On 01/09/15 11:07, Vijay Tandeker wrote:
> Dear Sir,
> - Downloaded "linux-3.18.20.tar.xz" from kernel.org
> - compiled it for UML using "ARCH=um".
> - Downloaded "Ubuntu-TrustyTahr-i386-root_fs" and
> "Fedora21-x86-root_fs".
> - Still same issue with Ubuntu as
On 27/08/15 20:14, Mark Morgan Lloyd wrote:
> The (unofficial?) FAQ at http://uml.devloop.org.uk/faq.html implies that
> it is possible to build UML on a 64-bit system to run a 32-bit guest.
You can build a 32-bit *kernel* from a 64-bit host.
> The best that I can manage on e.g. Debian "Jessie"
Hi all,
i have problem related to networking.
CASE 1:
My host machine has 2 ethernet cards.
Just a guess, but does it work if you disable the other network card?
If the 2 cards are connected to the same subnet/gateway, it is
impossible to make it work without source routing. AFAIK.
The
Hi all,
Just thought I'd let you know that Blaisorblade's patches for 2.4 apply
cleanly to 2.4.29, but compilation fails shortly on my dev box (looks
like some missing headers). What do i need to do?
Cheers
Antoine
---
This SF.Net email is
Hi all,
I was having some very unusual problems with 2.6.9-bs5 and -bb4 running
on an AMD 64 2.6.10-as1 host (2.6.10 + minimal security fixes): some
processes would not start with the correct SELinux context (sshd would
run in kernel_t instead of sshd_t for no apparent reason) but 2.6.10
vanilla
debug1: Authentications that can continue:
publickey,password,keyboard-interactive
debug1: Next authentication method: publickey
debug1: Trying private key: /home/franck/.ssh/identity
debug1: Trying private key: /home/franck/.ssh/id_rsa
debug1: Trying private key: /home/franck/.ssh/id_dsa
The 64-bit guest boots into existing 32-bit root_fs without problems, I
will rebuild all my 64-bit guests and let you know how this goes.
Actually, I got mixed up, the 64-bit filesystems worked but the 32-bit
didn't. I couldn't find any option to allow support for 32-bit binaries,
has it been
{error_exit+0}
8010ca47{__switch_to+311}
804f4de4{thread_return+0}
ff
On Fri, 2005-05-06 at 16:09 +0100, Antoine Martin wrote:
The 64-bit guest boots into existing 32-bit root_fs without problems, I
will rebuild all my 64-bit guests and let you know how this goes
On Sat, 2005-05-07 at 17:53 +0200, Blaisorblade wrote:
On Friday 06 May 2005 19:38, Antoine Martin wrote:
Now, I've hit some real problems. I've managed to crash the host twice
within a short period of time.
I captured the messages from a serial
console the second time
Would you check
On Tue, 2005-10-04 at 12:47 +0200, Blaisorblade wrote:
On Monday 03 October 2005 23:28, Antoine Martin wrote:
On Monday 03 October 2005 14:39, Jeff Dike wrote:
On Mon, Oct 03, 2005 at 09:28:04AM +0100, Chris Lightfoot wrote:
I can run Java on x86 but not on amd64 guests, have you got
Hmm, what's that? Faster than skas3 I guess?!
Is this going to work on x86/amd64 guests on amd64 hosts too?
No, SKAS4 would be as fast as SKAS3, but could be merged in mainline...
Nice, no more patching!
And amd64 hosts would be fixed as well.
Also, I've heard some reports of people
Quoting Blaisorblade [EMAIL PROTECTED]:
And I think that distributing usable root_fs images would be of
great help...
(actually, there are other sources but the site has no link to them
- plus I
don't like that much cherry-picking root_fs's from people I don't
trust and
putting UML's
Hi!
I installed a kernel um (2.4.26) from deb packages. In /var/log/syslog
of um
machine I read this error:
modprobe: modprobe: Can't open
dependencies
file /lib/modules/2.4.26-3um/modules.dep (No such file
or directory)
Because of
I've advanced a little more in running 2.4.19 on 2.6.x...
Aparently the exec-shield and randomize_va_space that comes plugged on
fedo=
ra weren't helping.
