diff options
| author |   Magnus Karlsson <magnus.karlsson@intel.com> | 2020-06-23 11:44:17 +0200 |
|---|---|---|
| committer |   Tony Nguyen <anthony.l.nguyen@intel.com> | 2020-07-01 14:27:11 -0700 |
| commit | 4b5539c01ddf838ea6f5b67d00ee035231c8248a (patch) | |
| tree | 58d71b8f7f4bef76d2fa7d921f051b3ecdb98fae /drivers/net/ethernet/intel/i40e/i40e_txrx.c | |
| parent | i40e: optimize AF_XDP Tx completion path (diff) | |
| download | linux-dev-4b5539c01ddf838ea6f5b67d00ee035231c8248a.tar.xz linux-dev-4b5539c01ddf838ea6f5b67d00ee035231c8248a.zip | |
i40e: eliminate division in napi_poll data path
Eliminate a division in the napi_poll data path. This division is
executed even though it is only needed in the rare case when there are
not enough interrupt lines so they have to be shared between queue
pairs. Instead, just test for this case and only execute the division
if needed. The code has been lifted from the ice driver.
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Diffstat (limited to 'drivers/net/ethernet/intel/i40e/i40e_txrx.c')
| -rw-r--r-- | drivers/net/ethernet/intel/i40e/i40e_txrx.c | 14 |
1 files changed, 10 insertions, 4 deletions
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c index aad4326ab0c9..3e5c566ceb01 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c +++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c @@ -2595,10 +2595,16 @@ int i40e_napi_poll(struct napi_struct *napi, int budget) if (budget <= 0) goto tx_only; - /* We attempt to distribute budget to each Rx queue fairly, but don't - * allow the budget to go below 1 because that would exit polling early. - */ - budget_per_ring = max(budget/q_vector->num_ringpairs, 1); + /* normally we have 1 Rx ring per q_vector */ + if (unlikely(q_vector->num_ringpairs > 1)) + /* We attempt to distribute budget to each Rx queue fairly, but + * don't allow the budget to go below 1 because that would exit + * polling early. + */ + budget_per_ring = max_t(int, budget / q_vector->num_ringpairs, 1); + else + /* Max of 1 Rx ring in this q_vector so give it the budget */ + budget_per_ring = budget; i40e_for_each_ring(ring, q_vector->rx) { int cleaned = ring->xsk_umem ? |