Re: [uml-user] Random Lockups

On 11 Jul 2008, Jeff Dike verbalised: > On Fri, Jul 11, 2008 at 06:01:32PM +0100, Jay Shah wrote: >> I'm running a 2.6.20.20 host and 2.6.24.3 guest. The guest will >> randomly 'freeze' and also stop/go doesn't work at all on any >> machines. Would upgrading to a new(er) kernel fix this [see botto

Re: [uml-user] Running UML over Vmware

On 7 Jul 2008, Jeff Dike told this: > On Sat, Jul 05, 2008 at 07:27:44PM +0200, David Fernández wrote: >> Any hint about the cause of this big performance difference? Should >> the installation of the SKAS3/4 patch on the host help to solve it? > > skas4 will certainly help. I get ~86% of nativ

Re: [uml-user] skas4 on x86-32 -> boom at startup, and fallback failing too

On 25 Jun 2008, Benedict Verheyen uttered the following: > I have skas4 working on 2.6.25.4. You need to patch both the guest & > host kernel but from you said i conclude that you already knew that. Yes, they were both patched. -- `If you are having a "ua luea luea le ua le" kind of day, I can

[uml-user] skas4 on x86-32 -> boom at startup, and fallback failing too

I felt unhappy recently because UML was working perfectly well and I had no complaints (except for the absence of skas3 on 2.6.25 hosts). Obviously this was unacceptable. So I tried out skas4, host+guest (the guest also has your rng patches, time fixes, and SIGWINCH fixes applied). The result: ,

Re: [uml-user] what is modified?

On 31 Mar 2008, fang zheng outgrape: > hi: >To port linux kernel to linux host,the linux kernel must be modified. > but what is modified? It's a new kind of architecture. Look in arch/um and include/asm-um. -- `The rest is a tale of post and counter-post.' --- Ian Rawlings

Re: [uml-user] skas3 on 2.6.23? switch_mm_skas - PTRACE_SWITCH_MM failed, errno = 22

On 28 Nov 2007, Jeff Dike said: > On Mon, Nov 26, 2007 at 07:05:21PM -0800, Jason Lunz wrote: >> Is anyone still using skas3 on modern kernels? I've been booting on >> 2.6.23 with noprocmm to avoid crashing with "Kernel panic - not syncing: >> switch_mm_skas - PTRACE_SWITCH_MM failed, errno = 22".

Re: [uml-user] /proc/mm in 2.6.23

On 21 Oct 2007, Antoine Martin verbalised: > Nix wrote: >> It works with `noprocmm'; thus, a trivial forward-porting of the skas >> patch is broken in 2.6.23. >> >> SKAS0 is tolerable (and far better than tt mode!) but it spams my >> process accounting logs

Re: [uml-user] /proc/mm in 2.6.23

On 20 Oct 2007, [EMAIL PROTECTED] said: > I'm afraid that on x86-32 I get an extremely efficient coredump from all > UML guests with this patch installed :/ however, I haven't tried > building a skasless 2.6.23 host kernel yet: maybe I get an extremely > efficient coredump there, too. It works wit

Re: [uml-user] /proc/mm in 2.6.23

On 14 Oct 2007, Antoine Martin verbalised: > It's very rough around the edges but works for me(tm): > http://uml.nagafix.co.uk/skas-2.6.23.patch.bz2 > (can't enabled visible process cmdlines - some struct has changed) I'm afraid that on x86-32 I get an extremely efficient coredump from all UML gue

Re: [uml-user] Strange clock problem

On 9 Jan 2007, Paolo Giarrusso told this: > Nix <[EMAIL PROTECTED]> ha scritto: >> Not so much `binaries', more `glibc'. (See the definition of >> DL_SYSDEP_OSCHECK in glibc-*/sysdeps/unix/sysv/linux/dl-osinfo.h.) > > My point (which maybe I didn't state c

Re: [uml-user] Strange clock problem

On 3 Jan 2007, [EMAIL PROTECTED] outgrape: > On Tuesday 26 December 2006 11:46, Stephane Bortzmeyer wrote: >> FATAL kernel is too old >> >> (And I cannot change the host at will, only the guest.) > > According to Nix, you're using binaries compiled on a more rece

Re: [uml-user] NPTL on Host Issue?

On 2 Dec 2006, eTecc Support spake thusly: > Wow; I feel silly. Enabling UBD support in the UML Kernel. But now I get > this: > > ubda: unknown partition table > kjournald starting. Commit interval 5 seconds > EXT3-fs: mounted filesystem with ordered data mode. > VFS: Mounted root (ext3 filesy

Re: [uml-user] Apache segfaults at every request

On 13 Nov 2006, [EMAIL PROTECTED] said: > On Sunday 12 November 2006 21:38, Stephane Bortzmeyer wrote: >> On Fri, Nov 10, 2006 at 11:53:36PM +0100, >> Blaisorblade <[EMAIL PROTECTED]> wrote >> >> a message of 34 lines which said: >> > 2.6 kernels can run on 2.4 systems >> >> Each time I try, I g

Re: [uml-user] Boot problem

On 26 Oct 2006, [EMAIL PROTECTED] uttered the following: > On Wednesday 25 October 2006 23:37, Nix wrote: >> On 25 Oct 2006, Christopher Marshall verbalised: >> > Thanks for all the information about initramfs, that was a very >> > informative post. >> >>

Re: [uml-user] Boot problem

On 25 Oct 2006, Christopher Marshall verbalised: > Thanks for all the information about initramfs, that was a very informative > post. I can but try. :) I've stuck my boot scripts online at . I don't know if they'll be any use, but at least it give

Re: [uml-user] Boot problem

On 12 Oct 2006, Christopher Marshall told this: > The other thing you can try is to make a simple initrd file with some > busybox commands in it, enough to give you the ash shell and some > commands like mount, pivot_root, chroot, so you can try to mount the > root filesystem yourself then pivot_ro

Re: [uml-user] Starting udev:

On Thu, 17 Aug 2006, Jeff Dike wrote: > On Thu, Aug 17, 2006 at 12:54:18PM -0400, Rick Spillane wrote: >> I was wondering why this phase of the boot process is so slow. Is >> there some way I can speed it up? If there is some configuration I can >> tweak to skip this step, that would be great, if i

Re: [uml-user] about mem=nM

On Tue, 27 Jun 2006, Jeff Dike mused: > On Wed, Jun 28, 2006 at 12:36:19AM +0100, Nix wrote: >> On Wed, 28 Jun 2006, paul john uttered the following: >> > if start um without mem=256M , it's ok , if start um with mem=256M it hangs >> > there .report "mount

Re: [uml-user] about mem=nM

On Wed, 28 Jun 2006, paul john uttered the following: > if start um without mem=256M , it's ok , if start um with mem=256M it hangs > there .report "mount devfs on /dev" Do you have at least 256Mb of space free in /dev/shm? (In older UMLs, pre-2.6.17, you'll need to make sure /tmp has enough room

Re: [uml-user] newbe questions

On Wed, 22 Mar 2006, David Lang uttered the following: > when you do ifup debian immediatly attempts to startup sendmail > (apparently by doing make as you see here). if you can't do name > resolution until other interfaces come up you end up with a 30 second > delay in your machine booting. I act

Re: [uml-user] newbe questions

On Wed, 22 Mar 2006, [EMAIL PROTECTED] whispered secretively: > On Wednesday 22 March 2006 17:46, David Lang wrote: >> Configuring network interfaces...make: /etc/mail/Makefile: No such file or >> directory >> make: *** No rule to make target `/etc/mail/Makefile'. Stop. >> make: /etc/mail/Makefile

Re: [uml-user] Sockets question

On Thu, 8 Dec 2005, Christopher Barry moaned: > when two virtual machines use sockets between one another on a single > local system, does that end up routed through the loopback interface? Well, it shouldn't end up routed through any physical network cards, but it shouldn't go through `lo'. At l