I don't know about randomize_va_space, but exec-shield has never caused
me any problems running uml in skas3 or skas0
With exec-shield
Which brings up another question. When I first booted the uml-guest I
got a modules.dep does not exist error. The recommendation I read for
dealing with this was to touch modules.dep, which certainly did get
rid
of the errors.
However, is this problem indicative of further problems I'm
The first major decision is whether to run in 32 or 64 bit mode.
If raw performance is what you are after, use 32-bit. Sad, but true.
(...) I can't expect SKAS3 to do so.
Correct, it doesn't work on 64-bit.
Running in 32-bit mode works almost as well. Chances are SKAS3 would work,
but I'd
Bob Manson wrote:
make ARCH=um
The output of the same is given below.
arch/um/sys-i386/built-in.o: In function `__crc_kunmap':
: multiple definition of `__crc_kunmap'
rm .config
make ARCH defconfig
http://www.mail-archive.com/user-mode-linux-user@lists.sourceforge.net/msg02185.html
Antoine
All,
I've updated the pages on the site (also updated most root_fs) and
finally added the SELinux documentation:
http://uml.nagafix.co.uk/SELinux
This should allow you to use UML to contain/limit UML guests.
Antoine
---
This SF.net email is
And why has the maschine been running with 2.6.11.12 6 months fine,
until this christmas when some guy (me) had the crazy idea of a new
kernel :-)
I had the exact same issue you describe, I guess you also used your
previous .config with a make ARCH=um oldconfig.
$ make defconfig
FYI, on amd64 hosts, using vanilla 2.6.15-rc7, then applying latest
sf.net patches gives:
make ARCH=um vmlinux
(...)
CC arch/um/kernel/skas/process_kern.o
arch/um/kernel/skas/process_kern.c: In function ‘copy_thread_skas’:
arch/um/kernel/skas/process_kern.c:127: error: ‘struct arch_thread’
Problematic things are in kernel hacking/debugging, and general features
(smp, 3-level pages, etc), watchdog? etc.
I still didn`t figure out how the watchdog stuff works. I have HW
watchdog on the host but a uml guest watchdog would be nice too. Any doc
or howto?
AFAIK, there isn't
On Sat, 2005-12-31 at 23:04 -0500, Jeff Dike wrote:
On Tue, Dec 27, 2005 at 06:31:43PM +, Antoine Martin wrote:
On Tue, 2005-12-27 at 17:53 +, Antoine Martin wrote:
FYI, on amd64 hosts, using vanilla 2.6.15-rc7, then applying latest
sf.net patches gives:
make ARCH=um SUBARCH
CC arch/um/os-Linux/tls.o
arch/um/os-Linux/tls.c: In function `os_set_thread_area':
arch/um/os-Linux/tls.c:22: error: dereferencing pointer to incomplete
type
arch/um/os-Linux/tls.c: In function `os_get_thread_area':
arch/um/os-Linux/tls.c:36: error: dereferencing pointer to
Failed right at the end... (TT disabled)
LD .tmp_vmlinux1
arch/um/sys-i386/built-in.o(.text+0x2f8e): In function `load_TLS':
: undefined reference to `indirect_set_thread_area'
collect2: ld returned 1 exit status
KSYM.tmp_kallsyms1.S
nm: '.tmp_vmlinux1': No such file
No valid symbol.
Thanks, that fixes it.
On Mon, 2006-01-16 at 12:10 -0500, D. Bahi wrote:
for the second problem try the attached patch of a patch.
Antoine Martin wrote:
On Fri, 2006-01-06 at 17:41 +, Antoine Martin wrote:
Failed right at the end... (TT disabled)
LD .tmp_vmlinux1
should it work? coz it doesn't /8-
Look at error messages and let them help you finding the road...
The error message is:
$ mknod /dev/udb0 b 98 0
$ mount -t ext2 /dev/udb0 /mnt
mount: /dev/udb0 is not a valid block device
Looks to me like this kernel does not have ubd
I'm new to usermode linux and I'm trying to get a guest up and
running on a Ubuntu Breezy based host. I'm running the kernel that
is included in Ubuntu and I'm attempting to build the 2.6.11 kernel
for uml. It builds correctly but when I go to run it, it gives me
the following
BUG: soft lockup detected on CPU#0!
(...)
Found another similar one on a 2.6.15.1 guest running in skas0 on amd64:
[43569638.90] BUG: soft lockup detected on CPU#0!
