From: Matthew Rosato <mjrosato@linux.ibm.com>
To: Cornelia Huck <cohuck@redhat.com>, Tony Krowiak <akrowiak@linux.ibm.com>
Cc: Alex Williamson <alex.williamson@redhat.com>,
kvm@vger.kernel.org, libvir-list@redhat.com,
Halil Pasic <pasic@linux.ibm.com>
Subject: Re: [PATCH RFC 0/1] mdevctl: further config for vfio-ap
Date: Thu, 13 Jun 2019 10:25:32 -0400 [thread overview]
Message-ID: <9e3a3955-f073-7c1c-9987-a1c7c3073a55@linux.ibm.com> (raw)
In-Reply-To: <20190613155418.7fa17ed2.cohuck@redhat.com>
On 6/13/19 9:54 AM, Cornelia Huck wrote:
> On Fri, 7 Jun 2019 14:30:48 -0400
> Tony Krowiak <akrowiak@linux.ibm.com> wrote:
>
>> On 6/7/19 10:56 AM, Cornelia Huck wrote:
>>> On Thu, 6 Jun 2019 12:45:29 -0400
>>> Matthew Rosato <mjrosato@linux.ibm.com> wrote:
>>>
>>>> On 6/6/19 10:44 AM, Cornelia Huck wrote:
>>>>> This patch adds a very rough implementation of additional config data
>>>>> for mdev devices. The idea is to make it possible to specify some
>>>>> type-specific key=value pairs in the config file for an mdev device.
>>>>> If a device is started automatically, the device is stopped and restarted
>>>>> after applying the config.
>>>>>
>>>>> The code has still some problems, like not doing a lot of error handling
>>>>> and being ugly in general; but most importantly, I can't really test it,
>>>>> as I don't have the needed hardware. Feedback welcome; would be good to
>>>>> know if the direction is sensible in general.
>>>>
>>>> Hi Connie,
>>>>
>>>> This is very similar to what I was looking to do in zdev (config via
>>>> key=value pairs), so I like your general approach.
>>>>
>>>> I pulled your code and took it for a spin on an LPAR with access to
>>>> crypto cards:
>>>>
>>>> # mdevctl create-mdev `uuidgen` matrix vfio_ap-passthrough
>>>> # mdevctl set-additional-config <uuid> ap_adapters=0x4,0x5
>>>> # mdevctl set-additional-config <uuid> ap_domains=0x36
>>>> # mdevctl set-additional-config <uuid> ap_control_domains=0x37
>>>>
>>>> Assuming all valid inputs, this successfully creates the appropriate
>>>> mdev and what looks to be a valid mdevctl.d entry. A subsequent reboot
>>>> successfully brings the same vfio_ap-passthrough device up again.
>>>
>>> Cool, thanks for checking!
>>
>> I also confirmed this. I realize this is a very early proof of concept,
>> if you will, but error handling could be an interesting endeavor in
>> the case of vfio_ap given the layers of configuration involved;
>
> I agree; that's why I mostly ignored it for now :)
>
>> for
>> example:
>>
>> * The necessity for the vfio_ap module to be installed
>> * The necessity that the /sys/bus/ap/apmask and /sys/bus/ap/aqmask must
>> be appropriately configured
>
> What do you think about outsourcing that configuration to some
> s390-specific tool (probably something in s390-tools)? While we can
> (and should) rely on driverctl for vfio-ccw, this does not look like
> something that can be easily served by some generic tooling.
>
+1 mdevctl should stick to manipulation of mdevs. I think any config
layer above that isn't directly modifying mdev parameters, like in this
case the kernel ap/aq masks, should be separate (and for vfio_ap I agree
s390-tools is the right place for this -- it's on my todo list)
>>
>>>
>>>>
>>>> Matt
>>>>
>>>>>
>>>>> Also available at
>>>>>
>>>>> https://github.com/cohuck/mdevctl conf-data
>>>>>
>>>>> Cornelia Huck (1):
>>>>> allow to specify additional config data
>>>>>
>>>>> mdevctl.libexec | 25 ++++++++++++++++++++++
>>>>> mdevctl.sbin | 56 ++++++++++++++++++++++++++++++++++++++++++++++++-
>>>>> 2 files changed, 80 insertions(+), 1 deletion(-)
>>>>>
>>>>
>>>
>>
>
next prev parent reply other threads:[~2019-06-13 15:02 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-06 14:44 [PATCH RFC 0/1] mdevctl: further config for vfio-ap Cornelia Huck
2019-06-06 14:44 ` [PATCH RFC 1/1] allow to specify additional config data Cornelia Huck
2019-06-06 15:32 ` Alex Williamson
2019-06-06 16:15 ` Alex Williamson
2019-06-07 18:26 ` Tony Krowiak
2019-06-07 20:03 ` Alex Williamson
2019-06-11 14:19 ` Tony Krowiak
2019-06-13 14:18 ` Cornelia Huck
2019-06-14 13:24 ` Tony Krowiak
2019-06-18 15:11 ` Cornelia Huck
2019-06-13 15:56 ` Halil Pasic
2019-06-06 16:45 ` [PATCH RFC 0/1] mdevctl: further config for vfio-ap Matthew Rosato
2019-06-07 14:56 ` Cornelia Huck
2019-06-07 18:30 ` Tony Krowiak
2019-06-13 13:54 ` Cornelia Huck
2019-06-13 14:25 ` Matthew Rosato [this message]
2019-06-14 13:36 ` Tony Krowiak
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=9e3a3955-f073-7c1c-9987-a1c7c3073a55@linux.ibm.com \
--to=mjrosato@linux.ibm.com \
--cc=akrowiak@linux.ibm.com \
--cc=alex.williamson@redhat.com \
--cc=cohuck@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=libvir-list@redhat.com \
--cc=pasic@linux.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).