Virtualization - May 2019 - [PATCH 01/10] virtio/s390: use vring_create_virtqueue

On 05.05.19 13:15, Cornelia Huck wrote:
> On Sat, 4 May 2019 16:03:40 +0200
> Halil Pasic <pasic at linux.ibm.com> wrote:
> 
>> On Fri, 3 May 2019 16:04:48 -0400
>> "Michael S. Tsirkin" <mst at redhat.com> wrote:
>>
>>> On Fri, May 03, 2019 at 11:17:24AM +0200, Cornelia Huck wrote:  
>>>> On Fri, 26 Apr 2019 20:32:36 +0200
>>>> Halil Pasic <pasic at linux.ibm.com> wrote:
>>>>   
>>>>> The commit 2a2d1382fe9d ("virtio: Add improved queue
allocation API")
>>>>> establishes a new way of allocating virtqueues (as a part
of the effort
>>>>> that taught DMA to virtio rings).
>>>>>
>>>>> In the future we will want virtio-ccw to use the DMA API as
well.
>>>>>
>>>>> Let us switch from the legacy method of allocating
virtqueues to
>>>>> vring_create_virtqueue() as the first step into that
direction.
>>>>>
>>>>> Signed-off-by: Halil Pasic <pasic at linux.ibm.com>
>>>>> ---
>>>>>  drivers/s390/virtio/virtio_ccw.c | 30
+++++++++++-------------------
>>>>>  1 file changed, 11 insertions(+), 19 deletions(-)  
>>>>
>>>> Reviewed-by: Cornelia Huck <cohuck at redhat.com>
>>>>
>>>> I'd vote for merging this patch right away for 5.2.  
>>>
>>> So which tree is this going through? mine?
>>>   
>>
>> Christian, what do you think? If the whole series is supposed to go in
>> in one go (which I hope it is), via Martin's tree could be the
simplest
>> route IMHO.
> 
> 
> The first three patches are virtio(-ccw) only and the those are the ones
> that I think are ready to go.
> 
> I'm not feeling comfortable going forward with the remainder as it
> stands now; waiting for some other folks to give feedback. (They are
> touching/interacting with code parts I'm not so familiar with, and lack
> of documentation, while not the developers' fault, does not make it
> easier.)
> 
> Michael, would you like to pick up 1-3 for your tree directly? That
> looks like the easiest way.
Agreed. Michael please pick 1-3.
We will continue to review 4- first and then see which tree is best.

On Tue, 7 May 2019 15:58:12 +0200
Christian Borntraeger <borntraeger at de.ibm.com> wrote:
> 
> 
> On 05.05.19 13:15, Cornelia Huck wrote:
> > On Sat, 4 May 2019 16:03:40 +0200
> > Halil Pasic <pasic at linux.ibm.com> wrote:
> > 
> >> On Fri, 3 May 2019 16:04:48 -0400
> >> "Michael S. Tsirkin" <mst at redhat.com> wrote:
> >>
> >>> On Fri, May 03, 2019 at 11:17:24AM +0200, Cornelia Huck wrote:
> >>>> On Fri, 26 Apr 2019 20:32:36 +0200
> >>>> Halil Pasic <pasic at linux.ibm.com> wrote:
> >>>>   
> >>>>> The commit 2a2d1382fe9d ("virtio: Add improved
queue allocation API")
> >>>>> establishes a new way of allocating virtqueues (as a
part of the effort
> >>>>> that taught DMA to virtio rings).
> >>>>>
> >>>>> In the future we will want virtio-ccw to use the DMA
API as well.
> >>>>>
> >>>>> Let us switch from the legacy method of allocating
virtqueues to
> >>>>> vring_create_virtqueue() as the first step into that
direction.
> >>>>>
> >>>>> Signed-off-by: Halil Pasic <pasic at
linux.ibm.com>
> >>>>> ---
> >>>>>  drivers/s390/virtio/virtio_ccw.c | 30
+++++++++++-------------------
> >>>>>  1 file changed, 11 insertions(+), 19 deletions(-)  
> >>>>
> >>>> Reviewed-by: Cornelia Huck <cohuck at redhat.com>
> >>>>
> >>>> I'd vote for merging this patch right away for 5.2.  
> >>>
> >>> So which tree is this going through? mine?
> >>>   
> >>
> >> Christian, what do you think? If the whole series is supposed to
go in
> >> in one go (which I hope it is), via Martin's tree could be the
simplest
> >> route IMHO.
> > 
> > 
> > The first three patches are virtio(-ccw) only and the those are the
ones
> > that I think are ready to go.
> > 
> > I'm not feeling comfortable going forward with the remainder as it
> > stands now; waiting for some other folks to give feedback. (They are
> > touching/interacting with code parts I'm not so familiar with, and
lack
> > of documentation, while not the developers' fault, does not make
it
> > easier.)
> > 
> > Michael, would you like to pick up 1-3 for your tree directly? That
> > looks like the easiest way.
> 
> Agreed. Michael please pick 1-3.
> We will continue to review 4- first and then see which tree is best.
Thanks Christian!

Guys, I broke my right arm on last Thursday (2nd may). You may
have noticed that I was not as responsive as I'm supposed to be.
Unfortunately this less than responsiveness is about to persist for a
couple more weeks. Fortunate the guys from IBM, and chiefly Michael
Mueller is going to help me drive this -- thanks Michael! Due to this,
I would generally prefer doing as few changes to this series as
necessary, and deferring as many of the beautifying to patches on top
(possibly authored by somebody else) as possible.