[43569638.90]
[43569638.90] EIP: :[] CPU: 0 Not tainted EFLAGS:
[43569638.90] Not
Excellent, the difference in speed is noticeable. no benchmarks yet, but
the load on the host seems lower.
Any idea what this is:
Kernel panic - not syncing: Kernel mode fault at addr 0x0, ip 0x4b20
deactivate_all_fds failed, errno = 9
It is 100% reproducible on a slackware10 filesystem
disassemble 0x4b20 0x4b40
(gdb) disassemble 0x4b20 0x4b40
Dump of assembler code from 0x4b20 to 0x4b40:
0x4b20: Cannot access memory at address 0x4b20
(gdb) c
Continuing.
Program received signal SIGUSR1, User defined signal 1.
0xa02aebb1 in kill () at err.h:24
On Tue, 2006-02-21 at 16:08 -0500, Jeff Dike wrote:
Kernel panic - not syncing: Kernel mode fault at addr 0x0, ip 0x4b20
deactivate_all_fds failed, errno = 9
This is the Slackware image from your site? And a stock i386 UML? Which
version?
ok, it wasn't the stock versions, but the
That hunk goes into
arch/um/kernel/skas/trap_user.c for 2.6.15, however.
I tried applying it but the first part was already applied in -bs2, the
second applied ok though.
Good news is that I made a mistake (yet again), tried to apply it twice.
The patch is fine.
Bad news is that the guest
Hi list,
I have added some new images here:
http://uml.nagafix.co.uk/
Fedora Core 4 AMD64, Mandriva-2006 x86 AMD64, OpenSuse x86.
All these now work well with TLS, tested with 2.6.15.6-bs3-tls.
The last images should be uploaded by the time you get this email.
I still have to rebuild some new
Hi,
Thanks for posting your root filesystem images and kernels (and configs),
it was very useful to find a working version first that I could then
use to rebuild my kernel and make sure it works.
Thanks for the feedback.
1. Could you post the script you use to create the Mandriva
When I boot the UML I notice it says I'm running in SKAS0 mode:
What's the command line you used?
Checking for the skas3 patch in the host:
- /proc/mm...not found
- PTRACE_FAULTINFO...not found
- PTRACE_LDT...not found
UML running in SKAS0 mode
It doesn't look like you're running on a
On Thu, Jun 29, 2006 at 11:56:05PM +0800, paul john wrote:
i tried to make a system ,that users can't modify it. is that often
be used
It can be done, but it is more difficult using a standard distribution
as a base: you're better off starting with something like busybox.
Early in the boot,
I've upgraded a Gentoo amd64 box to the latest gcc (4.1.1) and glibc
(2.4-r3). glibc 2.4 is nptlonly (at least on gentoo?)
The result is that all kernels just spin after mounting root:
sysrq-t shows all threads doing the same thing (switch_threads):
[42949405.03] kjournald S 400029FB
Sven wrote:
Hi,
has anybody of you gentoo-users out there (maybe Blaisorblade?) tried
UML with glibc 2.4 (which seems to be nptlonly!)?
Yes.
Does it compile?
Yes.
Does it work?
No. :(
See my previous post for details.
The exact same images run absolutely fine on x86 hosts with almost
identical settings (same gcc, glibc versions)
Can i conclude from your words, that UML runs fine on x86 32-bit with
gcc 4.1.1 and glibc 2.4
Works for me (tm) on x86 hosts.
Sorry about the confusion, the other mail should
Blaisorblade wrote:
On Tuesday 05 September 2006 02:47, Antoine Martin wrote:
I've upgraded a Gentoo amd64 box to the latest gcc (4.1.1) and glibc
(2.4-r3). glibc 2.4 is nptlonly (at least on gentoo?)
I'd be curious about which is the host kernel and its configuration - there
Host kernel
- actually_do_remove failed with err = -2
#
Looks like the 64-bit kernels work ok though. How weird.
Any ideas?
Antoine Martin wrote:
Blaisorblade wrote:
On Tuesday 05 September 2006 02:47, Antoine Martin wrote:
I've upgraded a Gentoo amd64 box to the latest gcc (4.1.1) and glibc
(2.4-r3). glibc 2.4
UML running in SKAS0 mode
remove_umid_dir - actually_do_remove failed with err = -2
Add stderr=1 to the command line and see what happens.
