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

Re: [Xen-devel] [PATCH] vm_event: make sure the domain is paused in key domctls


  • To: Tamas K Lengyel <tamas@xxxxxxxxxxxxx>
  • From: Razvan Cojocaru <rcojocaru@xxxxxxxxxxxxxxx>
  • Date: Thu, 28 Jan 2016 23:05:05 +0200
  • Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Keir Fraser <keir@xxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>, Xen-devel <xen-devel@xxxxxxxxxxxxx>
  • Comment: DomainKeys? See http://domainkeys.sourceforge.net/
  • Delivery-date: Thu, 28 Jan 2016 21:05:23 +0000
  • Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=default; d=bitdefender.com; b=PMDI6RC5NOd9FXzSnElWk/+YNVubLydNCXiyJI5XfFVsIf3TSQLSVj/9TsOpIKZKiaKXuuTYDrbWAcZAUk+bQCYnRjznJbikfw5szyeuHYde4GmQnQX6LJT0LFrKrtTqupEWgA0hJeaR/MlB0+l26M7Z2ZXr7S6BfM6HYv17d6S9tBXWnHuRY3cqVHF208SOwg4gwLEDO13qJZlLSl7lbkIUh/stOuwHkVIL549JKF7LZJoMTnKmKkiIDVA4vk9r5OurJdmNsujVwmczkQ9BKttMoEaWw27DPunRB7k3qQJmxlHhY2nmA9PnXGSX2kspV6I6Hsmi8UizEZCucpnHNg==; h=Received:Received:Received:Received:Received:Subject:To:References:Cc:From:X-Enigmail-Draft-Status:Message-ID:Date:User-Agent:MIME-Version:In-Reply-To:Content-Type:Content-Transfer-Encoding:X-BitDefender-Scanner:X-BitDefender-Spam:X-BitDefender-SpamStamp:X-BitDefender-CF-Stamp;
  • List-id: Xen developer discussion <xen-devel.lists.xen.org>

On 01/28/2016 10:09 PM, Tamas K Lengyel wrote:
> On Thu, Jan 28, 2016 at 6:52 AM, Razvan Cojocaru
> <rcojocaru@xxxxxxxxxxxxxxx <mailto:rcojocaru@xxxxxxxxxxxxxxx>> wrote:
> 
>     This patch pauses the domain for all writes through the 'ad'
>     pointer in monitor_domctl(), defers a domain_unpause() call until
>     after the CRs are updated for the MONITOR_EVENT_WRITE_CTRLREG
>     case, and makes sure that the domain is paused for both vm_event
>     enable and disable cases in vm_event_domctl().
>     Thanks go to Andrew Cooper for his review and suggestions.
> 
> 
> For vm_event_enable the domain is already paused by libxc before the
> domctl is issued. I don't see a problem in doing another pause in Xen,
> but given we have XSA-99, just doing this pause in Xen would not be
> enough. So is it really necessary/fixes anything?

This isn't about XSA-99, the problem here is related to my previous
patch "x86 vm_event: reset monitor in vm_event_cleanup_domain()". While
that improves matters and greatly reduces the chances of crashes due to
hvm_msr_write_intercept() or hvm_set_crX() dereferencing a NULL
v->arch.vm_event that's assumed to be OK, when the corresponding
v->domain->arch.monitor is non-zero, the foolproof way is to make sure
that functions such as vm_event_cleanup_domain() are always being called
only while the domain has been paused. So there should be a
domain_pause() call somewhere on the call path before that.

In fact, any writes to domain-scope variables that can affect the way a
guest runs (while the guest is running) should ideally be done with the
domain paused, hence the rest of the changes in the patch.


Thanks,
Razvan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

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