RE: [PATCH v2 3/7] vdpa: validate provisioned device features against specified attribute

[Parav Pandit]

> From: Si-Wei Liu <si-wei.liu@xxxxxxxxxx>
> Sent: Thursday, February 2, 2023 4:59 PM
> 
> On 2/1/2023 9:05 PM, Parav Pandit wrote:
> >
> >> From: Si-Wei Liu <si-wei.liu@xxxxxxxxxx>
> >> Sent: Tuesday, January 31, 2023 6:22 PM
> >>
> >> With device feature provisioning, there's a chance for
> >> misconfiguration that the vdpa feature attribute supplied in 'vdpa
> >> dev add' command doesn't get selected on the device_features to be
> >> provisioned. For instance, when a @mac attribute is specified, the
> >> corresponding feature bit _F_MAC in device_features should be set for
> >> consistency. If there's conflict on provisioned features against the attribute,
> it should be treated as an error to fail the ambiguous command.
> >> Noted the opposite is not necessarily true, for e.g. it's okay to
> >> have _F_MAC set in device_features without providing a corresponding
> >> @mac attribute, in which case the vdpa vendor driver could load
> >> certain default value for attribute that is not explicitly specified.
> >>
> >> Generalize this check in vdpa core so that there's no duplicate code
> >> in each vendor driver.
> >>
> >> Signed-off-by: Si-Wei Liu <si-wei.liu@xxxxxxxxxx>
> >> ---
> >>   drivers/vdpa/vdpa.c | 18 ++++++++++++++++++
> >>   1 file changed, 18 insertions(+)
> >>
> >> diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c index
> >> 21c8aa3..1eba978
> >> 100644
> >> --- a/drivers/vdpa/vdpa.c
> >> +++ b/drivers/vdpa/vdpa.c
> >> @@ -601,8 +601,26 @@ static int vdpa_nl_cmd_dev_add_set_doit(struct
> >> sk_buff *skb, struct genl_info *i
> >>   		config.mask |=
> >> BIT_ULL(VDPA_ATTR_DEV_NET_CFG_MAX_VQP);
> >>   	}
> >>   	if (nl_attrs[VDPA_ATTR_DEV_FEATURES]) {
> >> +		u64 missing = 0x0ULL;
> >> +
> >>   		config.device_features =
> >>   			nla_get_u64(nl_attrs[VDPA_ATTR_DEV_FEATURES]);
> >> +		if (nl_attrs[VDPA_ATTR_DEV_NET_CFG_MACADDR] &&
> >> +		    !(config.device_features & BIT_ULL(VIRTIO_NET_F_MAC)))
> >> +			missing |= BIT_ULL(VIRTIO_NET_F_MAC);
> >> +		if (nl_attrs[VDPA_ATTR_DEV_NET_CFG_MTU] &&
> >> +		    !(config.device_features & BIT_ULL(VIRTIO_NET_F_MTU)))
> >> +			missing |= BIT_ULL(VIRTIO_NET_F_MTU);
> >> +		if (nl_attrs[VDPA_ATTR_DEV_NET_CFG_MAX_VQP] &&
> >> +		    config.net.max_vq_pairs > 1 &&
> >> +		    !(config.device_features & BIT_ULL(VIRTIO_NET_F_MQ)))
> >> +			missing |= BIT_ULL(VIRTIO_NET_F_MQ);
> >> +		if (missing) {
> >> +			NL_SET_ERR_MSG_FMT_MOD(info->extack,
> >> +					       "Missing features 0x%llx for
> >> provided attributes",
> >> +					       missing);
> >> +			return -EINVAL;
> >> +		}
> >>   		config.mask |= BIT_ULL(VDPA_ATTR_DEV_FEATURES);
> >>   	}
> >>
> >> --
> >> 1.8.3.1
> > Vdpa this layer can likely derive the feature bits for the supplied config fields
> so that user doesn't need to keep track of both.
> > Only those feature bits which are unrelated to any config, is what user should
> be setting.
> It's not I can't do this, but Jason wanted to have clear semantics around
> migration compatibility for the driver, and for that users have to explicitly
> provide device_features that we may define new driver behavior (rather that
> inheritance which is implicit and not uniformly define across drivers) for
> compatibility using the new uAPI.
Make sense to explicitly tell, just requires more careful plumbing on the user space side.
Eventually it will get orchestrated by non user, so it should be fine to explicitly define it.