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

Re: [Xen-devel] VMX status report. Xen: #17501 & Xen0: #524 -- two new issues


  • To: "Li, Haicheng" <haicheng.li@xxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <keir.fraser@xxxxxxxxxxxxx>
  • Date: Wed, 23 Apr 2008 14:27:56 +0100
  • Delivery-date: Wed, 23 Apr 2008 06:28:19 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcilKK2Z9ZBbRqIESv6ia56nvm3K+AAE32DDAAJq/1o=
  • Thread-topic: [Xen-devel] VMX status report. Xen: #17501 & Xen0: #524 -- two new issues

These issues are now all resolved:

On 23/4/08 13:18, "Keir Fraser" <keir.fraser@xxxxxxxxxxxxx> wrote:

>> 1. Fail to create any HVM guest on IA32PAE host
>> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1221
> 
> Looks like it is due to Kevin Wolf's changeset 17499. Your attempt to go
> back to older tools obviously didn't work -- the backtrace in xend.log comes
> from a line introduced by c/s 17499.

I reverted this changeset.

>> 2. Dom0 hangs at booting if loglvl and guest_loglvl are set as all.
>> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1222
> 
> Well, the obvious answer is to print less drivel in the VT-d code. There's
> no reason at all to print so much useless text, even in a debug build.

I greatly reduced printk() count in vtd code. If this stuff was useful for
debugging (which for the most part it probably wasn't) you might add it back
via an ifdef'ed debug macro.

> Actually a second point is that we should drop characters from the 16kB
> serial buffer rather than spin in the hypervisor waiting for the buffer to
> empty. I'll implement that.

This is now done.

 -- Keir



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

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