[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [UNIKRAFT/LWIP PATCH v4 3/5] uknetdev: Per device RX/TX headroom



Hello Simon,

Thanks for the work.

Please find the comment inline.

Thanks & Regards

Sharan

On 10/30/20 7:31 PM, Simon Kuenzer wrote:
Store required rx and tx headroom on `lwip_data`.

Signed-off-by: Simon Kuenzer <simon.kuenzer@xxxxxxxxx>
---
  uknetdev.c | 39 ++++++++++++++-------------------------
  1 file changed, 14 insertions(+), 25 deletions(-)

diff --git a/uknetdev.c b/uknetdev.c
index ccb7368..ba7d41c 100644
--- a/uknetdev.c
+++ b/uknetdev.c
@@ -65,6 +65,8 @@
  struct lwip_netdev_data {
        uint32_t features;
        struct uk_alloc *pkt_a;
+       uint16_t rx_headroom;
+       uint16_t tx_headroom;
Can't we directly use the `uk_netdev_info` struct directly instead of copying the fields.
  #ifdef CONFIG_HAVE_SCHED
        struct uk_thread *poll_thread; /* Thread per device */
        char *_name; /* Thread name */
@@ -81,18 +83,6 @@ struct lwip_netdev_data {
   */
  UK_CTASSERT((sizeof(struct lwip_netdev_data)) <= 
CONFIG_UK_NETDEV_SCRATCH_SIZE);
-/*
- * Global headroom settings for buffer allocations used on receive
- * and transmit. We are taking the maximum of all uknetdev devices as
- * simplification.
- * TODO: A per-device setting might be more efficient but requires more data
- *       fields for `netif->state`. For now we point directly to the according
- *       `struct uk_netdev` in order to avoid another allocation for these
- *       per-device fields.
- */
-static uint16_t rx_headroom = ETH_PAD_SIZE;
-static uint16_t tx_headroom = ETH_PAD_SIZE;
-
  #define netif_to_uknetdev(nf) \
        ((struct uk_netdev *) (nf)->state)
@@ -109,7 +99,7 @@ static uint16_t netif_alloc_rxpkts(void *argp, struct uk_netbuf *nb[],
        for (i = 0; i < count; ++i) {
                nb[i] = lwip_alloc_netbuf(lwip_data->pkt_a,
                                          UKNETDEV_BUFLEN,
-                                         rx_headroom);
+                                         lwip_data->rx_headroom);
                if (!nb[i]) {
                        /* we run out of memory */
                        break;
@@ -139,7 +129,7 @@ static err_t uknetdev_output(struct netif *nf, struct pbuf 
*p)
        if (!allocation)
                return ERR_MEM;
        nb = uk_netbuf_prepare_buf(allocation, UKNETDEV_BUFLEN,
-                                  tx_headroom, 0, NULL);
+                                  lwip_data->tx_headroom, 0, NULL);
        UK_ASSERT(nb);
        nb->_a = a; /* register allocator for free operation */
        nb->_b = allocation;
@@ -495,21 +485,20 @@ err_t uknetdev_init(struct netif *nf)
        uk_netdev_info_get(dev, &info);
        if (!info.max_rx_queues || !info.max_tx_queues)
                return ERR_IF;
-       lwip_data->features = info.features;
-       lwip_data->pkt_a    = a;
+       lwip_data->features    = info.features;
+#if ETH_PAD_SIZE
+       lwip_data->rx_headroom = info.nb_encap_rx + ETH_PAD_SIZE;
+       lwip_data->tx_headroom = info.nb_encap_tx + ETH_PAD_SIZE;
+#else
+       lwip_data->rx_headroom = info.nb_encap_rx;
+       lwip_data->tx_headroom = info.nb_encap_tx;
+#endif
+       lwip_data->pkt_a       = a;
- /*
-        * Update our global (rx|tx)_headroom setting that we use for
-        * buffer allocations
-        */
-       rx_headroom = (rx_headroom < info.nb_encap_rx)
-                     ? info.nb_encap_rx : rx_headroom;
-       tx_headroom = (tx_headroom < info.nb_encap_tx)
-                     ? info.nb_encap_tx : tx_headroom;
        LWIP_DEBUGF(NETIF_DEBUG,
                    ("%s: %c%c%u: Need headroom rx:%"PRIu16", tx:%"PRIu16"\n",
                     __func__, nf->name[0], nf->name[1], nf->num,
-                    info.nb_encap_rx, info.nb_encap_tx));
+                    lwip_data->rx_headroom, lwip_data->tx_headroom));
/*
         * Device configuration,



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.