Re: [uml-user] Help compiling kernel 2.6.7

On Wed, 7 Dec 2005, noel anderson murmured woefully: > Am trying to compile kernel 2..6.7 with the equivlent patch 2.6.7-1. Any > help is appreciated. > I'm using FC4. There's *very* little chance that a UML patch of such great age will work reliably. IIRC, UML started working reliably around

Re: [uml-user] Three newbie questions

On Fri, 2 Dec 2005, Rob Landley gibbered uncontrollably: > On Friday 02 December 2005 07:09, Nix wrote: > >> > 1. There are many files created in the current directory on the host, >> > named like "1133497700-955364", containing fragments of the output, e.g. >

Re: [uml-user] Three newbie questions

On Fri, 2 Dec 2005, Klaus-J. Wolf suggested tentatively: > Hi, > > I am aware of the fact that my small troubles tend to be rather > unspecified, but maybe someone can help anyway. > > I run 2.6.14-ck6-skas3-v8.2 as host and 2.4.28-bs2 as guest OS. > > When I start "linux": > > 1. There are m

Re: [uml-user] line_write_room: tty1: no room left in buffer

On Wed, 30 Nov 2005, Jeff Dike uttered the following: > If the patchset were so excellent, it would be in mainline by now. It still > needs work, even though it is correct AFAICT. Merely the existence of the patchset is excellent :) -- `Y'know, London's nice at this time of year. If you like yo

Re: [uml-user] line_write_room: tty1: no room left in buffer

On Wed, 30 Nov 2005, Jeff Dike spake: > On Wed, Nov 30, 2005 at 11:03:56AM +0000, Nix wrote: >> There's actually two halves to this: >> >> - UML doesn't support NPTL in its guests, move /lib/tls away > > My current patchset is running NPTL just fine. Yea

Re: [uml-user] line_write_room: tty1: no room left in buffer

On Tue, 29 Nov 2005, [EMAIL PROTECTED] announced authoritatively: > Uff, it's the no. 1 UML FAQ - move /lib/tls away in the guest image, UML > doesn't support NPTL (yet). There's actually two halves to this: - UML doesn't support NPTL in its guests, move /lib/tls away - UML doesn't work if an N

Re: [uml-user] Gentoo stage3 for uml on amd64 host

On Mon, 17 Oct 2005, [EMAIL PROTECTED] stated: > On Sunday 16 October 2005 19:31, Nix wrote: >> These are synonyms for the same thing, as is -march=opteron and >> -march=athlon-fx. > >> -march=x86-64 is similar but disables use of 3DNow intrinsics (not that >> thos

Re: [uml-user] Gentoo stage3 for uml on amd64 host

On Sat, 15 Oct 2005, Joel Palmius wrote: > 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. These are synonyms for the same thing, as is -march=opteron and -march=athlon-fx.

Re: [uml-user] user-mode-linux 2.6.13.2-bs3 oddity

On Tue, 11 Oct 2005, [EMAIL PROTECTED] uttered the following: > On Monday 10 October 2005 00:01, Nix wrote: >> > Patches are attached - to apply in this order. Actually the first should >> > fix everything, but I'm gonna merge all three ones in -bs4, so you can >

Re: [uml-user] user-mode-linux 2.6.13.2-bs3 oddity

On Sun, 9 Oct 2005, [EMAIL PROTECTED] yowled: > On Monday 03 October 2005 21:50, Nix wrote: >> Well, 140789027962880 divides evenly into 2^32. Byte-ordering or a >> one-long misalignment of some kind or I'll eat my hat. (I haven't got a >> hat so I'll h

