public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Brian Ristuccia <brian@ristuccia.com>
To: linux-kernel@vger.kernel.org
Cc: user-mode-linux-devel@lists.sourceforge.net,
	Alex Pennace <alex@osiris.978.org>,
	Dennis Ristuccia <dennis@osiris.978.org>
Subject: usermode linux hoses 2.2.18-SMP host machine when run from regular user account
Date: Wed, 31 Jan 2001 23:29:44 -0500	[thread overview]
Message-ID: <20010131232944.J16908@osiris.978.org> (raw)

When run from a normal user account with its current working directory on a
NFS filesystem, usermode linux causes the host machine's kernel to enter a
hosed state. No processes (including UML) seem to respond, and the machine
becomes unusable.

I am running 2.2.18 built with SMP and hedrick's IDE patches on a dual
Celeron 366@550 box with 384mB of ram. I have also reproduced the crash on a
single Celeron 366@550 w/ 256mB of ram running 2.2.18-SMP w/ hedrick's IDE
patches, and a dual PPro 166 w/ 256mB of RAM running 2.2.18pre23 unmodified
from the version distributed by Alan Cox. The crash occurs with both the
usermode linux 2.4.0 binary from their web site and the linux 2.4.1 binary
there. I am using the root filesystem image file that comes in the tarball
named deb-package-2.4.0.tar.bz2. The crash comes after the line "INIT:
version 2.78 booting" during the virtual machine's boot sequence. The UML
web site is at http://user-mode-linux.sourceforge.net/

The UML people report that UML doesn't work when being run from an NFS
filesystem: 

  UML doesn't work when it is being run from an NFS filesystem: This seems to
  be a similar situation with the resierfs problem above. Some versions of
  NFS seems not to handle mmap correctly, which UML depends on.  The fix is
  to run UML in a non-NFS directory. The kernel binary and root filesystem
  be be mounted from NFS; UML itself just has to be running in a non-NFS
  directory.

But I can't see how something done in userspace should hang the whole
machine, especially when I set up low hard ulimits to prevent all system
resources from being consumed. The statistics I'm able to gather from
ctrl/shift-scrlock don't seem to indicate memory or process table overuse.

Alt-Fx still switches consoles unless X is running. Ctrl/shift-scrlock still
displays the usual debug stuff. The machine is still pingable and still
accepts tcp/ip connections but you don't get any response from the listening
daemon. Otherwise the box is hung. Ctrl-alt-del won't run the ctrlaltdel
entry in inittab. Hardware reset is required to restore machine to normal
operation.

-- 
Brian Ristuccia
brian@ristuccia.com
bristucc@cs.uml.edu
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

             reply	other threads:[~2001-02-01  4:30 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-02-01  4:29 Brian Ristuccia [this message]
2001-02-01  6:08 ` [uml-devel] usermode linux hoses 2.2.18-SMP host machine when run from regular user account Jeff Dike

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20010131232944.J16908@osiris.978.org \
    --to=brian@ristuccia.com \
    --cc=alex@osiris.978.org \
    --cc=dennis@osiris.978.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=user-mode-linux-devel@lists.sourceforge.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox