From mboxrd@z Thu Jan 1 00:00:00 1970 From: Darryl Veitch Subject: Re: RADclock on Xen (was Re: [PATCH] time-xen : Reset monotonic time when sync up time from dom0 to domU) Date: Thu, 21 Oct 2010 14:36:11 +1100 Message-ID: <63C6DF1D-933B-47CA-8FDF-9E15095740DE@unimelb.edu.au> References: <9C56B6AEEB2D60488B29B880DF7E53BE03C1B2C6B6@SSDEXCH2.websense.com> <4CB4672E020000780001C752@vpn.id2.novell.com> <0b6cc1b3-f838-482e-89dc-6f19cc8c47cc@default> <20101013161657.GJ4007@whitby.uk.xensource.com> <4CB5E2C3.9010500@goop.org> <20101014090737.GK4007@whitby.uk.xensource.com> <4CB793CE.50907@goop.org> <20101015083937.GM4007@whitby.uk.xensource.com> <5869AFE5-6E86-46D8-8817-98DCAF39F7FC@unimelb.edu.au> <4CB8E1E1.6080906@goop.org> Mime-Version: 1.0 (Apple Message framework v1081) Content-Type: multipart/mixed; boundary="===============0098287955==" Return-path: In-Reply-To: <4CB8E1E1.6080906@goop.org> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jeremy Fitzhardinge Cc: Dan Magenheimer , xen-devel@lists.xensource.com, Keir Fraser , Tim Deegan , Shunli Yi , Jan Beulich , Saipu Liu , Hang Du , Timothy Broomhead , Julien Ridoux List-Id: xen-devel@lists.xenproject.org --===============0098287955== Content-Type: multipart/alternative; boundary=Apple-Mail-23-87326388 --Apple-Mail-23-87326388 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii On 16/10/2010, at 10:21 AM, Jeremy Fitzhardinge wrote: > On 10/15/2010 05:10 AM, Darryl Veitch wrote: >> Hi Tim, >>=20 >> good to hear from you. Dom0 will need the usual RADclock patches >> which now integrate the Xen support. These supply two main things: =20= >> i) kernel level raw timestamping (ie reading of the chosen counter) >> of packet arrivals and departures in `good but safe' locations and an >> ability to pass those timestamps up (works transparently with pcap, >> and doesn't affect its normal operation) >=20 > Are you trying to get these changes into mainline Linux? We are intending to do this, but have not pushed comprehensively yet, = though we hope to start soon after we have the adoption into FreeBSD = finalised (theoretically around end-2010). We have had some discussions = with John Stultz in preparation. Would you be interested in being = involved in encouraging adoption here ? :-) =20 >=20 >> ii) the ability to access the raw counter of choice from both kernel >> and user context >=20 > The kernel already has this info, and I have patches for mapping it = into > userspace. Or would a syscall suffice? We use syscall's currently. Darryl ----------------------------------------------- Darryl Veitch Sabbatical Address: TREC INRIA 23 Avenue D'Italie 75214 Paris Cedex 13, France=20 Tel: +33 (0)1 39 63 55 40 & Technicolor Paris Research Lab 1, rue Jeanne d'Arc 92443 Issy-les-Moulineaux Cedex, France Tel: +33 (0)1 41 86 52 19 Global: dveitch AT unimelb edu au http://www.cubinlab.ee.unimelb.edu.au/~darryl/ AIM id: darrylveitch Skype: darrylveitch --Apple-Mail-23-87326388 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=us-ascii
On 10/15/2010 05:10 AM, Darryl Veitch = wrote:
Hi Tim,

good to hear = from you.  Dom0 will need the usual RADclock = patches
which now integrate = the Xen support. These supply two main things: =  
i)  kernel level = raw timestamping (ie reading of the chosen = counter)
of packet arrivals = and departures in `good but safe' locations and = an
ability to pass those = timestamps up (works transparently with = pcap,
and doesn't affect its = normal operation)

Are you trying to get these = changes into mainline Linux?

We are = intending to do this, but have not pushed comprehensively yet, though we = hope to start soon after we have the adoption into FreeBSD finalised = (theoretically around end-2010).  We have had some discussions with = John Stultz in preparation.   Would you be interested in being = involved in encouraging adoption here ? :-) =  



ii)  the ability = to access the raw counter of choice from both = kernel
and user = context

The kernel already has this info, and I have = patches for mapping it into
userspace.  Or would a syscall = suffice?

We use syscall's = currently.

Darryl



Darryl = Veitch
Sabbatical = AddressTREC = INRIA
23 Avenue D'Italie
75214 Paris Cedex = 13, Tel:  +33 (0)1 39 63 55 = 40
&
Technicolor Paris Research Lab
1, rue Jeanne = d'Arc
92443 Issy-les-Moulineaux Cedex, = France
Tel:  = +33 (0)1 41 86 52 = 19

dveitch = AT edu = au
AIM id:  = darrylveitch
Skype:  = darrylveitch










= --Apple-Mail-23-87326388-- --===============0098287955== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel --===============0098287955==--