I've been trying to get humfs to work, but I can't find it in 2.6.25. How
do you actually enable it in the kernel compile? I enabled hostfs, but
that doesn't seem to help:
[EMAIL PROTECTED]:~# mount none -t humfs -o /USERMODE /mnt
mount: unknown filesystem type 'humfs'
[EMAIL PROTECTED]
I use NFS for having the guests mount separate directories on the host,
then I rsync to these from the UMLs. Works for me, and is easily cron:able
from inside the UMLs. This was convenient since my host have a directory
with shared resources for the UMLs anyway, so NFS was already set up.
If yo
Host is in all tests
Dual-opteron amd64
Kernel 2.6.14-skas3-v9-pre7
Gentoo
Glibc 2.3.5-r2
Gcc 3.4.4-r1
Guest kernel is 2.6.16-rc5, with patches from uml homepage as of february
24, minus user-desc.
Guest kernel was compiled in 64-bit:
make mrproper ARCH=um
make defconfig ARCH=u
an_fs] Error 2
// Joel
On Thu, 9 Feb 2006, Blaisorblade wrote:
On Wednesday 08 February 2006 23:30, Joel Palmius wrote:
Well, after I updated I got another compile error at the same place:
CC arch/um/os-Linux/start_up.o
CC arch/um/os-Linux/time.o
CC arch/um/os-Linux/
l
On Wed, 8 Feb 2006, Jeff Dike wrote:
On Wed, Feb 08, 2006 at 12:44:22PM +0100, Joel Palmius wrote:
arch/um/os-Linux/tls.c: In function `os_set_thread_area':
arch/um/os-Linux/tls.c:20: error: `modify_ldt_t' undeclared (first use in
this function)
I had a fix for this, but I forgot
I've seen similar errors reported to and from on the lists, but none of
the suggestions seem to help. After applying patches as of february 06,
kernel with defconfig refuses to compile:
make defconfig ARCH=um
make linux ARCH=um
...
CC arch/um/os-Linux/start_up.o
CC arch/um/os-Linu
3
in 32bit with a slackware root_fs, I did not test this with 2.6.15 though)
This is an improvement, since I couldn't run unmodified stage3's in skas0
before due to the glibc tls issues. 2.6.14-SKAS0-32bit works fine in my
production environment anyway, so I'm pretty happy. I jus
root_fs is an unmodified gentoo 2005.1 athlon-xp stage3. Host is amd64
running 2.6.14-skas3-v9-pre7.
Host glibc is 2.3.5-r2. Host gcc is 3.4.4-r1. 32-bit chroot (see 32-bit
test below) glibc is 2.3.4.20041102-r1, 32-bit chroot's gcc is
3.3.5.20050130-r1. Root_fs is same image as chroot.
Gues
On Tue, 20 Dec 2005, Blaisorblade wrote:
So we need an utility doing just the setup and the writeout. Should be easy to
do, but I must learn the complete COW API. I will dismiss this task if I see
(as it seems) that there are no existing COW files.
Personally I stopped trusting COW files once
On Thu, 15 Dec 2005, Blaisorblade wrote:
uml_moo says "arithmetical error". I'm only guessing this is because the
root/cow files are too large..? Or is it a 64-bit host error?
...
Just for confirmation - I looked in your mails but I found a missing detail: I
expect you used a 32-bit kernel to
On Sun, 11 Dec 2005, Blaisorblade wrote:
Also, the iowait is on the host or on the guest? I fear on the host, but I'd
hope not.
No, the wait was on the guest. The guest could have 100% wait while the
host only had a few percent, leading me to think that stuff simply got
queued up somewhere a
ay 08 December 2005 08:06, Joel Palmius wrote:
Are there any general recommendations for making throughput on UBDs more
efficient? Such as kernel settings, recommended filesystems..?
Doing an "emerge sync" inside an uml raises iowait to 100% and load to 7,
making the machine more or le
Are there any general recommendations for making throughput on UBDs more
efficient? Such as kernel settings, recommended filesystems..?
