| Dave Frantz | FreeBSD 7.0-current Locks During Device Probe--Solved
I finally figured out what was causing FreeBSD 7.0-current to crash on my machine during the device probe. It turns out that if options CPUFREQ is compiled in, the system will not boot. However, it can be safely kldloaded in later.
I'm not entirely sure why my system dislikes CPUFREQ, but I can live without it. If you guys have any questions or anything, just let me know.
-Dave
---------------------------------
Building a website is a piece of cake.
Yahoo! Small Business gives you ...
| Oct 8, 8:21 pm 2007 |
| Randall Stewart | AMD64 panic on boot
Hi all:
I just updated to the latest.. and I am seeing a crash on
boot up of my 8-core Xeon machines (running in AMD mode).
its a
pagefault UA= 0x2b
sysctl_handle_int+0x27
syscall+0x254
Xfast_syscall+0xab
IP= 0xffffffff8048e7d7
I can't get a core unfortunately.. I get a core device
unavailable..
I have one of the machines sitting in a DDB> prompt
and its twin sitting on the old kernel..
Let me know how I can help.. and in the mean time I will poke
around in kgdb on the ...
| Oct 8, 7:11 pm 2007 |
| Randall Stewart | Re: AMD64 panic on boot
Hmm.. it appears to be in..
0xffffffff8048e2e7 is in sysctl_handle_int
(../../../kern/kern_sysctl.c:816).
811 */
812 if (arg1)
813 tmpout = *(int *)arg1;
814 else
815 tmpout = arg2;
816 error = SYSCTL_OUT(req, &tmpout, sizeof(int));
817
818 if (error || !req->newptr)
819 return (error);
So its trying to copy out and faulting.. not sure what..
Let me know if there ...
| Oct 8, 7:14 pm 2007 |
| Petr Holub |
| Oct 8, 12:32 pm 2007 |
| Petr Holub | RE: panic on boot
Despite you couldn't guarantee it, it really helped and it boots now.
I've used the binary loader available from ~kib site and just replaced
it on the original ISO: mdcondifig -> mount ->
tar cf - | (cd elswhere && tar xvpf -) -> replace the loader -> mkisomages.sh
Thanks a lot,
Petr
_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to ...
| Oct 8, 2:11 pm 2007 |
| Petr Holub | RE: panic on boot
After booting the machine with this patch, it works like a charm,
looks fairly stable, Myrinet 10GE card up and running. So I
wonder: would it be possible to include this realbtx into 7.0-RELEASE?
It seems rather important, as many machines are now being
installed off USB CD drives and USB keys...
Thanks a lot,
Petr
_______________________________________________
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any ...
| Oct 8, 4:47 pm 2007 |
| Petr Holub | panic on boot
Hi,
I was about to install September -CURRENT snapshot on a machine
with 10GE card. When attempting to install from USB CDROM (it's
a 1U machine without, I get something like a kernel panic
with a stacktrace running indefinitely and very fast. I must be
in a very early stage as I was unable to break it anyhow. When
I tried to break in using serial console (setup using BIOS), the
last message I get in the console is
CD Loader 1.2
and then nothing, but the screen keeps running with sth like ...
| Oct 8, 11:24 am 2007 |
| Kostik Belousov |
| Oct 8, 12:34 pm 2007 |
| Timur I. Bakeyev |
| Oct 8, 7:01 am 2007 |
| Michael Hanselmann |
| Oct 8, 7:04 am 2007 |
| Pawel Jakub Dawidek |
| Oct 8, 6:38 am 2007 |
| Dag-Erling Smørgrav |
| Oct 8, 8:28 am 2007 |
| Bernd Walter |
| Oct 8, 6:44 am 2007 |
| Dag-Erling Smørgrav |
| Oct 8, 7:42 am 2007 |
| Dag-Erling Smørgrav | Re: ZFS kmem_map too small.