Nothing happens, same output as above.
Also, if a 32-bit binary behaves differently on a 64-bit system than a
32-bit one, I would say that's a bug in
On Fri, 2006-09-15 at 12:59 -0400, Jeff Dike wrote:
On Tue, Sep 05, 2006 at 01:47:51AM +0100, Antoine Martin wrote:
I've upgraded a Gentoo amd64 box to the latest gcc (4.1.1) and glibc
(2.4-r3). glibc 2.4 is nptlonly (at least on gentoo?)
The host filesystem or the UML filesystem
There are a few servers I would like to be able to reboot without
loosing the UML instances running on them. At the moment, I am just
relying on the guests not needing to be restarted (most of them have
long uptime so that's no too bad)
I.e. they were started before the glibc update so they
Verify if dynamic linking helps at all - if something such is involved and
the
constant is linked in when linking with glibc, this would help.
I've tried non-static builds but it didn't help...
Antoine
-
Using Tomcat
Check out uml.nagafix.co.uk. Everything you need should be
there. Beware, though, that not all the guest kernels
hosted on that site work. The only configuration I have
been able to get working is x86 host kernel with
2.6.16.29-bs2 32 bit guest, and either my own debootstrapped
image or
On Mon, Sep 18, 2006 at 09:06:49PM -0500, Ben Wing wrote:
i'm trying to get uml working with modules. i downloaded the latest kernel
2.6.17.13, compiled using the `defconfig', obtained the root fs from
http://uml.nagafix.co.uk/Debian-3.1/ (which claims to have been updated sept
13 2006), and
iface eth0 inet dhcp
but that didn't work; when you `ifup eth0', you get an intro message
from the dhcp client and then it hangs the terminal entirely.
any ideas?
Failed to open /dev/net/tun, err = 13
crw-rw 1 root uml-net 10, 200 2006-08-24 09:17 /dev/net/tun
Can you access
Can you access /dev/net/tun as the user than runs the guest?
(are you running it as uml-net?)
What about the permissions on /dev/net, etc...
Yup, I can `touch /dev/net/tun' and write onto it. /dev/net is 755:
drwxr-xr-x 2 root root 60 2006-08-24 09:17 /dev/net/
but it doesn't
then tried the following:
# cd /usr/src/redhat/SPECS
# rpmbuild -ba --target=um kernel-2.6.spec
which gave me the following:
Building target platforms:um
Building for target um
error: Architecture is not included: um
So I then decided to try your suggestion:
# cd
Ah, 2.6.9...
I understand your motivation, but I wouldn't recommend it.
Use something more recent (more stable).
2.6.9 is way too old for what you're trying to do.
On Wed, 2006-09-20 at 05:49 -0700, Mathew Brown wrote:
Actually, all the rpmbuild does from my understanding is to patch the
stock
On Wed, 2006-09-20 at 06:57 -0700, Mathew Brown wrote:
So you would suggest I go with something like VMware Server or Xen for
prototyping / testing RHEL-based solutions? I would have preferred to
consolidate all my testing / prototyping on UML.
Personally I would go with UML and ~2.6.18
But
I've done some more tests with 2.6.18.
* building a x86 kernel on amd64 results in a binary that crashes early
(see strace and gdb below)
* building it on x86 and copying it across works better, but there are
still some TLS issues I believe.
the compilers is gcc 4.1.1 in both cases.
There
gdb it, put a breakpoint on panic, run it, and see what happens.
(gdb) break panic
Breakpoint 1 at 0x808de78
(gdb) run
Starting program: /uml/uml-3/chroot/kernel-3 noprocmm ubd0=./uml/root_fs
Checking that ptrace can change system call numbers...OK
Checking syscall emulation patch for
sure I'll have some more
questions.
-mike
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Antoine Martin
Sent: Tuesday, September 26, 2006 2:39 AM
To: user-mode-linux-user@lists.sourceforge.net; Mike Horn
Subject: [uml-user] 2.6.18 with FC5
Hi Mike,
I downloaded a 2.6.18 kernel and FC5 filesystem from Antoine Martin's
website and I see the same behavior. Are there any known issues with
running a 2.6.18 guest system with FC5 on a 2.6.18 host system with FC5?
