summaryrefslogtreecommitdiff
path: root/drivers/message
diff options
context:
space:
mode:
authorDavid S. Miller <davem@davemloft.net>2019-12-12 17:14:01 -0800
committerDavid S. Miller <davem@davemloft.net>2019-12-12 17:14:01 -0800
commit355979decf07d77e9f1fa59233b1e60ae51e307a (patch)
treeb79569d682e7e8c7c549193a2013ccec8adc16cf /drivers/message
parentaabf23d18322100720cb40002cdf05397f619914 (diff)
parent913b0bfd6c859aaf1735e1c8c69731e1d5359d31 (diff)
Merge branch 'Introduce-XDP-to-ena'
Sameeh Jubran says: ==================== Introduce XDP to ena This patchset includes 3 patches: * XDP_DROP implementation * XDP_TX implementation * A fix for an issue which might occur due to the XDP_TX patch. I see fit to place it as a standalone patch for clarity. Difference from v2: * Fixed the usage of rx headroom (XDP_PACKET_HEADROOM) * Aligned the page_offset of the packet when passing it to the stack * Switched to using xdp_frame in xdp xmit queue * Dropped the print for unsupported commands * Cosmetic changes Difference from RFC v1 (XDP_DROP patch): * Initialized xdp.rxq pointer * Updated max_mtu on attachment of xdp and removed the check from ena_change_mtu() * Moved the xdp execution from ena_rx_skb() to ena_clean_rx_irq() * Moved xdp buff (struct xdp_buff) from rx_ring to the local stack * Started using netlink's extack mechanism to deliver error messages to the user ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'drivers/message')
0 files changed, 0 insertions, 0 deletions