Speaking of which, I tried setting vm.kmem_size to 2G on a C2D system
with 4 GB RAM, but it simply panics:
OK set vm.kmem_size=3D2G
OK boot -s
GDB: no debug ports present
KDB: debugger backends: ddb
KDB: current backend: ddb
Copyright (c) 1992-2007 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD ...
| Oct 8, 6:30 am 2007 |
| Bjoern A. Zeeb |
| Oct 8, 12:04 am 2007 |
| Pyun YongHyeon | CFT: vr(4)
It seems that there had been several stability issues in vr(4).
Here is mimimal patch that make vr(4) work reliably under heavy
network loads. The patch does the following:
- Always check writability of mbuf before padding and make a
writable copy of the mbuf if mbuf is marked as read-only.
- Before padding is done check remaining bytes such that it can
safely extend buffer size of the mbuf.
- Before padding always check the return value of m_defrag(9).
- Zero out pad space to avoid ...
| Oct 7, 11:17 pm 2007 |
| Pyun YongHyeon | Re: CFT: vr(4)
On Mon, Oct 08, 2007 at 07:04:32AM +0000, Bjoern A. Zeeb wrote:
> On Mon, 8 Oct 2007, Pyun YongHyeon wrote:
>
> >It seems that there had been several stability issues in vr(4).
> >Here is mimimal patch that make vr(4) work reliably under heavy
> >network loads. The patch does the following:
> >- Always check writability of mbuf before padding and make a
> > writable copy of the mbuf if mbuf is marked as read-only.
> >- Before padding is done check remaining bytes such that it can
> > ...
| Oct 8, 12:29 am 2007 |
| Peter Jeremy | Re: bgfsck strikes again
I have hw.ata.wc=3D"0" and judging by the disk write performance, caching
really is off.
--=20
Peter Jeremy
| Oct 8, 12:32 am 2007 |
| Roman Divacky |
| Oct 8, 10:25 am 2007 |
| Kostik Belousov | Re: pad argument for mmap/lseek/etc.
You should lookup the corresponding commit that obviously touched
syscalls.master (seems to be rev. 1.232), and get the syscalls list from
there. Then, you should copy the old structure definitions to freebsd6_XXX
functions, and delete the paddings inplace.
The interesting question is whether the _SYS_SYSPROTO_H and absense of the
sys/sysproto.h supported. Or, shall we remove these #ifdef blocks instead
of fixing them in whole kernel.
| Oct 8, 8:31 am 2007 |
| Alexander Leidinger |
| Oct 8, 12:18 am 2007 |
| Stefan Esser |
| Oct 8, 3:02 am 2007 |
| Marcel Moolenaar | Re: cosmetics: bsdlabel fstype for ZFS partitions?
This is technically speaking not a good idea because the UUID may
be in use already. In this case, given that the MAC address is in
one of my machines, it can only be a UUID I can possibly generate
(legally) and I can tell you that the UUID is not in use :-)
If no-one has created their own UUID for ZFS, I'll re-use Craig's
one so that he doesn't have to change his GPT. Any objections for
using Craig's synthesized UUID for ZFS file systems?
--
Marcel ...
| Oct 8, 10:02 am 2007 |
| Craig Boston | Re: cosmetics: bsdlabel fstype for ZFS partitions?
I tend to abuse 'vinum' for this, and I suspect many use 4.2BSD just out
of paranoia of 'unused' being accidentally overwritten.
I also have ZFS on a few GPT-labeled disks and had a hard time deciding
what GUID to use. I couldn't find a good reference for what Sun uses
for ZFS disks, so I ended up making up a bogus one under FreeBSD's
range.
516e7cba-6ecf-11d6-8ff8-00022d09712b
Craig
_______________________________________________
freebsd-current@freebsd.org mailing ...
| Oct 8, 9:42 am 2007 |
| Jung-uk Kim |
| Oct 8, 10:58 am 2007 |
| Ralf S. Engelschall | Re: cosmetics: bsdlabel fstype for ZFS partitions?
Well, this UUID is...
