Undo part of the scalability commit. Many of the changes
in vm_fault had some performance enhancements not ready
for prime time. This commit backs out some of the changes.
peter [Sat, 27 Jul 1996 21:47:42 +0000 (21:47 +0000)]
General tidy-up of Makefile. The objective was to make it clear and
readable, as well as fix up some of the warts along the way. Among the
things fixed is the infamous "cd foo && make depend all install" bogon,
which is particularly important for users of the -DNOCLEAN option.
Also, a few typo and DESTDIR fixes and a -DNOGAMES hack from Gary Jennejohn.
Allow sequentially created mmap'ed anonymous regions to coalesce. There
is little or no reason to create a swap pager for small mmap's. The
vm_map_insert code will automatically create a swap pager if the object
becomes too large. This fix, per a request from phk.
Under the heavy load for transmiting condition, it will be write error
and then never accept for sending packet from upper layer anymore
(i.e. ping -f )
Reviewed by: David Greenman <dg@root.com>
Submitted by: amurai@spec.co.jp
andreas [Sat, 27 Jul 1996 11:54:30 +0000 (11:54 +0000)]
Fixed bug, that CONFIGURE_SCRIPT in a port's Makefile doesn't have
any effect. The target do-configure used ${SCRIPTDIR}/configure
instead of ${SCRIPTDIR}/${CONFIGURE_SCRIPT}.
-andreas
This commit is meant to solve a couple of VM system problems or
performance issues.
1) The pmap module has had too many inlines, and so the
object file is simply bigger than it needs to be.
Some common code is also merged into subroutines.
2) Removal of some *evil* PHYS_TO_VM_PAGE macro calls.
Unfortunately, a few have needed to be added also.
The removal caused the need for more vm_page_lookups.
I added lookup hints to minimize the need for the
page table lookup operations.
3) Removal of some bogus performance improvements, that
mostly made the code more complex (tracking individual
page table page updates unnecessarily). Those improvements
actually hurt 386 processors perf (not that people who
worry about perf use 386 processors anymore :-)).
4) Changed pv queue manipulations/structures to be TAILQ's.
5) The pv queue code has had some performance problems since
day one. Some significant scalability issues are resolved
by threading the pv entries from the pmap AND the physical
address instead of just the physical address. This makes
certain pmap operations run much faster. This does
not affect most micro-benchmarks, but should help loaded system
performance *significantly*. DG helped and came up with most
of the solution for this one.
6) Most if not all pmap bit operations follow the pattern:
pmap_test_bit();
pmap_clear_bit();
That made for twice the necessary pv list traversal. The
pmap interface now supports only pmap_tc_bit type operations:
pmap_[test/clear]_modified, pmap_[test/clear]_referenced.
Additionally, the modified routine now takes a vm_page_t arg
instead of a phys address. This eliminates a PHYS_TO_VM_PAGE
operation.
7) Several rewrites of routines that contain redundant code to
use common routines, so that there is a greater likelihood of
keeping the cache footprint smaller.
peter [Fri, 26 Jul 1996 13:47:38 +0000 (13:47 +0000)]
Apply a bandaid to a problem elsewhere in the driver, when the process is
blocked in a write() while waiting for the output to drain, sleep only
for tp->t_timeout, not forever. This only seems to happen when there is
either a modem lockup holding the hardware flow control down, or due to
some problem in the driver with processes attempting to write after the
modem has hung up (eg: elm, tf).
peter [Thu, 25 Jul 1996 19:32:37 +0000 (19:32 +0000)]
Fix the services.byname target so that it creates search keys for the
aliases of the "official" names as well, because now that getportbyname()
does a yp match, it no longer found the entries under the alias.
This broke rsh(1), because it looks up "shell/tcp" while the official
name in /etc/services is "cmd/tcp".
Refer to the ASCII character 000 by its proper name, `NUL', and not
`NULL' (which should only be used in reference to null pointers). Also
fix a cross-reference.
Fix a bug in ifa_ifwithnet() which caused a page fault in bcmp()
when attepmting to add certain types of routes. This problem
only manifested itself in the presence of unconfigured point-to-point
interfaces.
peter [Wed, 24 Jul 1996 14:04:57 +0000 (14:04 +0000)]
Fix typo in last commit, it seems that a hash comments out a \ at the end
of line.
Also, fix existing bug in ethers.byname, it was passing an unknown option
to yppush. This appears to have been a cut/paste slip intended for a
$(DBLOAD) command above it.
peter [Wed, 24 Jul 1996 10:48:07 +0000 (10:48 +0000)]
Add sample rules for amd.host, mostly from the AMD docs, but tweaked to fit
the FreeBSD Makefile.yp structure by me. This allows you to have a single
amd map for all machines in a cluster.
