Virtualization - Sep 2022 - [PATCH v1] vdpa/ifcvf: avoid waste ioremap area of vdpa hot migration

On Fri, Sep 23, 2022 at 12:15 PM Zhu, Lingshan <lingshan.zhu at intel.com>
wrote:
>
>
>
> On 9/23/2022 12:09 PM, Jason Wang wrote:
> > On Fri, Sep 23, 2022 at 11:53 AM Angus Chen <angus.chen at
jaguarmicro.com> wrote:
> >> The array capacity should be queuepair,
> > If this is true, we need a better name for the ifcvf_vring_lm_cfg
structure.
> Not true AFAIK. And we are re-designing this cap.
>
> Thanks
> Zhu Lingshan
Ok, but I think we still need to keep this for backward compatibility
and need to differ the generations via some Intel specific registers.

Thanks
> >
> > Thanks
> >
> >> and the queuepairs should
> >> be half of IFCVF_MAX_QUEUES without control queue,
> >> or should be (IFCVF_MAX_QUEUES+1)/2 with the control queue.
> >> So the definition of ifcvf_lm_cfg waste some memory,
> >> and it will waste some ioremap area also.
> >>
> >> Fixes: 2ddae773c93b ("vDPA/ifcvf: detect and use the onboard
number of queues directly")
> >> Signed-off-by: Angus Chen <angus.chen at jaguarmicro.com>
> >> ---
> >>   drivers/vdpa/ifcvf/ifcvf_base.h | 2 +-
> >>   1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/vdpa/ifcvf/ifcvf_base.h
b/drivers/vdpa/ifcvf/ifcvf_base.h
> >> index f5563f665cc6..563c04f9b7ac 100644
> >> --- a/drivers/vdpa/ifcvf/ifcvf_base.h
> >> +++ b/drivers/vdpa/ifcvf/ifcvf_base.h
> >> @@ -104,7 +104,7 @@ struct ifcvf_vring_lm_cfg {
> >>
> >>   struct ifcvf_lm_cfg {
> >>          u8 reserved[IFCVF_LM_RING_STATE_OFFSET];
> >> -       struct ifcvf_vring_lm_cfg vring_lm_cfg[IFCVF_MAX_QUEUES];
> >> +       struct ifcvf_vring_lm_cfg
vring_lm_cfg[(IFCVF_MAX_QUEUES+1)/2];
> >>   };
> >>
> >>   struct ifcvf_vdpa_mgmt_dev {
> >> --
> >> 2.17.1
> >>
>