Re: [uml-user] user-mode-linux 2.6.13.2-bs3 oddity

On Mon, 3 Oct 2005, [EMAIL PROTECTED] yowled: > On Sunday 02 October 2005 22:52, Nix wrote: >> Size mismatch (140789027962880 vs 209715200) of COW header vs backing file >> Failed to open '/mirror/uml/esperi-root-cow.image', errno = 22 >> VFS: Cannot open root devi

[uml-user] user-mode-linux 2.6.13.2-bs3 oddity

I'm using a COWed block device as my root filesystem. After boot, I see this: Size mismatch (140789027962880 vs 209715200) of COW header vs backing file Failed to open '/mirror/uml/esperi-root-cow.image', errno = 22 VFS: Cannot open root device "98:0" or unknown-block(98,0) Please append a correct

[uml-user] 2.6.12.[35]-bs{7,11} boot failure

(Those are all the 2.6.12 UMLs I've tried.) Basically, userspace never starts: the last two lines I see are VFS: Mounted root (ext2 filesystem) readonly. line_ioctl: tty0: ioctl KDSIGACCEPT called [halt] (If I reboot into 2.6.11.9-bs5 with the same root fs, the next line I see is init's banner.)

Re: [uml-user] Hardware Requirements -Some Questions about UML by a Newbie

On Thu, 11 Aug 2005, eric leitner wondered: > - SKAS / TT > Is SKAS really worth the trouble? I mean is the performance better? Or > are there other issues why one should use SKAS Mode? Drastically better performance, easier debuggability, lower load on the host, better security. Use it unless you

Re: [uml-user] UML using 100% CPU after mounting root filesystem

On Fri, 22 Jul 2005, [EMAIL PROTECTED] yowled: > On Friday 22 July 2005 18:53, Ruaidhri Power wrote: >> On Thu, Jul 21, 2005 at 01:23:46AM +0200, Blaisorblade wrote: >> > Which was the old UML version? Which is your guest distro? >> > Also, have you tested the old UML releases on this new host kern

Re: [uml-user] Running SKAS: 3 processes ok, one terminated

On Sat, 4 Jun 2005, [EMAIL PROTECTED] moaned: > Hi all, > > I am running UML with the SKAS patch applied. All works fine although my > host system shows one UML-process as terminated: > > 2605 pts/1S 0:00 ./linux (uml1) [/sbin/getty] > 2607 pts/1T 0:00 [linux] > 2612 pts/1

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Tue, 17 May 2005, [EMAIL PROTECTED] moaned: > On Tuesday 17 May 2005 01:46, Blaisorblade wrote: >> Also, about TLS and such: do you know where I can download the userspace >> example code? > Here I refer to futexes, sorry. /pub/linux/kernel/people/rusty/futex-2.2.tar.gz, I think. -- `Once aga

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

[Sorry for the delay: holiday and illness collide.] On Tue, 17 May 2005, [EMAIL PROTECTED] announced authoritatively: > On Tuesday 17 May 2005 01:17, Nix wrote: >> On Tue, 17 May 2005, [EMAIL PROTECTED] yowled: >> > On Saturday 14 May 2005 14:00, Nix wrote: >> >> At

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Tue, 17 May 2005, [EMAIL PROTECTED] suggested tentatively: > On Tue, 17 May 2005, [EMAIL PROTECTED] yowled: >>So, we could conditionalize that (maybe -D_SEARCH_PATH=). I'm >> not going to merge it in this form, so send me the cleaned up version. > > If I could work out what to

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Tue, 17 May 2005, [EMAIL PROTECTED] yowled: > On Saturday 14 May 2005 14:00, Nix wrote: >> At any rate, specifying the paths explicitly won't do any harm >> anywhere. :) > On 2.6 we just removed -L/usr/lib from the linker command line because of > some warnings. EXP

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Sat, 14 May 2005, Jeff Dike announced authoritatively: > On Fri, May 13, 2005 at 10:18:22PM +0100, Nix wrote: >> Now on a 2.4 UML that works fine (well, obviously *that* line panics >> because it doesn't specify a root fs. But the longer line works, and in >> 2.6 both

