +0000 (06:10 +0000)]
Add basic support for the 398X cards as multi-channel SCSI host adapters.
This involves expanding the support of the SEEPROM routines to deal with
the larger SEEPROMs on these cards and providing a mechanism to share
SCB arrays between multiple controllers.
Most of the 398X support came from Dan Eischer.
ahc_data -> ahc_softc
Clean up some more type bogons I missed from the last pass.
Be more clear when handing the NO_MATCH condition. NO_MATCH can also
happen when the sequencer encounters an SCB we've asked to be aborted.
Justin T. Gibbs [Mon, 28 Oct 1996 06:05:58 +0000 (06:05 +0000)]
Add two new aic7xxx driver options:
AHC_FORCE_PIO - This forces the driver to use PIO even on systems that
say they have memory mapped the controller's registers. This
seems to fix Ken Lam's problems. I've also placed this option
in the GENERIC kernel file so that we are guaranteed to install
even on these flakey machines.
AHC_SHARE_SCBS - This option attempts to share the external SCB SRAM on
the 398X controllers allowing a totoll of 255 non-paged SCBs.
This doesn't work quite yet, so this option is mostly here to
help 398X owners to experiment and give me feedback until this
works properly.
Justin T. Gibbs [Mon, 28 Oct 1996 06:01:08 +0000 (06:01 +0000)]
Fix problems dealing with non-tagged devices when SCB paging is enabled.
Mostly this involved changing the semantics of the findSCB routine so that
it could be used at times other than handling a reconnection.
The way we get a vnode for swapdev is not quite kosher. In particular
it breaks in the DEVFS_ROOT case. replicate a bit too much of bdevvp()
in here to circumvent the problem. The real problem is the magic that
lives in bdevsw[1].
Søren Schmidt [Sat, 26 Oct 1996 20:16:58 +0000 (20:16 +0000)]
Use the calibrated/adjustable i8254 frequency `timer_freq' instead of
TIMER_FREQ.
Fixed missing splx() in scrn_timer(). The bug was harmless because of the
undocumented behaviour that the ipl is automatically restored for timeout
functions (see softclock()). Perhaps we should depend on this behaviour.
Fixed the ddb fix in rev.1.176. The in_debugger flag was no use because
it only works when the debugger is entered via the keyboard hotkey. The
debugger may be entered for breakpoints and traps, and the console putc
routine has no way of knowing when it was, so the console putc routine
must (almost?) always remove the cursor image.
Not fixed: console switching in ddb doesn't work (ISTR it working), and
console 0 shouldn't be switched to for the debugger hotkey unless console
0 is /dev/console.
Fixed side effects from calling add_keyboard_randomness() in the console
getc routine by not calling it. add_keyboard_randomness() currently
always reenables interrupts on 386's and 486's. This is very bad if the
console getc routine is called from the debugger and the debugger was
entered with interrupts disabled.
Fixed preservation of initial screen and now-bogus comment about it. It
was broken by setting the initial scr_buf to `buffer' instead of Crtat.
`buffer' was full of nulls and the first scroll cleared everything above
the things written through syscons.
Mark Murray [Sat, 26 Oct 1996 17:20:52 +0000 (17:20 +0000)]
First round of moving secure telnet (AKA kerberised telnet) to its new home
in eBones/. Peter has already (ages ago) done the repository copy. These
are mainly -Wall cleanups. The makefile changes will follow in a day or so.
Warner Losh [Fri, 25 Oct 1996 18:14:48 +0000 (18:14 +0000)]
Fix a problem where the command line could be used to overflow a stack
buffer which could be made to lead to a root shell. This patch is
OpenBSD's solution to the problem, and will silently truncate the
output rather than overflow the buffer.
Bill Fenner [Fri, 25 Oct 1996 17:57:53 +0000 (17:57 +0000)]
Don't allow reassembly to create packets bigger than IP_MAXPACKET, and count
attempts to do so.