I am seeing the same thing here on amd64 host - I have reported it with
On Tue, 2006-09-26 at 11:39 -0400, Jeff Dike wrote:
On Tue, Sep 26, 2006 at 03:31:00PM +0100, Antoine Martin wrote:
Ah, so it is TLS on the host that's causing the problem, quite odd.
The problem I have is that glibc 2.4 is nptlonly and so I just can't
disable TLS... I'm really stuck. I
Hi,
On Mon, 2006-10-09 at 11:06 -0400, Fei Liu wrote:
Hi, I visited your website http://uml.nagafix.co.uk/ about UML, it's
excellent.
Thanks very much!
Note it's subnet 192.168.0.0 not 192.168.0.1 (dhcpd will complain/stop
when it's 192.168.0.1)
Oops sorry about that, I've updated the
try adding mem=256 to the kernel command line, your guest is out of
memory.
[42950509.72] Out of Memory: Kill process 2175 (yum) score 564 and
children.
[42950509.72] Out of memory: Killed process 2175 (yum).
Killed
Antoine
The other thing you can try is to make a simple initrd file with some
busybox
I don't have any idea to do that. I don't know procedures of any type in
this case.
You can download one ready-made here:
http://uml.nagafix.co.uk/busybox/
/do-chroot . 1004 ./kernel ubd0=/disk1 mem=96M uml_dir=/tmp
The user 'chtest' has the uid 1004, 'kernel' is a recent 2.6.18 with a skas3
patch.
So far, there's no problem, everything works fine. Just one thing:
SKAS3 won't work
On startup, there's the skas3 capability test and a
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
make ARCH=um SUBARCH=i386
(...)
LD .tmp_vmlinux1
arch/um/sys-i386/built-in.o: In function `kunmap_atomic':
: undefined reference to `kpte_clear_flush'
collect2: ld returned 1 exit status
KSYM.tmp_kallsyms1.S
nm: '.tmp_vmlinux1': No such
the pcap code?) but they stopped working after the host
was upgraded to glib 2.4.
On Wed, 2006-12-20 at 12:58 -0500, Jeff Dike wrote:
On Thu, Dec 14, 2006 at 12:02:19PM +, Antoine Martin wrote:
I believe this issue is linked to Fedora's glibc 2.4 or/and gcc 4.1 on
amd64 platforms only (I have
Downgrading to skas-v8.2 makes things work fine again.
Ah, did you get around to generate a v8.2 patch for 2.6.19?
I think the key to reproducing this is a skas-v9 host. It doesn't
happen for me on every file system, which is somewhat interesting.
Yep, it seems like a nptl/glibc issue (just
I've run into out of memory errors, when
Out of memory: Killed process 895 (xsltproc).
/xmlto: line 379: 895 Killed
xsltproc $XSLTOPTS -o $XSLT_PROCESSED $STYLESHEET
$INPUT_FILE
xslt often consumes large amounts of memory (no matter what language is
used to implement it), I
I am actually writing about UML. I also have to demonstrate how a
malicious user can break out of a UML-instance. My question is if there
exist tools which make breakouts available. E.g. to bypass ptrac() or
bad kernel-modules.
Using code similar to hostfs makes it pretty easy to tamper
I got it working with mem=512M.
I got it by searching google.
This option should be put in the howto on the web site
I am sure it is there somewhere, just not very well organized.
so that new user knows about it.
Like all quality unix command line tools:
'--help' is your friend.
Is there a
Warning: unable to open an initial console.
Am I missing something?
Yes, a console!
crw--- 1 root tty 5, 1 /dev/console
-
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and
Magesh,
I can't remember if I did find a solution to this problem.
Which version are you trying to build? (host settings, etc)
Antoine
Magesh mannar wrote:
Dear Antoine
This mail is in reference to one of your postings to
the uml mailing list
anukalp desai wrote:
after issuing command ./linux ubda=root_fs_nameoffilesystem
which fs is this?
uml process starts
and after VFS: mount root(ext2) file system read only
Is the filesystem read-only?
the terminal hangs
can anyone please help
http://uml.nagafix.co.uk/FedoraCore6/
Full installation scripts included.
Please report any problems.
-
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to
http://uml.nagafix.co.uk/Slackware-11/
Full installation scripts included.
Please report any problems.
