xsk: Eliminate the RX batch size (4b638f13) · Commits · 戴 / test

In the xsk consumer ring code there is a variable called RX_BATCH_SIZE that dictates the minimum number of entries that we try to grab from the fill and Tx rings. In fact, the code always try to grab the maximum amount of entries from these rings. The only thing this variable does is to throw an error if there is less than 16 (as it is defined) entries on the ring. There is no reason to do this and it will just lead to weird behavior from user space's point of view. So eliminate this variable. With this change, we will be able to simplify the xskq_nb_free and xskq_nb_avail code in the next commit. Signed-off-by:

Magnus Karlsson <magnus.karlsson@intel.com> Signed-off-by:

Alexei Starovoitov <ast@kernel.org> Link: https://lore.kernel.org/bpf/1576759171-28550-6-git-send-email-magnus.karlsson@intel.com

net/xdp/xsk_queue.h

+2 −4

Original line number	Diff line number	Diff line
		@@ -10,8 +10,6 @@
		#include <linux/if_xdp.h>
		#include <net/xdp_sock.h>

		#define RX_BATCH_SIZE 16

		struct xdp_ring {
		u32 producer ____cacheline_aligned_in_smp;
		u32 consumer ____cacheline_aligned_in_smp;
		@@ -202,7 +200,7 @@ static inline u64 xskq_peek_addr(struct xsk_queue q, u64 *addr,
		if (q->cons_tail == q->cons_head) {
		smp_mb(); /* D, matches A */
		WRITE_ONCE(q->ring->consumer, q->cons_tail);
		q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
		q->cons_head = q->cons_tail + xskq_nb_avail(q, 1);

		/* Order consumer and data */
		smp_rmb();
		@@ -320,7 +318,7 @@ static inline struct xdp_desc xskq_peek_desc(struct xsk_queue q,
		if (q->cons_tail == q->cons_head) {
		smp_mb(); /* D, matches A */
		WRITE_ONCE(q->ring->consumer, q->cons_tail);
		q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
		q->cons_head = q->cons_tail + xskq_nb_avail(q, 1);

		/* Order consumer and data */
		smp_rmb(); /* C, matches B */

Admin message