Don't allow users to source packets bigger than IP_MAXPACKET.
Make UDP length and ipovly's protocol length unsigned short.
Bill Paul [Fri, 25 Oct 1996 16:13:09 +0000 (16:13 +0000)]
Give ypxfr the ability to detect the presence of the YP_INTERDOMAIN
and YP_SECURE flags so that it can properly add them to newly created
maps when needed. This applies only when using the 'standard' method
for map transfers. When using rpc.ypxfrd, the whole map is copied
verbatim, along with any special entries that may be encoded in it.
Also made -Wall a little quieter for ypxfrd_getmap.c.
Bruce Evans [Fri, 25 Oct 1996 15:50:17 +0000 (15:50 +0000)]
Use the same format for the "expect N shift reduce conflicts" message
as in the one other place in /usr/src that prints such an "expect"
message (amd).
Bruce Evans [Fri, 25 Oct 1996 15:36:42 +0000 (15:36 +0000)]
Made this work for non-root again. The security fixes in rev.1.11
were equivalent to disabling its setgid'ness, since the default
kernel was not distinguished from a user-specified kernel.
Bruce Evans [Fri, 25 Oct 1996 14:49:24 +0000 (14:49 +0000)]
Build intermediate object files for mkinit, mknodes and mksyntax
so that simple regresssion tests based on `cmp' work. mkdep still
doesn't work right for these tools. They should probably be in
separate directories.
Bruce Evans [Fri, 25 Oct 1996 13:01:56 +0000 (13:01 +0000)]
Improved biasing of i586 clock by adjusting for hardclock() latency.
I decided to do this for every hardclock() call instead of lazily
in microtime(). The lazy method is simpler but has more overhead
if microtime() is called a lot.
CPU_THISTICKLEN() is now a no-op and should probably go away.
Previously it did nothing directly but had the side effect of
setting i586_last_tick for CPU_CLOCKUPDATE() and i586_avg_tick for
debugging. CPU_CLOCKUPDATE() now uses a better method and
i586_avg_tick is too much trouble to maintain.
Reduced nesting of #includes in the usual case.
Increased nesting of #includes when CLOCK_HAIR is defined. This
is a kludge to get typedefs for inline functions only when the
inline functions are used. Normally only kern_clock.c defines
this. kern_clock.c can't include the i386 headers directly.
Bruce Evans [Fri, 25 Oct 1996 12:26:19 +0000 (12:26 +0000)]
Removed #include of <machine/clock.h>. It is no longer used, and would
break when I remove LOCORE support from clock.h.
I586_CTR_MULTIPLIER_SHIFT = 32 from clock.h is actually still used, but
32 is so magic that it doesn't get used explicitly.
Justin T. Gibbs [Fri, 25 Oct 1996 06:42:53 +0000 (06:42 +0000)]
- KNF cleanup.
- Add support for memory mapped I/O.
- Use DMA to get SCBs down to the adapters.
- Remove old paging code.
- Be much smarter about how we allocate SCB space. The old, simple method
wasted almost half a page per SCB. Ooops.
- Make command complete interrupt processing more efficient.
- Break the monolithic ahc_intr into sub-routines. The sub-routines handle
rare, special case events so the function call is not a penalty and the
removal of the code from the main routine most likely improves performance
instruction prefech will work better and less code is pushed into the cache.
- Never, ever allow tagged queueing if a device has disconnection disabled.
- Clean up and simplify timeout code. Many of the changes are to handle the
new DMA scheme.
Justin T. Gibbs [Fri, 25 Oct 1996 06:34:59 +0000 (06:34 +0000)]
Go back to using DMA to get SCBs down to the adapter.
SCB paging is now handled almost entirely by the sequencer and also uses
DMA. This should make SCB paging at least an order of magnitude more
efficient and vastly simplifies the implementation.
Add a few space optimizations so this code still fits on aic7770 chips.