Doing an "emerge sync" inside an uml raises iowait to 100% and load to 7,
making the machine more or less inaccessible. Just wondering if it's me
having bad s
While uml now mostly works, and keeps working happily as soon as it has
booted, sometimes it refuses to boot. Seems pretty random to me. Sometimes
it boots, sometimes it doesn't. If I retry a couple of times it usually
decides to boot after a while. Strange thing is it hangs before even
trying
-xp stage3. Haven't
seen any soft lockup at all so far.
// Joel
On Sat, 26 Nov 2005, Joel Palmius wrote:
As far as I can see, the machine now hangs (busy hang) indefinitely instead.
Pity.. things worked so well in skas0 on my opteron machine. I'll try to
reinstall this (athlon64
y did not
help all that much. :-)
// Joel
On Fri, 25 Nov 2005, Jeff Dike wrote:
On Thu, Nov 24, 2005 at 05:43:38PM +0100, Joel Palmius wrote:
What more?
Disable CONFIG_DETECT_SOFTLOCKUP until we figure out what's happening.
e problem still happens.
What more?
// Joel
On Thu, 24 Nov 2005, Joel Palmius wrote:
It didn't.
Changing to 2.6.13.4-bs5 solved the problem though. I haven't tested a lot
yet, but so far I haven't seen any soft lockups.
// Joel
On Thu, 24 Nov 2005, Joel Palmius wrote:
S
It didn't.
Changing to 2.6.13.4-bs5 solved the problem though. I haven't tested a lot
yet, but so far I haven't seen any soft lockups.
// Joel
On Thu, 24 Nov 2005, Joel Palmius wrote:
So what do I do? Might enabling pre-emption help?
// Joel
On Thu, 24 Nov 2005,
So what do I do? Might enabling pre-emption help?
// Joel
On Thu, 24 Nov 2005, Jeff Dike wrote:
On Wed, Nov 23, 2005 at 07:53:23PM +0100, Joel Palmius wrote:
The below happens more often than not (but not always). The host is
athlon64 running in 32-bit mode. Host kernel is 2.6.14-skas3-v9
The below happens more often than not (but not always). The host is
athlon64 running in 32-bit mode. Host kernel is 2.6.14-skas3-v9-pre7.
Guest kernel is 2.6.14-bs1. Same thing happens in skas0 and skas3.
When the machine boots correctly it seems stable though.
If I don't try to enable swap (i
I'm planning to set up and use a server with UMLs for more serious
production use. The server I have is a dual-core athlon64 with some disk
and 4gb of memory. Now question is how to configure it in order to get the
most out of the UMLs.
The first major decision is whether to run in 32 or 64 bi
nd re-test 2.6.15-rc some time later.
// Joel
On Sun, 20 Nov 2005, Blaisorblade wrote:
On Saturday 19 November 2005 14:11, Joel Palmius wrote:
These tests are running on the following:
* Dual opteron running in amd64 mode
* Host kernel 2.6.14-skas3-v9-pre7
* Host OS is g
These tests are running on the following:
* Dual opteron running in amd64 mode
* Host kernel 2.6.14-skas3-v9-pre7
* Host OS is gentoo
* Guest OS is vanilla stage3-athlonxp
(suite 1)
* Guest kernel 2.6.15-rc1 + jeff's patches as of nov17
* Guest kernel config is defconfig + enabled
The patches on
http://user-mode-linux.sourceforge.net/patches.html
.. are they only for guest or are some for host?
// Joel
---
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today
Register for a JBoss Training Course. F
How about dropping COW and ubds altogether and run root on NFS instead?
Works fine for me.
// Joel
On Wed, 16 Nov 2005, Dan Lund wrote:
Hi,
I'm centralizing my image(s) to an NFS share currently to save space
on my 3 UML hosts.