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Sat, 14 May 2005, [EMAIL PROTECTED] prattled cheerily: > On Saturday 14 May 2005 13:00, Nix wrote: >> On Sat, 14 May 2005, [EMAIL PROTECTED] whispered secretively: >> > On Saturday 14 May 2005 02:00, Nix wrote: >> >> On Sat, 14 May 2005, [EMAIL PROTECTED] sp

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Sat, 14 May 2005, [EMAIL PROTECTED] whispered secretively: > On Saturday 14 May 2005 02:00, Nix wrote: >> On Sat, 14 May 2005, [EMAIL PROTECTED] spake: >> > What's your host processor? Can you try upgrading to UML 2.6.11-bs5? I've >> > recently fixed

Re: [uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

On Sat, 14 May 2005, [EMAIL PROTECTED] spake: > On Friday 13 May 2005 23:18, Nix wrote: >> Checking for /proc/mm...found >> Checking for the skas3 patch in the host...found >> [dead] > add stderr=1 to make it more verbose. Hm, nothing new appears. >> The

[uml-user] user-mode-linux 2.4.27 -> 2.6.11 woes

So I'm trying to migrate from 2.4 to 2.6 across the board. I've been at 2.6 everywhere but UML for some time, but this problem is defeating me. I'm running UML on a GCC-3.4.3-compiled 2.6.11.7-skas3-v8-rc5 host (to be .9 when next I reboot) running glibc-2.3.5, like this (simplified from longer li

Re: [uml-user] tls patch

On Tue, 10 May 2005, [EMAIL PROTECTED] wrote: > On Tuesday 10 May 2005 17:23, Nix wrote: >> Indeed: configuring glibc with --without-__thread, putting the result in >> /lib/tls, and linking its dynamic linker into /lib yields an NPTL >> installation that works on i586 witho

Re: [uml-user] tls patch

On Tue, 10 May 2005, [EMAIL PROTECTED] mused: > Exporting in the environment LD_ASSUME_KERNEL=2.4.1 should also work, and is > a > bit more reliable in theory (because a distro *could* compile /lib/tls to > even work on i486, I've heard; since no distro does it, this remains > theory). Indeed

Re: [uml-user] uml and iptables.

On Fri, 1 Apr 2005, Jim Carter spake: > A maximally paranoid sysop will disable module loading, but this gives only > a small benefit in security, because having done a root exploit the hacker > can write nefarious code into /dev/kmem, as easily as he can load an > inimical module or install a h

Re: [uml-user] UML scheduling

On Thu, 24 Feb 2005, Gordon Russell said: > monitor that by creating the swap with sparse blocks and using du. Last I tried (in 2.4), sparse swapfiles were a no-no. -- > ...Hires Root Beer... What we need these days is a stable, fast, anti-aliased root beer with dynamic shading. Not that you can

Re: [uml-user] re: peculiar intermittent network seizures with uml-2.4.27-1 on 2.4.28

On Fri, 17 Dec 2004, roland whispered secretively: > maybe this isn`t a networking problem but a problem with the UML process > itself ? > Keep in mind that a UML is a program which runs in userspace - and it`s no > realtime > application. So, under certain circumstances the response of that appl

[uml-user] peculiar intermittent network seizures with uml-2.4.27-1 on 2.4.28

Ever since upgrading to the 2.4.27-1 UML-applied-to-2.4.28 (from 2.4.24-2-applied-to-2.4.26), I've been seeing decidedly peculiar network interaction. Notably, the UML ceases responding to network packets for up to a minute and a half at a time, then restarts again just as quietly: the console-on

Re: [uml-user] re: peculiar intermittent network seizures with uml-2.4.27-1 on 2.4.28

las, I don't think it's that simple :( My arp tables are simple; virtually everything has the MAC of the ADSL router. Every packet going out of the wrong interface gets syslogged, and I see no such syslog messages, just heaps of NFS timeout messages. (See <http://www.esperi.demon.c