-
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to
the SAKS patch used is from
http://www.user-mode-linux.org/~blaisorblade/patches/skas3-2.6/skas-2.6.18-v9-pre9/skas-2.6.18-v9-pre9.patch.bz2
Try using 8.2 as there are problems with v9-pre9
my_machine%./linux ../../Debian-3.0r0.ext2
Checking that ptrace can change system call numbers...OK
$ ./linux ./nagafix_try/Debian-3.1-x86-root_fs
Checking that ptrace can change system call numbers...OK
Checking syscall emulation patch for ptrace...OK
Checking advanced syscall emulation patch for ptrace...OK
Checking for tmpfs mount on /dev/shm...OK
Checking PROT_EXEC mmap in
I had recompiled my host kernel with setting the EXT2_FS_SECURITY and
EXT3_FS_SECURITY in filesystem configuration. My UML is OK before doing this,
but now I can not boot the UML in the new-recompiled kernel. It just went to
the step...
That's very interesting! Thanks for reporting it.
I am
Between stop and go the filesystem is still mounted on the guest right?
If so, you are screwing the filesystems by mounting them on the host:
the mount command will run fixups on the (journaled?) filesystem.
(You could try mounting read-only - but still, bad idea IMO)
Antoine
Jonas Meyer wrote:
How about attaching a disk with mconsole, mount it in the guest, copy
the data, umount, detach disk. done?
Chris Lightfoot wrote:
On Tue, Jan 23, 2007 at 12:14:53PM -0500, Jonas Meyer wrote:
Hmm... What is the preferred method of backup for a VM then? As I
said, I'd rather not install ssh
Hi,
thanks for that, it should help figure out which release/patch caused
the problem.
Could you try the same test using plain kernel.org kernels?
(2.6.17 and 2.6.18 for a start)
That should tell us if it is was a redhat patch that introduced the problem.
Antoine
JuanJo Ciarlante wrote:
Hi!
You are using a modular kernel but you haven't added the modules to the
filesystem image.
Your options:
* use a static kernel (like the ones on my site - let me know if these
do not work for you)
* loop mount the image and add the modules
Maren Peasley wrote:
Hello,
I am attempting to setup
Hi,
You are not doing anything wrong, if you check at the top of the page at
http://uml.nagafix.co.uk/
It does mention the fact that there are problems with recent host
kernels (=2.6.16). I can only suggest that you try an older host kernel
until the x86 bug is fixed.
Antoine
(I'll top post,
and
enable
SELinux in same 2.6.19.2 kernel which is working ??
The SELinux options are in the Security section, just enable them.
If the kernel you compiled yourself works it should make little difference.
Antoine
On 2/1/07, Antoine Martin [EMAIL PROTECTED] wrote:
Hi,
You are not doing
and choose any combination of kernel and fs
(well, that's the theory anyway)
Antoine
That line of reasoning is correct ?
Regards
Ernani
2007/2/1, Antoine Martin [EMAIL PROTECTED]:
Pravin wrote:
Thanx for reply.
I changed my host kernel to 2.6.15 and tried to run those
precompiled
host kernel ?
It varies. Depends on which machine.
Is desirable that unl-kernel version match with the host kernel version ?
No. If at all possible use an up-to-date host and guest.
Note: there are issues with x86 host kernels = 2.6.16
Antoine
Regards
Ernani
2007/2/1, Antoine Martin [EMAIL
kernels = 2.6.16
Ok. Could you suggest a good stable 2.6.xx version ?
Host:
On x86: 2.6.15-skas3-v8.2
On amd64: 2.6.19.2 +skas3 + patch Jeff posted yesterday
Guest:
latest.
Antoine
Regards
Ernani
2007/2/1, Antoine Martin [EMAIL PROTECTED]:
Please always CC the list, other UML people could
AUDIT, NETWORK_SECMARK are not very intutive to me to be
associated with SELINUX
and now its working fine...
On 2/1/07, Antoine Martin [EMAIL PROTECTED] wrote:
Pravin wrote:
Thanx for reply.
I changed my host kernel to 2.6.15 and tried to run those
precompiled and
manually compiled
Haywood Floyd wrote:
Hi Sirs.
I am new in uml affairs and wish to clear some doubts.