$ uuid -d 516e7cba-6ecf-11d6-8ff8-00022d09712b
encode: STR: 516e7cba-6ecf-11d6-8ff8-00022d09712b
SIV: 108241150102539780167861756444952588587
decode: variant: DCE 1.1, ISO/IEC 11578:1996
version: 1 (time and node based)
content: time: 2002-05-24 04:32:55.748933.8 UTC
clock: 4088 (usually random)
Well, if a really synthesized UUID for ZFS under FreeBSD is wished, I
would propose to generate a UUID v3 for this purpose ...
| Oct 8, 10:28 am 2007 |
| Craig Boston | Re: ZFS kmem_map too small.
I'll bet it's this change that provokes the problem; as even when
manually tuning the ARC size to 1/2 kmem_size or lower I still sometimes
get panics. Probably what's happening is kmem usage gets high from
other things, and when there's a sudden spike zfs can't react fast
enough and shrink the ARC. At 50% it acts more conservatively so
there's more memory available for burst usage.
I noticed that some of the time, on my mostly-stable system, the panic
happens when the nvidia driver is ...
| Oct 8, 9:27 am 2007 |
| Craig Boston |
| Oct 8, 11:14 am 2007 |
| Claus Guttesen |
| Oct 8, 10:18 am 2007 |
| Craig Boston | Re: ZFS kmem_map too small.
Sorry to spam the thread again, but one thing I noticed is that zfs does
an awful lot of allocations of various sizes:
$ uptime
11:30AM up 2 days, 18:06, 11 users, load averages: 0.05, 0.12, 0.15
(was idle most of the weekend)
$ vmstat -m | grep solaris
solaris 83176 341456K - 53216607 16,32,64,128,256,512,1024,2048,4096
I'm not completely up-to-date on what algorithm the kernel allocator is
using these days, but is it possible that kernel memory is getting
fragmented by ...
| Oct 8, 9:33 am 2007 |
| Micha Mutschler | Re: ZFS kmem_map too small.
Hi!
I was doing a ports rsync from a zfs to /usr/ports (ufs). That worked
fine. The crash happend as I started to copy (parallel to the rsync) a
huge file from a nfs server to the zfs.
For 512 MB RAM I've set:
kern.maxvnodes="25000"
vfs.zfs.prefetch_disable="1"
vfs.zfs.arc_max="52428800"
vm.kmem_size_max="402653184"
vfs.zfs.zil_disable="1"
[root@filer ~]# zpool status
pool: xport
state: ONLINE
scrub: none requested
config:
NAME STATE READ WRITE CKSUM
...
| Oct 8, 5:26 am 2007 |
| Pawel Jakub Dawidek | Re: ZFS kmem_map too small.
Here are some updates:
I was able to reproduce the panic by rsyncing big files and trying
bonnie++ test suggested in this thread.
Can you guys retry with this patch:
http://people.freebsd.org/~pjd/patches/vm_kern.c.2.patch
It's a hack, yes, but allows to mitigate the problem quite well. I'm
looking for a solution that can be used for 7.0 before we find a better
fix.
BTW. To use ZFS you _must_ increase vm.kmem_size/vm.kmem_size_max.
If you have the problem discussed here and you're ...
| Oct 8, 5:15 am 2007 |
| Guy Brand | Re: ZFS kmem_map too small.
Hi,
On a Lenovo X61s from yesterday (FreeBSD 7.0-CURRENT #1: Sun Oct 7
21:55:14 CEST 2007) with
hw.model: Intel(R) Core(TM)2 Duo CPU L7500 @ 1.60GHz
hw.physmem: 2091008000
hw.machine_arch: i386
hw.realmem: 2104164352
and a single zpool containing ad4s2 slice (all FS except /). I tuned
vm.kmem_size and ran ad vitam:
- vt0: rsync /usr/src /tmp/. && rm -rf /tmp/src
- vt1: dd if=/dev/zero of=/tmp/file bs=1k count=1000000
The kernel panics until ...
| Oct 8, 1:27 pm 2007 |
| Ivan Voras | Re: ZFS kmem_map too small.
