|Dutch Meyer||Feb 18, 2009 2:48 pm|
|Dutch Meyer||Mar 26, 2009 11:08 pm|
|Dutch Meyer||Apr 2, 2009 11:14 pm|
|Ben Guthro||Apr 3, 2009 7:52 am|
|Ben Guthro||Apr 3, 2009 7:54 am|
|Dutch Meyer||Apr 3, 2009 8:27 am|
|Ben Guthro||Apr 3, 2009 9:58 am||.patch|
|Dutch Meyer||Apr 3, 2009 1:24 pm|
|Dutch Meyer||Apr 3, 2009 5:25 pm|
|Keir Fraser||May 18, 2009 6:26 pm|
|Dutch Meyer||May 22, 2009 11:46 am|
|Subject:||[Xen-devel] [5 Patches] Synchronize blktap with citrix blktap2.|
|From:||Dutch Meyer (dme...@cs.ubc.ca)|
|Date:||Feb 18, 2009 2:48:52 pm|
This is an updated patch set to the one I posted a few months back.
These patches contain a completely rewritten blktap implementation and are an open source release of what Citrix intends to use in future releases of XenServer. They contain a number of bug fixes from my previous patchset.
They also contain Citrix's implementation of the VHD image format. VHD is what XenServer uses to store file-based images, and this code is considerably more robust and efficient than the qcow implementation that is in the tree today.
Benefits to blktap2 over the old version of blktap:
* Isolation from xenstore - Blktap devices are now created directly on the linux dom0 command line, rather than being spawned in response to XenStore events. This is handy for debugging, makes blktap generally easier to work with, and is a step toward a generic user-level block device implementation that is not Xen-specific.
* Improved tapdisk infrastructure: simpler request forwarding, new request scheduler, request merging, more efficient use of AIO.
* Improved tapdisk error handling and memory management. No allocations on the block data path, IO retry logic to protect guests transient block device failures. This has been tested and is known to work on weird environments such as NFS soft mounts.
* Pause and snapshot of live virtual disks (see xmsnap script).
* VHD support. The VHD code in this release has been rigorously tested, and represents a very mature implementation of the VHD image format.
* No more duplication of mechanism with blkback. The blktap kernel module has changed dramatically from the original blktap. Blkback is now always used to talk to Xen guests, blktap just presents a Linux gendisk that blkback can export. This is done while preserving the zero-copy data path from domU to physical device.
These patches deprecate the old blktap code, which can hopefully be removed from the tree completely at some point in the future.
Two blktap Patches (the xen tree): http://www.cs.ubc.ca/~dmeyer/blktap2/01-deprecate-blktap http://www.cs.ubc.ca/~dmeyer/blktap2/02-add-blktap2
And 3 kernel module patches (the linux tree): http://www.cs.ubc.ca/~dmeyer/blktap2/01-blktap-patch http://www.cs.ubc.ca/~dmeyer/blktap2/02-blkback-pagemap-patch http://www.cs.ubc.ca/~dmeyer/blktap2/03-blktap2-patch
In addition to the above patches, the (external) ioemu git repository currently has some very heavy dependencies on blktap, and does not like the change of the blktap directory name to deprecate that code. The following patch needs to be applied to fix ioemu:
Ideally we can come up with a better integration between blktap2 and ioemu than what is currently in place with the old tapdisk code. For instance, blktap2 makes it possible to just treat the tapdisk-based devices as raw devices for qemu, severing these (now unnecessary) dependencies.