Re: [PATCH RFC 0/4] tools: Upstream blktap "2.5" as an external repo

[George Dunlap <george.dunlap@eu.citrix.com>]

On 03/26/2015 05:12 PM, Andrew Cooper wrote:
> On 26/03/15 17:05, George Dunlap wrote:
>> On 03/26/2015 04:25 PM, Felipe Franciosi wrote:
>>>> Another thing I'd like to explore (since this took all of about an
>>>> afternoon to get
>>>> working) is what it would take to switch to using
>>>> blktap3 instead.  As I understand from my conversations with the
>>>> XenServer
>>>> team, they use a kernel module in XenServer when mounting an image
>>>> in dom0
>>>> for scalability reasons; but there's no reason XenProject need to do
>>>> the same
>>>> thing.
>>> Today, to access virtual disks in dom0 we use the blktap2 kernel
>>> module that Jonathan Ludlam mentioned. This provides a block device
>>> (in /dev/xen/blktap-2/tapdev<minor>). In the (somewhat distant) past,
>>> we actually had blkfront in dom0.
>> OK -- so the XenServer blktap2 kernel module is definitely being
>> maintained.  That's good to know.
>>
>>> What is also needed to get blktap3 working is a gntdev supporting
>>> grant copy. I believe this is the order they are applied in 3.10:
>>> https://github.com/xenserver/linux-3.x.pg/blob/master/master/0001-xen-install-xen-gntdev.h-and-xen-gntalloc.h.patch
>>>
>>> https://github.com/xenserver/linux-3.x.pg/blob/master/master/xen-gntdev-grant-copy.patch
>>>
>>> https://github.com/xenserver/linux-3.x.pg/blob/master/master/0002-xen-gntdev-mark-userspace-PTEs-as-special-on-x86-PV-.patch
>>>
>>> https://github.com/xenserver/linux-3.x.pg/blob/master/master/0003-xen-gntdev-provide-a-set-of-pages-for-the-VMA.patch
>>>
>>>
>>> The _main_ difference between tapdisk2 and 3 is that tapdisk3 can
>>> connect directly to blkfront. It does all I/O via grant copies which
>>> has some implications in the way memory is handled in dom0.
>> Right, I didn't realize blktap3 also required kernel changes.  Are those
>> reasonably upstream-able, and is there any plan by the XenServer team to
>> upstream them?
> 
> They are all in Linux 3.19 iirc.  These are not patches for blktap3
> specially; they are all patches fixing the use of userspace grant maps
> with /dev/xen/gnttab, which never functioned correctly in the past.

OK, good.  So with my XenProject hat on, nothing needs to be done per se
to use blktap3; we just need to tell people to use linux >= 3.19.  (And
make sure there's a robust way to detect the necessary kernel support
before trying to use blktap3.)

With my CentOS hat on, if I want to use an older version of Linux, I
need to import / port those patches to whatever version I'm using; but
that's not directly relevant to this discussion.

 -George