Commit 114bdef0 authored by Jakub Kicinski's avatar Jakub Kicinski Committed by David S. Miller

nfp: preallocate RX buffers early in .ndo_open

We want the .ndo_open() to have following structure:
 - allocate resources;
 - configure HW/FW;
 - enable the device from stack perspective.
Therefore filling RX rings needs to be moved to the beginning
of .ndo_open().
Signed-off-by: default avatarJakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 1934680f
...@@ -1666,28 +1666,19 @@ static void nfp_net_clear_config_and_disable(struct nfp_net *nn) ...@@ -1666,28 +1666,19 @@ static void nfp_net_clear_config_and_disable(struct nfp_net *nn)
* @nn: NFP Net device structure * @nn: NFP Net device structure
* @r_vec: Ring vector to be started * @r_vec: Ring vector to be started
*/ */
static int nfp_net_start_vec(struct nfp_net *nn, struct nfp_net_r_vector *r_vec) static void
nfp_net_start_vec(struct nfp_net *nn, struct nfp_net_r_vector *r_vec)
{ {
unsigned int irq_vec; unsigned int irq_vec;
int err = 0;
irq_vec = nn->irq_entries[r_vec->irq_idx].vector; irq_vec = nn->irq_entries[r_vec->irq_idx].vector;
disable_irq(irq_vec); disable_irq(irq_vec);
err = nfp_net_rx_ring_bufs_alloc(r_vec->nfp_net, r_vec->rx_ring);
if (err) {
nn_err(nn, "RV%02d: couldn't allocate enough buffers\n",
r_vec->irq_idx);
goto out;
}
nfp_net_rx_ring_fill_freelist(r_vec->rx_ring); nfp_net_rx_ring_fill_freelist(r_vec->rx_ring);
napi_enable(&r_vec->napi); napi_enable(&r_vec->napi);
out:
enable_irq(irq_vec);
return err; enable_irq(irq_vec);
} }
static int nfp_net_netdev_open(struct net_device *netdev) static int nfp_net_netdev_open(struct net_device *netdev)
...@@ -1742,6 +1733,10 @@ static int nfp_net_netdev_open(struct net_device *netdev) ...@@ -1742,6 +1733,10 @@ static int nfp_net_netdev_open(struct net_device *netdev)
err = nfp_net_rx_ring_alloc(nn->r_vecs[r].rx_ring); err = nfp_net_rx_ring_alloc(nn->r_vecs[r].rx_ring);
if (err) if (err)
goto err_free_tx_ring_p; goto err_free_tx_ring_p;
err = nfp_net_rx_ring_bufs_alloc(nn, nn->r_vecs[r].rx_ring);
if (err)
goto err_flush_rx_ring_p;
} }
err = netif_set_real_num_tx_queues(netdev, nn->num_tx_rings); err = netif_set_real_num_tx_queues(netdev, nn->num_tx_rings);
...@@ -1814,11 +1809,8 @@ static int nfp_net_netdev_open(struct net_device *netdev) ...@@ -1814,11 +1809,8 @@ static int nfp_net_netdev_open(struct net_device *netdev)
* - enable all TX queues * - enable all TX queues
* - set link state * - set link state
*/ */
for (r = 0; r < nn->num_r_vecs; r++) { for (r = 0; r < nn->num_r_vecs; r++)
err = nfp_net_start_vec(nn, &nn->r_vecs[r]); nfp_net_start_vec(nn, &nn->r_vecs[r]);
if (err)
goto err_disable_napi;
}
netif_tx_wake_all_queues(netdev); netif_tx_wake_all_queues(netdev);
...@@ -1827,18 +1819,14 @@ static int nfp_net_netdev_open(struct net_device *netdev) ...@@ -1827,18 +1819,14 @@ static int nfp_net_netdev_open(struct net_device *netdev)
return 0; return 0;
err_disable_napi:
while (r--) {
napi_disable(&nn->r_vecs[r].napi);
nfp_net_rx_ring_reset(nn->r_vecs[r].rx_ring);
nfp_net_rx_ring_bufs_free(nn, nn->r_vecs[r].rx_ring);
}
err_clear_config: err_clear_config:
nfp_net_clear_config_and_disable(nn); nfp_net_clear_config_and_disable(nn);
err_free_rings: err_free_rings:
r = nn->num_r_vecs; r = nn->num_r_vecs;
err_free_prev_vecs: err_free_prev_vecs:
while (r--) { while (r--) {
nfp_net_rx_ring_bufs_free(nn, nn->r_vecs[r].rx_ring);
err_flush_rx_ring_p:
nfp_net_rx_ring_free(nn->r_vecs[r].rx_ring); nfp_net_rx_ring_free(nn->r_vecs[r].rx_ring);
err_free_tx_ring_p: err_free_tx_ring_p:
nfp_net_tx_ring_free(nn->r_vecs[r].tx_ring); nfp_net_tx_ring_free(nn->r_vecs[r].tx_ring);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment