> On Thu, 2015-12-17 at 06:51 +0000, kan.li...@intel.com wrote:
> [...]
> > --- a/net/core/ethtool.c
> > +++ b/net/core/ethtool.c
> > @@ -1778,6 +1778,37 @@ static int
> ethtool_get_per_queue_coalesce(struct net_device *dev,
> > return 0;
> > }
> >
> > +static int ethtool_set_per_queue_coalesce(struct net_device *dev,
> > + void __user *useraddr,
> > + struct ethtool_per_queue_op
> *per_queue_opt) {
> > + u64 queue_mask;
> > + int bit, i, ret;
> > +
> > + if (!dev->ethtool_ops->set_per_queue_coalesce)
> > + return -EOPNOTSUPP;
> > +
> > + useraddr += sizeof(*per_queue_opt);
> > + for (i = 0; i < MAX_QUEUE_MASK; i++) {
> > + queue_mask = per_queue_opt->queue_mask[i];
> > + if (queue_mask > 0) {
> > + for_each_set_bit(bit, (unsigned long
> *)&queue_mask, 64) {
> > + struct ethtool_coalesce coalesce;
> > +
> > + if (copy_from_user(&coalesce, useraddr,
> sizeof(coalesce)))
> > + return -EFAULT;
> > +
> > + ret = dev->ethtool_ops-
> >set_per_queue_coalesce(dev, bit + i * 64, &coalesce);
> > + if (ret != 0)
> > + return ret;
> [...]
>
> So there's no attempt to roll back on failure?
>
No, ethtool core doesn’t save the old value for this version. So it's hard to
Rollback.
If we want this feature, the kernel will alloc a chunk of memory to store
coalesce value for masked queues. Is it OK?
> Then, what is the benefit of doing this iteration in the kernel rather than
> userland?
>
It should save many ioctls.
Thanks,
Kan
