/* Test for nonblocking read and write. Copyright (C) 2011-2023 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see . */ /* A data block ought to be larger than the size of the in-kernel buffer. Working values of SOCKET_DATA_BLOCK_SIZE, depending on kernel: Platform SOCKET_DATA_BLOCK_SIZE Linux >= 7350000 (depends on circumstances) Linux/Android >= 1700000 (approx.) FreeBSD >= 107521 OpenBSD >= 106430 (depends on circumstances) Mac OS X >= 680000 (depends on circumstances) AIX 5.1 >= 125713 AIX 7.1 >= 200000 (depends on circumstances) HP-UX >= 114689 IRIX >= 61089 OSF/1 >= 122881 Solaris 7 >= 63000 (depends on circumstances) Solaris 8 >= 49153 Solaris 9 >= 73729 Solaris 10 >= 98305 Solaris 11 2010-11 >= 73729 Cygwin 1.5.x >= 66294401 but then write() fails with ENOBUFS Cygwin 1.7.x >= 163838 (depends on circumstances) native Windows >= 66294401 */ #define SOCKET_DATA_BLOCK_SIZE 1000000 /* On Linux, Mac OS X, Cygwin 1.5.x, native Windows, sockets have very large buffers in the kernel, so that write() calls succeed before the reader has started reading, even if fd is blocking and the amount of data is larger than 1 MB. */ #if defined __linux__ || defined __ANDROID__ || (defined __APPLE__ && defined __MACH__) || defined _WIN32 || defined __CYGWIN__ # define SOCKET_HAS_LARGE_BUFFER 1 #else # define SOCKET_HAS_LARGE_BUFFER 0 #endif