martin hudec wrote:
amber# dd if=/dev/zero of=/usr/testfile bs=64k count=2000
131072000 bytes transferred in 2.546718 secs (51467023 bytes/sec)
These are not very usefull tests of raw disk bandwith because they use
the filesystem (and are hence subject to the vagiaries of block
Rob wrote:
You'd better cite your source and / or reasoning, as ~3*10^8m/s =is=
the
accepted constant speed of light in vacuum.
It's deeper than that. The second and the meter are both defined in
terms of wavelengths of light, which (as a consequence) fixes the speed
of light _by definition_,
Thanks all for the enlightening response. There is one bit I still don't get:
How do I tell the kernel which port to look for the PPS signal on with
PPS_SYNC?
___
[EMAIL PROTECTED] mailing list
Would you please point out some cases when HTT with machdep.hlt_logical_cpus=
0
causes worse performance than machdep.hlt_logical_cpus=1?
Typically, running a single, large, CPU-bound process, especially
FPU-instensive simulations.
___
[EMAIL
Really sounds like there is another machine on the same subnet with the same
(new) IP address, and they are having ARP duels. Sometimes you get machine A,
sometimes machine B. When you get A, it works, when you get B, there is no
HTTP/IMAP/ etc daemons listening and the host key is wrong.
First thing to check is rDNS - make sure the box can resolve the IP addresses
used to connect to it. That's often why ssh/telnet/ftp/etc hang up - waiting
for rDNS results to write log records.
___
[EMAIL PROTECTED] mailing list
Any module compiled without PAE will probably panic a PAE kernel. (See recent
threads about pnc and rl ethernet drivers). Given the fragility of the module
build process, a good rule of thumb might be never use modules in a PAE
kernel (or an INVARIENTS kernel, or).
At the very least,
However we have a situation where if I set MAXDSIZ to 2048 or above then
things break, so FreeBSD right now has an effectivce limit of 2GB per
process.
Is this to be considered a bug or a feature?
It orta work above 2Gb, and has been known to work to about 3Gb. The fact that
tcsh is
Has anyone got hints on using KVM switches with XFree86? I'm using a wheel
mouse and moused, but something gets out of whack when I switch away then back
to X. Next time I touch the mouse, all sorts of weird things happen for a
fraction of a second, then the mouse starts working OK. This is
DUMP: 101.44% done, finished in 0:-1
Id' bet that's caused by activity on the FS between when dump starts (and makes
it's estimate of how much to dump) and it finishes.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-stable in the body of the message
I had, same error occured. Even with -O0, which turns off any
optimization whatsoever.
Umm, it said 'revert to -O', not 'without optimization'. And for a good
reason. Compiling with anything but the system compiler and with optimisation
set to anything except -O is _not supported_.
To
You have chosen to maintain systems which stretch FreeBSD to its limits
and uncover bugs lurking in the code. This is great. But you cannot do
so on the one hand and refuse to face the administrative work on the
other hand.
And this is not a FreeBSD problem either - if you are doing
David O'Brien merged the relevant changes yesterday. If you think
Looks like I CVSUP'd before the patch made it to the mirror I was using. I've
just CVSUPed again and got the changes.
Thanks!
It's a huuuge pity this didn't happen before the change to the Ports
makefiles, tho...
To
I have a DLINK 520 PCI card on a FreeBSD server (running 4.6-stable from around
20 July) and a DLink 650 PCMCIA card in a WinMe laptop.
I can get these systems to communicate in Ad-Hoc mode with and without WEP,
and in hostap mode on the FreeBSD end without WEP, but I cannot seem to get the
9660: /dev/acd0c: Device not configured
Did you remake the devices? (cd /dev sh MAKEDEV acd0)
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-stable in the body of the message
I have a system with buildworld/installworld last done on about 1 June. It's
a pretty plain-jayne HP P4 with an on-board fxp0 interface.
Did a cvsup, buildworld and buildkernel and installkernel on RELENG_4 from
about 1900 UTC on Satuday 10th. Rebooted before doing installworld (just to
make
Has anyone tried these patches on 3.5-STABLE? I applied the patch
and it only errored on the name6.c portion. I'm building world with
it but would like to hear from someone who has used this on such an
old system.
I've tried it on 3.2-Release. The name6.c stuff is from IPv6 which is not
This line in rc.conf used to work in 4.5-STABLE:
static_routes=-net 192.168.1.0 10.1.1.1 -netmask 255.255.255.0
I can't imagine how. That is never how the static_routes rc.conf variable
has worked.
see man rc.conf, what you want is something like
static_routes=mynet
route_mynet=-net
hand carved round table with metal chain link in the middle
Every time you think you've seen the worst the net can offer, along comes
another AOLer
[Hey, the list was [EMAIL PROTECTED]!]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-stable in the body of the
You can always boot off of boot floppies by making kern.flp and mfsroot.flp
images. It's just that the kernel running during install might be stripped
down and thus might not have support for some newer devices.