Regards,
Halil

On Tue, 7 May 2019 15:58:12 +0200
Christian Borntraeger <borntraeger at de.ibm.com> wrote:
> On 05.05.19 13:15, Cornelia Huck wrote:
> > On Sat, 4 May 2019 16:03:40 +0200
> > Halil Pasic <pasic at linux.ibm.com> wrote:
> >   
> >> On Fri, 3 May 2019 16:04:48 -0400
> >> "Michael S. Tsirkin" <mst at redhat.com> wrote:
> >>  
> >>> On Fri, May 03, 2019 at 11:17:24AM +0200, Cornelia Huck wrote:
> >>>> On Fri, 26 Apr 2019 20:32:36 +0200
> >>>> Halil Pasic <pasic at linux.ibm.com> wrote:
> >>>>     
> >>>>> The commit 2a2d1382fe9d ("virtio: Add improved
queue allocation API")
> >>>>> establishes a new way of allocating virtqueues (as a
part of the effort
> >>>>> that taught DMA to virtio rings).
> >>>>>
> >>>>> In the future we will want virtio-ccw to use the DMA
API as well.
> >>>>>
> >>>>> Let us switch from the legacy method of allocating
virtqueues to
> >>>>> vring_create_virtqueue() as the first step into that
direction.
> >>>>>
> >>>>> Signed-off-by: Halil Pasic <pasic at
linux.ibm.com>
> >>>>> ---
> >>>>>  drivers/s390/virtio/virtio_ccw.c | 30
+++++++++++-------------------
> >>>>>  1 file changed, 11 insertions(+), 19 deletions(-)    
> >>>>
> >>>> Reviewed-by: Cornelia Huck <cohuck at redhat.com>
> >>>>
> >>>> I'd vote for merging this patch right away for 5.2.
> >>>
> >>> So which tree is this going through? mine?
> >>>     
> >>
> >> Christian, what do you think? If the whole series is supposed to
go in
> >> in one go (which I hope it is), via Martin's tree could be the
simplest
> >> route IMHO.  
> > 
> > 
> > The first three patches are virtio(-ccw) only and the those are the
ones
> > that I think are ready to go.
> > 
> > I'm not feeling comfortable going forward with the remainder as it
> > stands now; waiting for some other folks to give feedback. (They are
> > touching/interacting with code parts I'm not so familiar with, and
lack
> > of documentation, while not the developers' fault, does not make
it
> > easier.)
> > 
> > Michael, would you like to pick up 1-3 for your tree directly? That
> > looks like the easiest way.  
> 
> Agreed. Michael please pick 1-3.
> We will continue to review 4- first and then see which tree is best.
Michael, please let me know if you'll pick directly or whether I should
post a series.

[Given that the patches are from one virtio-ccw maintainer and reviewed
by the other, picking directly would eliminate an unnecessary
indirection :)]

On Fri, May 10, 2019 at 04:07:44PM +0200, Cornelia Huck
wrote:
> On Tue, 7 May 2019 15:58:12 +0200
> Christian Borntraeger <borntraeger at de.ibm.com> wrote:
> 
> > On 05.05.19 13:15, Cornelia Huck wrote:
> > > On Sat, 4 May 2019 16:03:40 +0200
> > > Halil Pasic <pasic at linux.ibm.com> wrote:
> > >   
> > >> On Fri, 3 May 2019 16:04:48 -0400
> > >> "Michael S. Tsirkin" <mst at redhat.com>
wrote:
> > >>  
> > >>> On Fri, May 03, 2019 at 11:17:24AM +0200, Cornelia Huck
wrote:
> > >>>> On Fri, 26 Apr 2019 20:32:36 +0200
> > >>>> Halil Pasic <pasic at linux.ibm.com> wrote:
> > >>>>     
> > >>>>> The commit 2a2d1382fe9d ("virtio: Add
improved queue allocation API")
> > >>>>> establishes a new way of allocating virtqueues
(as a part of the effort
> > >>>>> that taught DMA to virtio rings).
> > >>>>>
> > >>>>> In the future we will want virtio-ccw to use the
DMA API as well.
> > >>>>>
> > >>>>> Let us switch from the legacy method of
allocating virtqueues to
> > >>>>> vring_create_virtqueue() as the first step into
that direction.
> > >>>>>
> > >>>>> Signed-off-by: Halil Pasic <pasic at
linux.ibm.com>
> > >>>>> ---
> > >>>>>  drivers/s390/virtio/virtio_ccw.c | 30
+++++++++++-------------------
> > >>>>>  1 file changed, 11 insertions(+), 19
deletions(-)
> > >>>>
> > >>>> Reviewed-by: Cornelia Huck <cohuck at
redhat.com>
> > >>>>
> > >>>> I'd vote for merging this patch right away for
5.2.
> > >>>
> > >>> So which tree is this going through? mine?
> > >>>     
> > >>
> > >> Christian, what do you think? If the whole series is supposed
to go in
> > >> in one go (which I hope it is), via Martin's tree could
be the simplest
> > >> route IMHO.  
> > > 
> > > 
> > > The first three patches are virtio(-ccw) only and the those are
the ones
> > > that I think are ready to go.
> > > 
> > > I'm not feeling comfortable going forward with the remainder
as it
> > > stands now; waiting for some other folks to give feedback. (They
are
> > > touching/interacting with code parts I'm not so familiar
with, and lack
> > > of documentation, while not the developers' fault, does not
make it
> > > easier.)
> > > 
> > > Michael, would you like to pick up 1-3 for your tree directly?
That
> > > looks like the easiest way.  
> > 
> > Agreed. Michael please pick 1-3.
> > We will continue to review 4- first and then see which tree is best.
> 
> Michael, please let me know if you'll pick directly or whether I should
> post a series.
> 
> [Given that the patches are from one virtio-ccw maintainer and reviewed
> by the other, picking directly would eliminate an unnecessary
> indirection :)]
picked them