Lines Matching +full:cost +full:- +full:effective

14 -----------------------
22 it replaces per byte copy cost with page accounting and completion
24 effective at writes over around 10 KB.
39 ---------
42 netdev 2.1. For more in-depth information see that paper and talk,
52     [PATCH net-next v4 0/9] socket sendmsg MSG_ZEROCOPY
53     https://lore.kernel.org/netdev/20170803202945.70750-1-willemdebruijn.kernel@gmail.com
63 ------------
76 ------------
85 A zerocopy failure will return -1 with errno ENOBUFS. This happens if
100 -------------
107 maintains an internal unsigned 32-bit counter. Each send call with
120 Reading from the error queue is always a non-blocking operation. The
129 	if (poll(&pfd, 1, -1) != 1 || pfd.revents & POLLERR == 0)
133 	if (ret == -1)
161 For protocols that acknowledge data in-order, like TCP, each
183 The 32-bit notification range is encoded as [ee_info, ee_data]. This
193 	if (cm->cmsg_level != SOL_IP &&
194 	    cm->cmsg_type != IP_RECVERR)
198 	if (serr->ee_errno != 0 ||
199 	    serr->ee_origin != SO_EE_ORIGIN_ZEROCOPY)
202 	printf("completed: %u..%u\n", serr->ee_info, serr->ee_data);
213 scatter-gather I/O cannot send packets made up of kernel generated
225 also incurs notification processing cost for no benefit. For this
236 --------