From: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Date: Thu, 24 Oct 2019 01:11:23 -0700
Subject: ice: introduce frame padding computation logic
Patch-mainline: v5.5-rc1
Git-commit: 59bb08080557589aaf577a99d329ccea38b55c95
References: jsc#SLE-7926
Take into account the underlying architecture specific settings and
based on that calculate the possible padding that can be supplied.
Typically, for x86 and standard MTU size we will end up with 192 bytes
of headroom. This is the same behavior as our other drivers have and we
can dedicate it for XDP purposes.
Furthermore, introduce the Rx ring flag for indicating whether build_skb
is used on particular. Based on that invoke the routines for padding
calculation.
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Acked-by: Thomas Bogendoerfer <tbogendoerfer@suse.de>
---
drivers/net/ethernet/intel/ice/ice_base.c | 6 ++
drivers/net/ethernet/intel/ice/ice_ethtool.c | 2
drivers/net/ethernet/intel/ice/ice_lib.c | 3 -
drivers/net/ethernet/intel/ice/ice_txrx.c | 42 ++++++++------
drivers/net/ethernet/intel/ice/ice_txrx.h | 81 +++++++++++++++++++++++++++
5 files changed, 114 insertions(+), 20 deletions(-)
--- a/drivers/net/ethernet/intel/ice/ice_base.c
+++ b/drivers/net/ethernet/intel/ice/ice_base.c
@@ -406,6 +406,12 @@ int ice_setup_rx_ctx(struct ice_ring *ri
if (vsi->type == ICE_VSI_VF)
return 0;
+ /* configure Rx buffer alignment */
+ if (!vsi->netdev || test_bit(ICE_FLAG_LEGACY_RX, vsi->back->flags))
+ ice_clear_ring_build_skb_ena(ring);
+ else
+ ice_set_ring_build_skb_ena(ring);
+
/* init queue specific tail register */
ring->tail = hw->hw_addr + QRX_TAIL(pf_q);
writel(0, ring->tail);
--- a/drivers/net/ethernet/intel/ice/ice_ethtool.c
+++ b/drivers/net/ethernet/intel/ice/ice_ethtool.c
@@ -624,7 +624,7 @@ static int ice_lbtest_receive_frames(str
continue;
rx_buf = &rx_ring->rx_buf[i];
- received_buf = page_address(rx_buf->page);
+ received_buf = page_address(rx_buf->page) + rx_buf->page_offset;
if (ice_lbtest_check_frame(received_buf))
valid_frames++;
--- a/drivers/net/ethernet/intel/ice/ice_lib.c
+++ b/drivers/net/ethernet/intel/ice/ice_lib.c
@@ -1194,7 +1194,8 @@ void ice_vsi_cfg_frame_size(struct ice_v
vsi->max_frame = ICE_AQ_SET_MAC_FRAME_SIZE_MAX;
vsi->rx_buf_len = ICE_RXBUF_2048;
#if (PAGE_SIZE < 8192)
- } else if (vsi->netdev->mtu <= ETH_DATA_LEN) {
+ } else if (!ICE_2K_TOO_SMALL_WITH_PADDING &&
+ (vsi->netdev->mtu <= ETH_DATA_LEN)) {
vsi->max_frame = ICE_RXBUF_1536 - NET_IP_ALIGN;
vsi->rx_buf_len = ICE_RXBUF_1536 - NET_IP_ALIGN;
#endif
--- a/drivers/net/ethernet/intel/ice/ice_txrx.c
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.c
@@ -415,7 +415,12 @@ err:
*/
static unsigned int ice_rx_offset(struct ice_ring *rx_ring)
{
- return ice_is_xdp_ena_vsi(rx_ring->vsi) ? XDP_PACKET_HEADROOM : 0;
+ if (ice_ring_uses_build_skb(rx_ring))
+ return ICE_SKB_PAD;
+ else if (ice_is_xdp_ena_vsi(rx_ring->vsi))
+ return XDP_PACKET_HEADROOM;
+
+ return 0;
}
/**
@@ -710,7 +715,7 @@ ice_add_rx_frag(struct ice_ring *rx_ring
struct sk_buff *skb, unsigned int size)
{
#if (PAGE_SIZE >= 8192)
- unsigned int truesize = SKB_DATA_ALIGN(size);
+ unsigned int truesize = SKB_DATA_ALIGN(size + ice_rx_offset(rx_ring));
#else
unsigned int truesize = ice_rx_pg_size(rx_ring) / 2;
#endif
@@ -1008,27 +1013,28 @@ static int ice_clean_rx_irq(struct ice_r
xdp_res = ice_run_xdp(rx_ring, &xdp, xdp_prog);
rcu_read_unlock();
- if (xdp_res) {
- if (xdp_res & (ICE_XDP_TX | ICE_XDP_REDIR)) {
- unsigned int truesize;
+ if (!xdp_res)
+ goto construct_skb;
+ if (xdp_res & (ICE_XDP_TX | ICE_XDP_REDIR)) {
+ unsigned int truesize;
#if (PAGE_SIZE < 8192)
- truesize = ice_rx_pg_size(rx_ring) / 2;
+ truesize = ice_rx_pg_size(rx_ring) / 2;
#else
- truesize = SKB_DATA_ALIGN(size);
+ truesize = SKB_DATA_ALIGN(ice_rx_offset(rx_ring) +
+ size);
#endif
- xdp_xmit |= xdp_res;
- ice_rx_buf_adjust_pg_offset(rx_buf, truesize);
- } else {
- rx_buf->pagecnt_bias++;
- }
- total_rx_bytes += size;
- total_rx_pkts++;
-
- cleaned_count++;
- ice_put_rx_buf(rx_ring, rx_buf);
- continue;
+ xdp_xmit |= xdp_res;
+ ice_rx_buf_adjust_pg_offset(rx_buf, truesize);
+ } else {
+ rx_buf->pagecnt_bias++;
}
+ total_rx_bytes += size;
+ total_rx_pkts++;
+
+ cleaned_count++;
+ ice_put_rx_buf(rx_ring, rx_buf);
+ continue;
construct_skb:
if (skb)
ice_add_rx_frag(rx_ring, rx_buf, skb, size);
--- a/drivers/net/ethernet/intel/ice/ice_txrx.h
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.h
@@ -26,6 +26,71 @@
#define ICE_RX_BUF_WRITE 16 /* Must be power of 2 */
#define ICE_MAX_TXQ_PER_TXQG 128
+/* Attempt to maximize the headroom available for incoming frames. We use a 2K
+ * buffer for MTUs <= 1500 and need 1536/1534 to store the data for the frame.
