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

Re: [Xen-devel] AFS-based VBD backend

  • To: Ian Pratt <m+Ian.Pratt@xxxxxxxxxxxx>
  • From: Andrew Warfield <andrew.warfield@xxxxxxxxx>
  • Date: Thu, 23 Dec 2004 12:36:30 +0000
  • Cc: Steve Traugott <stevegt@xxxxxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxxx
  • Delivery-date: Thu, 23 Dec 2004 12:38:14 +0000
  • Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references; b=Puy0q3Nvr0rljEmdA4hm7HOzLPLXhSuRBYmdrbdDrN8FlVCty/uD0Q0/MMcRXXWlu7uzI2xc/O3YQt2lVF1zSGHjWBG7Yow4An1VcJli9D3rfMK0woqWSBnbmct92SUI3SZmjVEFsG1k0JMzKRC6FdMaQB4qGScjadrxqEGx8u8=
  • List-id: List for Xen developers <xen-devel.lists.sourceforge.net>

> The blocktap backend driver is what you want. I'm not sure of it's
> current state, but the plan is to enable you to terminate a blk device
> channel in user-space.

The driver currently does this.  There is a rather dated version of it
in the unstable tree...  It lets you either terminate a blkif
connection in userspace of a tap domain, or intercept in-flight
requests by proxying them to user-space before they are passed along
to a normal backend.  Data pages provided by the front end are mapped
up to user space by the tap driver, so there is no copying along that

I've got a newer and more, um, robust ;) version of the blktap stuff
in my repository... it's a bit tied up with some of the other device
channel updates that we have been looking at lately though -- so it's
not in a state to integrate directly to the public trees.  I'd imagine
that much of this stuff should be ready to go into unstable in the
next few weeks...  hopefully along with a nettap as well.

Steve, if you need the tap more urgently than this, I can likely sort
out an interim version.  If you can wait until early january though,
that might be best.

> A future revision of blocktap could use kiovec's to avoid having to copy
> data into user space, and thus would give good performance.

arranging direct i/o to the mapped data pages is clearly the thing to
do.  I'll take a look at the kvec stuff over the next few weeks as


SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now. 
Xen-devel mailing list



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