Is kmem / arc allocation static or dynamic? I.e. the "regular" file
cache can use up all (or at least big chunk of) memory and then return
it to applications when they need it. I suppose this is not the case
with ZFS?
| Oct 8, 5:47 am 2007 |
| Stephan Uphoff |
| Oct 8, 8:49 am 2007 |
| Marc Fonvieille |
| Oct 8, 7:31 am 2007 |
| Marc Fonvieille | Re: ULE/yielding patch for testing.
Without SMP and apic, the number of context switches is lower:
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr ad4 cd0 in sy cs us sy id
0 0 0 501656 89424 0 0 0 0 0 0 0 0 1266 3532 1511 8 7 86
0 0 0 501656 89424 0 0 0 0 0 0 0 0 1252 8127 964 11 6 83
1 0 0 501656 89424 0 0 0 0 0 0 0 0 1253 1514 973 2 7 92
3 0 0 501656 89424 2647 0 0 0 ...
| Oct 8, 7:27 am 2007 |
| Andrey V. Elsukov |
| Oct 8, 5:06 am 2007 |
| Andrey V. Elsukov |
| Oct 8, 8:55 pm 2007 |
| Stefan Esser |
| Oct 8, 6:05 am 2007 |
| TooMany Secrets | Re: Need motherboard for home fileserver
MjAwNy8xMC84LCBTdGVmYW4gRXNzZXIgPHNlQGZyZWVic2Qub3JnPjoKPiBXZWxsLCB5b3UgbWF5
IHdhbnQgdG8gdGFrZSBhIGxvb2sgYXQgdGhlIEFiaXQgQU4tTTIgKG9yIC1NMkhELCBpZiB5b3UK
PiBhcmUgaW50ZXJlc3RlZCBpbiBmaXJld2lyZSBzdXBwb3J0LCBvciBwb3NzaWJseSBpbiBIRE1J
L0hEQ1ApLiBBbGwKPiBmdW5jdGlvbnMgYXBwZWFyIHRvIGJlIHdlbGwgc3VwcG9ydGVkIChkaXNr
LCB2aWRlbywgc291bmQsIGludGVncmF0ZWQKPiBQQ0llIEdpZ2FiaXQgRXRoZXJuZXQpLiBDb21i
aW5lZCB3aXRoIGFuICJBcmN0aWMgQ29vbGluZyBGcmVlemVyIDY0IFBybwo+IFBXTSIgaXQgZ2l2
ZXMgYSBjaGVhcCwgY29tcGFjdCwgYW5kIHF1aWV0 ...
| Oct 8, 4:28 am 2007 |
| Stefan Esser | Re: Need motherboard for home fileserver
What made you come to that conclusion???
I use an Abit AN-M2HD (and I know I'm not alone with that main-board)
which is based on the nForce 7050/630a and works very well. I had one
issue with an LCD connected to the HDMI port (a HDMI-to-DVI cable is
bundled) and since it happened outside X11 (i.e., with a text console),
it might have been a bogus enabling of HDCP. (Hard to describe, the
screen showed a picture like an unsynchronized TV, but steady; only
The video integrated into the ...
| Oct 8, 2:56 am 2007 |
| Tijl Coosemans |
| Oct 8, 3:35 am 2007 |
| Eric Schuele |
| Oct 7, 10:41 pm 2007 |
| Danny Braniss | Re: iscsi port in /etc/services?
> Hey guys, my first post here, so don't shoot me. (Don't know if this is=
As far as I can tell, iscsi-target does NOT use /etc/services, it specifi=
cally
requests port 3260, unless it's changed by the the -p flag.
So your problem is somewhere in the configuration file.
danny
PS: the nice thing about FreeBSD is that you have the sources, of course =
that
needs some basic computer language knowledge :-)
_______________________________________________
freebsd-current@freebsd.org ...
| Oct 8, 7:08 am 2007 |
