dillon [Sat, 9 Jun 2001 18:06:58 +0000 (18:06 +0000)]
Two fixes to the out-of-swap process termination code. First, start killing
processes a little earlier to avoid a deadlock. Second, when calculating
the 'largest process' do not just count RSS. Instead count the RSS + SWAP
used by the process. Without this the code tended to kill small
inconsequential processes like, oh, sshd, rather then one of the many
'eatmem 200MB' I run on a whim :-). This fix has been extensively tested on
-stable and somewhat tested on -current and will be MFCd in a few days.
imp [Sat, 9 Jun 2001 06:43:54 +0000 (06:43 +0000)]
Add PC9801-102 CBUS card to the list of plug and play devices. Some
mapping of irq 6 may be required to use that irq, but if so,
additional commits will follow.
Since we use getopt (in rev. 1.12), there's no need for the case's for
the individual options to increment argv and decrement argc. This
caused the -T option to swallow an extra argument.
PR: 27982
Submitted by: Samuel Greear <sgreear@vsni.com>
bmah [Fri, 8 Jun 2001 18:35:09 +0000 (18:35 +0000)]
MFS: Apply edits from 4-STABLE errata file to -CURRENT. Also add some
comments and notes to the effect that this file won't have any real
content until after 5.0-RELEASE.
bmah [Fri, 8 Jun 2001 15:31:54 +0000 (15:31 +0000)]
Forced commit:
The last commit to this file introduced some text that was "Copyright (C)
2000 BSDI". It was asserted (correctly) that permission should have been
sought from someone at BSDi / WindRiver first.
In response to a query on this point, jkh wrote:
> Hmmm. I'm not sure we ever asserted any particular copyright, but in
> any case it's certainly not clear who one would ask. Me? If so,
> I give my permission. :)
gallatin [Fri, 8 Jun 2001 13:38:02 +0000 (13:38 +0000)]
Sync up to v 1.9 of NetBSD's db_trace.c to get access to the
greatly improved traceback code from Ross Harvey. This code
requires the use of more traceback friendly temporary labels
at kernel entry points, hence the changes to exception.s and
asm.h
When changing an indirect route, kernel routing code allocates
a route to the gateway and caches it in the route structure.
It may happen (if the routing table is screwed) that the gateway
route is the same route as the one being modified, in which case
a kernel reports EDQUOT. Be more verbose about this:
phk [Fri, 8 Jun 2001 11:53:58 +0000 (11:53 +0000)]
The date for Australia day in
src/usr.bin/caldendar/calenders/calendar.holiday
is incorrect.
From looking through webcvs it seems like the error is in Open/NetBSD also.
PR: 27960
Submitted by: Harley Anderson <Warragul@selcon.com.au>
imp [Fri, 8 Jun 2001 07:16:56 +0000 (07:16 +0000)]
The TI-1031 is more like the TI-113x chips rather than the 12xx or
higher chips. Treat it as if it were a 113x. This is correct as far
as 16-bit cards go, at least how we're using it.
# It appears that my TI-1031 based pci card that YAMAMOTO shigeru-san gave
# me on my trip to Japan now works.
peter [Fri, 8 Jun 2001 05:24:21 +0000 (05:24 +0000)]
"Fix" the previous initial attempt at fixing TUNABLE_INT(). This time
around, use a common function for looking up and extracting the tunables
from the kernel environment. This saves duplicating the same function
over and over again. This way typically has an overhead of 8 bytes + the
path string, versus about 26 bytes + the path string.
Remove xref to blackhole(4). sysctl has nothing to do with
blackhole(4), except that blackhole(4) uses sysctl's. This xref
obviously isn't appropriate unless we want to xref all the other man
pages which mention sysctls, which we obviously don't (we may want to
list those sysctls, but that's another story).
lock sound device when adding/removing channels
implement setblocksize for vchans
don't panic when doing certain ioctls or aborting on a vchan
xmms now works with vchans
- Exit 1 if "add", "change", or "delete" operation fails.
PR: bin/12489
- Use inet_ntoa(3) where it should have been used. This
part of code simply wasn't converted to the "new" style
after the routename() function was converted from the
protocol-generic version to protocol-specific version
in CSRG revision 5.6.
- Renamed the ``proxy'' modifier of the -d flag to ``pub'',
to be consistent with the -s flag. Updated documentation
on what this modifier does.