Which hardly seems fatal, as any system with those newer devices orta have a
Yes, but then who do you target the ISO at? I'm trying to judge how widely
used the older machines are and if we should still use boot.flp on the ISO's
to
accomodate them.
It depends on the nature and ubiquity of the newer devices that get dropped
off kern.flp. If we get to the stage
What I'd like to know is if anyone has figured out a way to do a
g++ equivalent to Microsoft's precompiled headers.
Not really needed for gcc. The preprocessor has special-case logic to
recognise header files of the form
/* optional comments */
#ifndef SYMBOL
Yes, it's a hack,
So send in the clone()s.
:
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-stable in the body of the message
Cheers for any light anyone can shed,
As others have said:
- check login.conf
- force a salt like '$1$$'
also:
check /etc/auth.conf for a crypt_default line.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-stable in the body of the message
[Cc'd to stable and bugs. Please reply to bugs only]
The driver for Stallion multiport cards in 4-Stable is quite old. It doesn't
support many of the newer cards, in particular the EasyIO-PCI card, nor any
card manufactured since about 1998, which have a newer and unsupported UART
chip.
The
There is an interesting difference in malloc. It defines the symbol
free and if you use that symbol in your program then it tends to
write over some part of your program. By changing the symbol in my
program (an int) to a different name, the problem goes away and it
works properly. I
I have been experiencing some strange behavior with netstat.
First check: are you sure that your installed userland and running kernel are
built from the same set of sources? IE you have built and installed a kernel
and rebooted when you last diod a make world?
To Unsubscribe: send
I've got a weird little problem... trying to play wavefiles via
'play' (usr/ports/audio/play) results in the following (SB Live):
play: /dev/dsp: Invalid argument
Hmm, I'm tracking down a similar problem with waveplay (/usr/ports/audio/
waveplay). The first time this is run after a reboot
While it is technically free, you won't be able to use it in day-to-day
operations. It is reserved for the superuser. I believe it is so that if
something should flood the filesystem, the superuser still has space to
operate on the partition.
Partially. It is also a performace hack - the
Jan 11 00:56:11 culverk /kernel: proc: table is full
Jan 11 00:56:14 culverk last message repeated 1022 times
This says you have run out of process slots in the kernel - running too many
jobs. Either run fewer processes (e.g. use "make -j4" rather than "make -j8")
or build a new kernel with
Can't you just increase the kern.maxproc sysctl?
Nope, it's readonly, because various kernel internal hash tables are sized on
it. See src/sys/kern/kern_{proc,resource,mib}.c
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
hrmm, I tried that, same error (that was the first thing I tried actually)
Hmm, it may be that the particular program you are building has some kind of
error (in build scripts, or makefile, or whatever) that is causing an infinite
loop somewhere. Examine the last N lines of output before the
I thought about what I had read
regarding MFS... that I could make a filesystem in RAM. This being the
case, couldn't it also be shared via samba?
Samba interracts with the local filesystem via the normal read()/write()/
open()/close() system calls so doesn't care what the underlying
Are there any limits for size/quantity of partitions at freebsd
slice?
4 slices per disk, 8 partitions per slice (by convention, partition "c" covers
the whole disk so usually only 7 usable partitions).
AFAIK FreeBSD cannot use the so-called "extended partitions" (i.e.
I've just done another cvsup and it looks like rsa_eay.c has been added back to
the cvs-crypto collection. It looks like it's all fixed; I'm running a
buildworld so I should have a definitive answer sometime this afternoon.
Thanks all.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with
For the record/mail-list archives:
I sumbitted this as a PR (http://www.freebsd.org/cgi/query-pr.cgi?pr=19553)
and David Malone pointed out that using KLDs compiled without INVARIANTS
in a kernel compiled with INVARIANTS will cause this. See the PR for details.
To Unsubscribe: send mail to
I recently started to track 4.X Stable. How often does one need
to run mergemaster, or even better how does one even know that
it needs to be run?
Every time you run "make installworld" or "make world". It's no big chore, I
usually do it while I'm waiting for the new kernel to compile.
Copy it to /etc/rc.conf and change YES to NO
If by this you mean "copy /etc/defaults/rc.conf to /etc/rc.conf" - this is not
a good idea. The system will not boot ("out of file descriptors" from
infinitely recursing rc.conf scripts.)
just add the relevent line ('sendmail_enable="NO"' in this
USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND
root 156 0.0 0.2 262968 256 ?? Is Mon05PM 0:00.21 rpc.statd
rpc.statd mmap()'s a 256Mb chunk of anon memory but never uses most of it.
This is 100% normal.
[Is this in the FAQ?!?]
To Unsubscribe:
today I fetched the fresh 3.2 stable code using cvsup, starting make
buildworld. It failed. Since I´m out of my depth here, maybe you could
help.
You missed the HEADS UP messages posted by John Polstra on -stable on 03 Jul?
[you do follow the stable mailinglist, don't you!?]
The softupdates
40 matches
Mail list logo