In /etc/sysconfig, it would look something like:
amdflags="-p -a /net -c 1800 -l syslog /host amd.host"
When running 'rrestore foo', you get a segmentation fault because
the obsolete() function to convert dump-style args to getopt-style
args doesn't check to see that 'f' really has an argument following
the option string in argv[1].
Add a new, better mechanism for sticking packets onto ifqueues.
The old system had the misfeature that the only policy it could implement
was tail-drop; the new IF_ENQ_DROP macro/function makes it possible
to implement more sophisticated queueing policies on a system-wide
basis. No code actually uses this yet (although on my machine
I have converted the ethernet and (polled) loopback to use it).
Bring in some fixes from NetBSD and re-hack our syslogd to be option-compatible
with theirs (change the -I option to -s (but leave -I in for backwards compat.)
Also eliminate an make sane some magic numbers, and fix a small bug where we'd
send to an unopened socket.
I have added a new option -p to the mount command. This was
inspired by SunOS version of mount which uses option -p to
indicate that the mount information should be printed in fstab
format.
This is a neat way to create a new fstab file to use later when
one has modified the mount points or mount options or added or
removed mount some mount points. You just type
mount -p > /etc/fstab.new
and there is your new fstab file ready to be used though you
will of course have to add any necessary noauto flags manually.
[Committers note: This also seems to do the wrong thing for AMD
mounts, but in the more average case this is a nifty feature nonetheless
and one can always edit the bogus entries out]
Replace the annoying calls to Debugger() by panic()'s in the
#ifdef DIAGNOSTIC case, and a warning only otherwise.
People who want them to break into the debugger can always set the
breakpoint explicitly. The existing behaviour was a misfeature from
the beginning, in the (wrong) assumption that the SCSI controller must
always be of essential importance to the entire system.
Major cleanup of the timerX_{acquire,release} stuff. In particular,
make it more intelligible, improve the partially bogus locking, and
allow for a ``quick re-acquiration'' from a pending release of timer 0
that happened ``recently'', so it was not processed yet by clkintr().
This latter modification now finally allows to play XBoing over
pcaudio without losing sounds or getting complaints. ;-) (XBoing
opens/writes/closes the sound device all over the day.)
Check for, and disallow, duplicate tags in the "cvs import" command.
RCS cannot deal with duplicate tags; the extra one always becomes
inaccessible and useless.
This will prevent the common mistake of specifying the same name
for the vendor tag and the release tag. The FreeBSD CVS repository
already contains zillions of files with this error. We don't need
any more of them.
The previous change to alarm.3 to improve the wording
was still somewhat confusing and poorly worded. So I took
the alarm.3 man page from NetBSD, which looked much
better all around.
se [Fri, 19 Jul 1996 17:18:20 +0000 (17:18 +0000)]
Last night (local time :) I accidently checked in an intermediate version
of this patch, which had not actually been reviewed by Joerg or Paul!
(I'll better stop committing files after midnight ...)
I'm now commiting the latest code, which has been reported to work.
Minor correction to the previous commit message for this file:
The first PCI Lance in a system will get a name of lnc1, the second
will be known as lnc2 and so on. An arbitrary number of cards is
supported in a system ...
Add interrupt register set up with original register value
before attaching. Without this fix, 3c579(EISA) never make
any H/W inturrupt.
Reviewed by: "Justin T. Gibbs" <gibbs@freefall.freebsd.org>, nao@sbl.cl.nec.co.jp and owner-current on mailing list ;-)
Submitted by: amurai@spec.co.jp, nao@sbl.cl.nec.co.jp
se [Thu, 18 Jul 1996 22:03:47 +0000 (22:03 +0000)]
FINALLY: PCI support for the Lance Ethernet driver.
This code applies to several systems with integrated Ethernet
chip, for example from HP or Compaq. It should also support
PCI Ethernet cards based on the AMD PCI Lance chip.
This code has been reviewed (visually) by Paul Richards and
tested (using an ISA Lance board) by Joerg Wunsch.
Since the parameters to nearly each and every single function
had to be changed (generally from unit number to lnc_soft*),
there is some potential for buglets having crept in ...
BEWARE: If you had lnc0 configured to have the ISA probe find
your PCI Lance, then it should now be found by the PCI probe,
and should be automatically configured as pci1 (!!! note the "1").
projects / FreeBSD / FreeBSD.git / log