Lines Matching +full:cost +full:- +full:effective
15 -----------------------
23 it replaces per byte copy cost with page accounting and completion
25 effective at writes over around 10 KB.
40 ---------
43 netdev 2.1. For more in-depth information see that paper and talk,
53 [PATCH net-next v4 0/9] socket sendmsg MSG_ZEROCOPY
54 https://lore.kernel.org/netdev/20170803202945.70750-1-[email protected]
64 ------------
77 ------------
86 A zerocopy failure will return -1 with errno ENOBUFS. This happens if
101 -------------
108 maintains an internal unsigned 32-bit counter. Each send call with
121 Reading from the error queue is always a non-blocking operation. The
130 if (poll(&pfd, 1, -1) != 1 || pfd.revents & POLLERR == 0)
134 if (ret == -1)
162 For protocols that acknowledge data in-order, like TCP, each
186 The 32-bit notification range is encoded as [ee_info, ee_data]. This
196 if (cm->cmsg_level != SOL_IP &&
197 cm->cmsg_type != IP_RECVERR)
201 if (serr->ee_errno != 0 ||
202 serr->ee_origin != SO_EE_ORIGIN_ZEROCOPY)
205 printf("completed: %u..%u\n", serr->ee_info, serr->ee_data);
216 scatter-gather I/O cannot send packets made up of kernel generated
228 also incurs notification processing cost for no benefit. For this
239 --------
249 Data path sent to local sockets is the same as for non-local sockets.