[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] Who sets my p2m entries to INVALID?

  • To: xen-devel <Xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Jacob Gorm Hansen <jacobg@xxxxxxx>
  • Date: Fri, 27 Apr 2007 17:17:30 +0200
  • Delivery-date: Fri, 27 Apr 2007 08:17:37 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:subject:from:to:content-type:date:message-id:mime-version:x-mailer:content-transfer-encoding:sender; b=IbwFuPPO13rlMOjop/7MWFf3nUTrT+mcO1nla0JFOWv2a9w2lmvgmpkhOl7KXTVcDwtYwz0KXBmdxhfdI94zLSg3R+ytDgOt4D5gGZODuKQbiwP655y4Y3ijRmAtQtYVmzZ1wJlKHNM+9NLGjXt79akJWzRL8AO9bOidzUx7g3g=
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>


I am having a strange problem here, where exactly half my XenLinux PFNs
are recorded as INVALID_P2M_ENTRY in phys_to_machine_mapping. This is
with xen 3.0.3, and only happens on some of my machines, all 2GHz P4's
with 512 megs of memory, 64 of which are given to dom0.

The lupper half of the phys_to_machine array is filled with 0xFFFFFFFF
entries, but a  check for this input value in set_phys_to_machine()
never logs any invalid entries written through that function. The VM
uses the network quite heavily, so I was thinking that the problem could
be due to improper recovery of incoming netfront pages. The problem
seems to occur with both flipping and copying receiver.

On another machine, an Intel(R) Core(TM)2 CPU, with more, memory, and
the exact same binaries for xen, vmlinux0 and vmlinuxU, the problem
never occurs.

Has anyone seen problem before?


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.