+ * This leaves us with 512 bytes of room. From that we need to deduct the
+ * space needed for the shared info and the padding needed to IP align the
+ * frame.
+ *
+ * Note: For cache line sizes 256 or larger this value is going to end
+ * up negative. In these cases we should fall back to the legacy
+ * receive path.
+ */
+#if (PAGE_SIZE < 8192)
+#define ICE_2K_TOO_SMALL_WITH_PADDING \
+((NET_SKB_PAD + ICE_RXBUF_1536) > SKB_WITH_OVERHEAD(ICE_RXBUF_2048))
+
+/**
+ * ice_compute_pad - compute the padding
+ * rx_buf_len: buffer length
+ *
+ * Figure out the size of half page based on given buffer length and
+ * then subtract the skb_shared_info followed by subtraction of the
+ * actual buffer length; this in turn results in the actual space that
+ * is left for padding usage
+ */
+static inline int ice_compute_pad(int rx_buf_len)
+{
+ int half_page_size;
+
+ half_page_size = ALIGN(rx_buf_len, PAGE_SIZE / 2);
+ return SKB_WITH_OVERHEAD(half_page_size) - rx_buf_len;
+}
+
+/**
+ * ice_skb_pad - determine the padding that we can supply
+ *
+ * Figure out the right Rx buffer size and based on that calculate the
+ * padding
+ */
+static inline int ice_skb_pad(void)
+{
+ int rx_buf_len;
+
+ /* If a 2K buffer cannot handle a standard Ethernet frame then
+ * optimize padding for a 3K buffer instead of a 1.5K buffer.
+ *
+ * For a 3K buffer we need to add enough padding to allow for
+ * tailroom due to NET_IP_ALIGN possibly shifting us out of
+ * cache-line alignment.
+ */
+ if (ICE_2K_TOO_SMALL_WITH_PADDING)
+ rx_buf_len = ICE_RXBUF_3072 + SKB_DATA_ALIGN(NET_IP_ALIGN);
+ else
+ rx_buf_len = ICE_RXBUF_1536;
+
+ /* if needed make room for NET_IP_ALIGN */
+ rx_buf_len -= NET_IP_ALIGN;
+
+ return ice_compute_pad(rx_buf_len);
+}
+
+#define ICE_SKB_PAD ice_skb_pad()
+#else
+#define ICE_2K_TOO_SMALL_WITH_PADDING false
+#define ICE_SKB_PAD (NET_SKB_PAD + NET_IP_ALIGN)
+#endif
+
/* We are assuming that the cache line is always 64 Bytes here for ice.
* In order to make sure that is a correct assumption there is a check in probe
* to print a warning if the read from GLPCI_CNF2 tells us that the cache line
@@ -231,6 +296,7 @@ struct ice_ring {
* in their own cache line if possible
*/
#define ICE_TX_FLAGS_RING_XDP BIT(0)
+#define ICE_RX_FLAGS_RING_BUILD_SKB BIT(1)
u8 flags;
dma_addr_t dma; /* physical address of ring */
unsigned int size; /* length of descriptor ring in bytes */
@@ -239,6 +305,21 @@ struct ice_ring {
u8 dcb_tc; /* Traffic class of ring */
} ____cacheline_internodealigned_in_smp;
+static inline bool ice_ring_uses_build_skb(struct ice_ring *ring)
+{
+ return !!(ring->flags & ICE_RX_FLAGS_RING_BUILD_SKB);
+}
+
+static inline void ice_set_ring_build_skb_ena(struct ice_ring *ring)
+{
+ ring->flags |= ICE_RX_FLAGS_RING_BUILD_SKB;
+}
+
+static inline void ice_clear_ring_build_skb_ena(struct ice_ring *ring)
+{
+ ring->flags &= ~ICE_RX_FLAGS_RING_BUILD_SKB;
+}
+
static inline bool ice_ring_is_xdp(struct ice_ring *ring)
{
return !!(ring->flags & ICE_TX_FLAGS_RING_XDP);