This is the first time I've used NFS as a partition to store b
Is it possible to pass or change xterm settings somehow? For example I'd
like a bigger font and white text on black background in the xterms uml
pops up for me.
// Joel
---
SF.Net email is sponsored by:
Tame your development challenges wi
For me it worked to build the guest system from scratch as long as glibc
was built with "--without-tls". I don't think you need to rebuild the
host.
I've built a root_fs for i686, based on gentoo, with tls disabled. Works
for me. If you're interested it's at
http://gathering.itm.mh.se/~joe
I've now played around a bit more with different ways of booting UML, and
I've successfully booted with root both on hostfs and on nfs.
For practical reasons, the hostfs approach seems to require less arcane
configuration options, but on the other hand, the nfs approach doesn't
require running
esday 19 October 2005 11:32, Joel Palmius wrote:
No, V3 is in use since long time - a lot before 2.4.24 surely.
Then there's another (minor) bug in uml_moo :-)
[EMAIL PROTECTED] /export/virtual/usermode/machine.test $ ./uml_moo
...
./uml_moo supports version 1 and 2 COW files.
Yes, saw
No, V3 is in use since long time - a lot before 2.4.24 surely.
Then there's another (minor) bug in uml_moo :-)
[EMAIL PROTECTED] /export/virtual/usermode/machine.test $ ./uml_moo
...
./uml_moo supports version 1 and 2 COW files.
[EMAIL PROTECTED] /export/virtual/usermode/machine.test $
Is thi
ero thingie?
This is how far my C knowledge runs atm. I actually programmed some C like
seven years ago, but that obviously was a bit too long ago. :-)
// Joel
On Tue, 18 Oct 2005, Blaisorblade wrote:
On Tuesday 18 October 2005 00:29, Joel Palmius wrote:
It was altogether too long since
pread(3, "OOOM\0\0\0\3CR\232%\0\0\0\1\0\20\0\0\0\0\2\0\0\0\20\0\0"..., 4136, 0)
= 4136
--- SIGFPE (Floating point exception) @ 0 (0) ---
+++ killed by SIGFPE +++
This didn't tell me a whole lot either. :-)
I hope any of this was more intelligible to someone else than it was
[EMAIL PROTECTED] /export/virtual/usermode/machine.master $ ls -lh
totalt 4,3G
-rwxr-xr-x 1 joepal users 671 16 okt 16.09 config.pl*
lrwxrwxrwx 1 joepal users 12 16 okt 16.09 master.uml -> ../lib/linux*
-rw-r--r-- 1 joepal users 4,1G 14 okt 21.02 root_fs
-rw-r--r-- 1 joepal users 4,1G 16 ok
On Sat, 15 Oct 2005, Blaisorblade wrote:
One question - -march=athlon64 is also supported by GCC (even if I don't know
how much it helps).
Yup. Or -march=k8. The recommendation for building a 32-bit chroot in the
gentoo docs says -march=athlon-xp though, possibly because there was some
confu
I've put up a stage3 tarball optimized for UML on amd64 hosts. You can
find it on:
http://gathering.itm.mh.se/~joepal/uml/
This boots in SKAS0 without hanging on the TLS issue. The whole system is
compiled with:
USE="sse 3dnow mmx -acpi -apm -directfb -svga -fbcon"
CHOST="i386-pc-linux-gnu"
c info.
Also, I now have a good infrastructure for running test cases with
different hostkernel/guestkernel/mode/root_fs combos, so suggest more test
cases.
// Joel
On Fri, 7 Oct 2005, Blaisorblade wrote:
On Friday 07 October 2005 13:00, Joel Palmius wrote:
I plan on doing a bit more st
I plan on doing a bit more structured testing and playing-around with
UML+gentoo this weekend. What is the as-of-today best setup for an amd64
machine if I want to run a gentoo guest?
I'm guessing:
host kernel: 2.6.14-rc3 (unpatched)
guest kernel: 2.6.13.3-bs3
Or should I patch the host kerne
de on
an amd64 smp host. (Trying to boot a gentoo root with same setup pukes
horribly though)
// Joel
On Wed, 28 Sep 2005, Blaisorblade wrote:
On Tuesday 27 September 2005 22:21, Joel Palmius wrote:
Whups, wrong script. Well, it works with 2.6.13.2-bs2 too:
Linux version 2.6.13.
tl". Anything else?
Or is there a guest kernel patch which might enable me to run an
unmodified gentoo tarball (compiled for athlon-xp) under skas0?
// Joel
On Tue, 27 Sep 2005, Joel Palmius wrote:
As an addition: The stage tarball is a x86 (i386) one. Trying to use for
example an ath
)
NET: Registered protocol family 16
mconsole (version 2) initialized on /root/.uml/Z7Pwz6/mconsole
ubd: Synchronous mode
(and I can log in and at run various command without a crash)
// Joel
On Tue, 27 Sep 2005, Joel Palmius wrote:
Well, excepting I might have been doing something wrong
ay 27 September 2005 16:02, Joel Palmius wrote:
Lengthy.. Summary is that on an amd64 host, uml hangs on skas0 mode if the
host kernel has been patched with -v9-pre7.
Using an unpatched vanilla
host kernel, skas0 boots. Why this is so I have no clue, ask someone else.
Have you correctly added &
in/sshd
Welcome to Linux 2.6.12-usermode-bs1 (tty0)
darkstar login:
---
At this point I can log in.. Haven't done a whole lot with it though.
// Joel
On Tue, 27 Sep 2005, Blaisorblade wrote:
On Monday 26 September 2005 15:08, Joel Palmius wrote:
Update: I've now tested -bs
As an addition: The stage tarball is a x86 (i386) one. Trying to use for
example an athlon-xp stage tarball crashes with "cannot set up LDT for
thread-local storage"
// Joel
On Tue, 27 Sep 2005, Joel Palmius wrote:
Lengthy.. Summary is that on an amd64 host, uml hangs on skas0 m
Lengthy.. Summary is that on an amd64 host, uml hangs on skas0 mode if the
host kernel has been patched with -v9-pre7. Using an unpatched vanilla
host kernel, skas0 boots. Why this is so I have no clue, ask someone else.
:-)
Further, 2.6.12-bs11 refuses to attempt running in skas0 on an unpatc
The boot works in tt mode but not in skas0 and skas3. I'm assuming this
means the inittab is decent..?
// Joel
On Mon, 26 Sep 2005, Mikael Löfstrand wrote:
Have you tried adding a tty0 in /etc/inittab?
/Micke
"UML Newbie"
Joel Palmius skrev:
Update: I've now teste
at my gentoo-based root_fs sucks somehow, since I only get
the busy hang when using that.
Wild guess anyone? What is it most likely I've done wrong with the gentoo
setup? :-)
// Joel
On Mon, 26 Sep 2005, Joel Palmius wrote:
Does this affect SKAS0 too? I did some more testing, and I ge
Does this affect SKAS0 too? I did some more testing, and I get the same
hangs as previously reported for amd64 (SKAS0 and SKAS3) when having built
up a new system on a 32-bit host with the same patches.
// Joel
On Sat, 24 Sep 2005, Blaisorblade wrote:
Ok, I've now went debugging the SKAS3
and test a new unpatched tree the next few days to confirm
that the same thing happens in both patched and unpatched.
Is there anything else I can do to help debugging?
// Joel
On Fri, 23 Sep 2005, Blaisorblade wrote:
On Friday 23 September 2005 10:24, Joel Palmius wrote:
Hello, I'm tr
Hello, I'm trying to get uml working on my opteron machine. It does, but
only in TT mode. Both host and guest runs reasonably updated versions of
gentoo. The guest system was compiled completely in i686 mode, and so was
the guest kernel. The host runs compiled in x86_64 mode. The guest kernel
i
49 matches
Mail list logo
