> On Thu, 2015-12-17 at 06:51 +0000, [email protected] wrote:
> > From: Kan Liang <[email protected]>
> >
> > Device driver has to provide an interface to get per queue coalesce.
> > The interrupt coalescing parameters of each masked queue will be
> > copied back to user space one by one.
> >
> > Signed-off-by: Kan Liang <[email protected]>
> > ---
> > include/linux/ethtool.h | 5 ++++-
> > net/core/ethtool.c | 33 ++++++++++++++++++++++++++++++++-
> > 2 files changed, 36 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h index
> > 653dc9c..107f75f 100644
> > --- a/include/linux/ethtool.h
> > +++ b/include/linux/ethtool.h
> > @@ -201,6 +201,8 @@ static inline u32 ethtool_rxfh_indir_default(u32
> > index, u32 n_rx_rings)
> > * @get_module_eeprom: Get the eeprom information from the plug-in
> > module
> > * @get_eee: Get Energy-Efficient (EEE) supported and status.
> > * @set_eee: Set EEE status (enable/disable) as well as LPI timers.
> > + * @get_per_queue_coalesce: Get interrupt coalescing parameters per
> queue.
> > + * Returns a negative error code or zero.
>
> Is the ethtool core or the driver responsible for range-checking the queue
> number?
>
> What is reported if the queue number is in range for RX queues but not for
> TX queues (or vice versa)?
>
The driver should do the range-checking.
I think it's also driver to determine which should be reported
in tx_coalesce_usecs, if RX in range but TX not. Maybe ~0 is a good return
for tx_coalesce_usecs.
> > *
> > * All operations are optional (i.e. the function pointer may be set
> > * to %NULL) and callers must take this into account. Callers must
> > @@ -279,7 +281,8 @@ struct ethtool_ops {
> > const struct ethtool_tunable *, void *);
> > int (*set_tunable)(struct net_device *,
> > const struct ethtool_tunable *, const void *);
> > -
> > + int (*get_per_queue_coalesce)(struct net_device *, int,
> > + struct ethtool_coalesce *);
> >
> > };
> > #endif /* _LINUX_ETHTOOL_H */
> > diff --git a/net/core/ethtool.c b/net/core/ethtool.c index
> > 125fb32..22ff69a 100644
> > --- a/net/core/ethtool.c
> > +++ b/net/core/ethtool.c
> > @@ -1748,6 +1748,36 @@ out:
> > return ret;
> > }
> >
> > +static int ethtool_get_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->get_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) {
> [...]
>
> This assumes a little-endian system.
>
> You should probably build on top of David Decotigny's recent patch series
> that adds conversion of u32-based bitmaps to kernel bitmaps.
>
Thanks for pointing me to the patch.
I will change the code based on his patch.
Thanks,
Kan
