summary refs log tree commit diff
path: root/net
diff options
context:
space:
mode:
authorMagnus Karlsson <magnus.karlsson@intel.com>2020-10-08 16:12:18 +0200
committerDaniel Borkmann <daniel@iogearbox.net>2020-10-09 16:35:01 +0200
commitc3f01fdcedf9459b2a27b22287414ab3a0d5d818 (patch)
tree3e656fa260264bbc9e0f7518a030ba39d05c6e89 /net
parent1e9259eca8fdf04b8d12ec093711a73cde19aed5 (diff)
downloadlinux-c3f01fdcedf9459b2a27b22287414ab3a0d5d818.tar.gz
xsk: Introduce padding between ring pointers
Introduce one cache line worth of padding between the producer and
consumer pointers in all the lockless rings. This so that the HW
adjacency prefetcher will not prefetch the consumer pointer when the
producer pointer is used and vice versa. This improves throughput
performance for the l2fwd sample app with 2% on my machine with HW
prefetching turned on.

Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/1602166338-21378-1-git-send-email-magnus.karlsson@gmail.com
Diffstat (limited to 'net')
-rw-r--r--net/xdp/xsk_queue.h4
1 files changed, 4 insertions, 0 deletions
diff --git a/net/xdp/xsk_queue.h b/net/xdp/xsk_queue.h
index dc1dd5ef70d1..cdb9cf3cd136 100644
--- a/net/xdp/xsk_queue.h
+++ b/net/xdp/xsk_queue.h
@@ -15,6 +15,10 @@
 
 struct xdp_ring {
 	u32 producer ____cacheline_aligned_in_smp;
+	/* Hinder the adjacent cache prefetcher to prefetch the consumer
+	 * pointer if the producer pointer is touched and vice versa.
+	 */
+	u32 pad ____cacheline_aligned_in_smp;
 	u32 consumer ____cacheline_aligned_in_smp;
 	u32 flags;
 };