Re: [PATCH net] ixgbe: napi_poll must return the work done

On Wed, 2016-06-15 at 09:34 -0700, Venkatesh Srinivas wrote:

On Wed, Jun 15, 2016 at 6:37 AM, Paolo Abeni [off-list ref]
wrote:

quoted

Currently the function ixgbe_poll() returns 0 when it clean
completely
the rx rings, but this foul budget accounting in core code.
Fix this returning the actual work done, capped to weight - 1,
since
the core doesn't allow to return the full budget when the driver
modifies
the napi status

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c

b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 088c47c..8bebd86 100644

--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c

@@ -2887,7 +2887,7 @@ int ixgbe_poll(struct napi_struct *napi, int

budget)
        if (!test_bit(__IXGBE_DOWN, &adapter->state))
                ixgbe_irq_enable_queues(adapter, BIT_ULL(q_vector-

quoted

v_idx));

-       return 0;
+       return min(work_done, budget - 1);
 }

 /**

Reviewed-by: Venkatesh Srinivas <redacted>

The same bit of code appears in fm10k and i40e/i40evf. ixgb appears
to
correctly return work_done.

ixgbe_poll also appears to return an (minor) incorrect work_done in
another case, BTW. It divides its
budget between Rx rings associated with a vector. If any ring exceeds
its share of the budget, ixgbe_poll
claims to have consumed the full budget, even if a full budget of
frames was not received in a single
pass.

-- vs;

For the record, I couldn't find any documentation on this in
Documentatino/networking, or as a function header. Where would be the
best place to document the expectations of the napi core? I'd like to
submit a patch so that future it will be easier to determine what a new
driver should do (without just blindly copying from other drivers as
has caused these so far).

Thanks,
Jake