- Added the ``only'' keyword to the -s and -S flags, that
could be used to created "proxy-only" published entries.
Previously, arp(8) created an entry of this type only
in the absence of the route to a destination.
In the "Message from Talk_Daemon" announcement, print the date as well
as the current time. It's nice to know whether the talk request you
see was sent just a few minutes ago (assuming you didn't hear the
bell), or if it's been decaying for days (weeks?).
peter [Thu, 7 Jun 2001 03:17:26 +0000 (03:17 +0000)]
Back out part of my previous commit. This was a last minute change
and I botched testing. This is a perfect example of how NOT to do
this sort of thing. :-(
Don't allow the user to change the subject; instead, always generate
it from the Synopsis field. There's no reason for the subject to be
different, since all that does is cause confusion. Users may get
confused because they may think the subject and synopsis are supposed
to be different, and developers may get confused because it may look
like there are two different problems.
peter [Wed, 6 Jun 2001 22:17:08 +0000 (22:17 +0000)]
Make the TUNABLE_*() macros look and behave more consistantly like the
SYSCTL_*() macros. TUNABLE_INT_DECL() was an odd name because it didn't
actually declare the int, which is what the name suggests it would do.
wpaul [Wed, 6 Jun 2001 22:16:23 +0000 (22:16 +0000)]
- Remember to set the 'extsts enable' bit in the CFG register to enable the
use of the extsts field in DMA descriptors. We need this to tell the chip
to calculate TCP/IP checksums in hardware on a per-packet basis.
- Fix the unions in DMA descriptor structures. Breakage on alpha led
me to realize I'd done it wrong the first time.
peter [Wed, 6 Jun 2001 22:00:03 +0000 (22:00 +0000)]
Activate the kthread mechanism for doing usb bus discovery. This means
that device add/remove will work without usbd running. usbd is still
used for execing stuff, but that is all now. Ideally it could be replaced
by a devd some day. Until now, usbd had to be running so that the
USB_DISCOVER ioctl could be called to walk the tree when an attachment
status change was noticed.
Among the changes:
- when a detach happens, remove any pending 'attach' messages or the system
suffers from whiplash from exec moused / kill moused loops if you do lots
of attach/detach and later start usbd.
- tweaks related to kthread differences
- disable the select handler for the old interface (never return success).
I have not removed it yet or old usbd's will abort. That can get removed
later once usbd is cleaned up and things have stabilized for a few weeks.
- get Giant in the kthread.
- a couple of minor potential bug fixes (usb_nevents vs malloc failure etc)
jesper [Wed, 6 Jun 2001 19:41:51 +0000 (19:41 +0000)]
Silby's take one on increasing FreeBSD's resistance to SYN floods:
One way we can reduce the amount of traffic we send in response to a SYN
flood is to eliminate the RST we send when removing a connection from
the listen queue. Since we are being flooded, we can assume that the
majority of connections in the queue are bogus. Our RST is unwanted
by these hosts, just as our SYN-ACK was. Genuine connection attempts
will result in hosts responding to our SYN-ACK with an ACK packet. We
will automatically return a RST response to their ACK when it gets to us
if the connection has been dropped, so the early RST doesn't serve the
genuine class of connections much. In summary, we can reduce the number
of packets we send by a factor of two without any loss in functionality
by ensuring that RST packets are not sent when dropping a connection
from the listen queue.
wpaul [Wed, 6 Jun 2001 19:16:02 +0000 (19:16 +0000)]
Use LGE_INC() macro to increment tx producer index in lge_encap().
Disable the extra TCP/UCP checksum checking in lge_rxeof() since it
doesn't appear to actually work as advertised.
jlemon [Wed, 6 Jun 2001 17:47:13 +0000 (17:47 +0000)]
o If the file being tailed is moved back to the original location,
resume using kqueue to monitor it, if possible.
o auto-clear the read filter on EOF; this is needed when tailing a FIFO.
bde [Wed, 6 Jun 2001 06:58:13 +0000 (06:58 +0000)]
Fixed missing parentheses in the definition of KTR_COMPILE. KTR_COMPILE
is usually (always?) used in expressions like (KTR_COMPILE & KTR_FOO).
Defining it as KTR_INTR|KTR_PROC gave the wrong value in approximately
8497 places according to error output for compiling LINT.