I hesitated in send this message because the level of threads posted
here is far beyond the respective message subject. However I
think...well, everybody started from beginning...so:
I am building an uml
Haywood Floyd wrote:
bs stands for Blairsorblade's Stable branch.
tls is required to run modern distributions that use newer glibcs (with
nptl).
How can i find, based on kernel source version that i want to use,
which one pick ?
With 2.6.19.2 there aren't any bs or tls patches to apply.
Haywood Floyd wrote:
Ok. But...
With 2.6.19.2 there aren't any bs or tls patches to apply.
I remember that you suggested to me a 2.6.15 kernel version.
You say ..there are issues with x86 host kernels = 2.6.16.
Is it only for host kernel ?
Yes
For uml kernels its better use the last
Haywood Floyd wrote:
Thus,
for host kernel: 2.6.15 version
for guest (uml) kernel: 2.6.20 version
I have to ask: uml is a linux (working in user mode). Why the difference ?
Because there are issues with host kernels 2.6.15 (otherwise you would
use them - you could try and see if it works
Haywood Floyd wrote:
Hum.
For guest (uml) kernel: What patch file is more appropriated ?
I find no patch file matching the 2.6.20 version in Mr Blasorblade
directories.
As I said, patching is not needed for the latest versions.
The bug fixes in the previous patches (and more) have been
Haywood Floyd wrote:
I'm sorry. I dont make myself clear. Its my fault.
I mean, which uml patch i must apply to a 2.6.20 kernel to turn it
into an uml (guest) kernel ?
None.
My fault again.
When i say 2.6.20 kernel i mean the 2.6.20 kernel source downloaded
from www.kernel.org. A host
(guest) kernel ?
You can't. As I said, start with a brand new clean source tree.
The hole difference is made para ARCH=um ?
Yes.
Antoine
Thanks
2007/2/5, Antoine Martin [EMAIL PROTECTED]:
Haywood Floyd wrote:
I'm sorry. I dont make myself clear. Its my fault.
I mean, which uml patch i
, Antoine Martin [EMAIL PROTECTED]:
Haywood Floyd wrote:
Very enlightening.
I think i had get the concept in a wrong form.
Two questions:
1) what is the patch files for ?
The ones we have been discussing would improve uml support.
2) where resides the modifications that turn a originally
for 2.6
Is correct to say : if i get a kernel source, apply no patch at all,
compile with make ARCH=um + etc.. then i get a fully operational uml
guest ?
Yes.
Antoine
2007/2/5, Antoine Martin [EMAIL PROTECTED]:
Haywood Floyd wrote:
I think i get a little confused.
The key aspect here
Jochen Antesberger wrote:
Hi,
I'm using uml with the tuntap option. I.e. I have eth0=tuntap,tap1 and
eth1=tuntap,tap2 as part of the command line.
Try specifying the hardware mac address on the command line every time:
eth0=tuntap,tap1,fe:f0:00:00:00:01,10.0.1.254
Antoine
When I updated
Hi,
Which version of uml utilities are you using?
I've successfully used non standard paths for years.
ie: uml_mconsole /path/to/mconsole
Antoine
ghannemann wrote:
Hi
Sadly I did not get any responses for the question a while back. Here it is
again - in short.
Why can I not set the
It was true when it was written, I believe some (but not all) fixes have
made it into 2.6.20. Better ask Jeff.
I'll happily remove the warning when I get a chance to re-test.
Antoine
n schembr wrote:
I found a warning on http://uml.nagafix.co.uk/
main page.
Warning: because of a recent
[snip]
Checking PROT_EXEC mmap in /dev/shm/...failed: Operation not permitted
/dev/shm/ must be not mounted noexec
The message is as explicit as can be.
Can someone help me understand this?
Check /dev/shm in your /etc/fstab and remove noexec.
This has been covered many many times on the ML,
Better ask Jeff, I've been quite busy for the last 3 weeks and I know
there were lots of patches flying around.
Antoine
---BeginMessage---
Hi!
First of all, my congratulations for your web page, it's very useful for
we the begginers in UML.
Next, my question: I was trying to find
Brad Brock wrote:
Hi List, I'm running uml on debian. I installed uml
from debian packages. I've patched the host kernel
with skas. When I ran multiple uml for the first time,
some of them was disconnected.
I had tried to enlarge the tmpfs, but there wasn't
much difference. I ran 180 uml
1 - 100 of 299 matches
Mail list logo