1This file is here for people to list their contributions to the 2netperf benchmark. When you enhance or fix something, put your name 3and a description of the what/where/whey/why/how here. If you like, 4feel free to include an email address. 5 6I would like to apologize in advance to anyone I've forgotten to 7include. 8 9Rick Jones <raj@cup.hp.com> 10Responsible for initial packaging and release of netperf and 11"editorial" continuity for subsequent releases. 12 13Karen Choy 14Code to allow netserver to run as a standalone daemon. 15 16Dave Shield <daves@csc.liv.ac.uk> 17Wrote the first revision of the netperf and netserver manpages. 18 19Sarr Blumson <sarr@citi.umich.edu> 20Fixes for AIX 3.1 and 3.2. Also fixes for Solaris 2.1 without 21realizing it ;-) 22 23Jeff Smits <jgs@usl.com> 24Fixes for TCP_RR and UDP_RR on sytems with an htonl that is not a no-op. 25 26Warren Burnett <somewhere in Kansas> 27Example code for DLPI tests. 28 29Several Folks 30Code to tell SunOS 4 to *not* restart system calls on receipt of a 31signal. 32 33Fore Systems Inc. 34Manpages for the FORE API and question answering 35 36David Channin 37Access to systems running the Fore ATM API 38 39Jonathan Stone <jonathan@DSG.Stanford.EDU> 40Include file fixes for Ultrix 41 42Bruce Barnett <barnett@grymoire.crd.ge.com> 43Bunches of warnings fixes and lint picks for Solaris 2.3 44 45Herman Dierks et al <dierks@austin.ibm.com> 46Code to calculate confidence intervals for tests 47 48Hal Murray <murray@pa.dec.com> 49Helpful suggestions for the scripts to make them more compatible with 50the netperf database (http://www.cup.hp.com/netperf/NetperfPage.html). 51His prompting finally got me off my whatever to put the confidence 52interval stuff from the guys at IBM into netperf 53 54Peter Skopp pds@cs.columbia.edu 55Fixes to make netperf more secure. 56 57Tom Wilson <tcwilson@newbridge.com> 58A fix to send_udp_rr to correct bogus throughput values. 59 60Thorsten Lockert <tholo@tetherless.com> 61A bunch of clean-up for the *BSD OSes 62 63Serge Pachkovsky <ps@ocisgi7.unizh.ch> 64Code for low-priority soaker process for AIX and SGI 65 66The fine folks at Adaptec 67The initial port of netperf (1.9PL4) to Windows NT 3.51. 68 69Robin Callender <robin_callender@VNET.IBM.COM> 70The PPC binaries for the 1.9PL4 port of netperf to NT and for assorted 71code clean-ups and help with CPU utilization measurements. 72 73"Todd J. Derr" <tjd@fore.com> 74For offering to take-on support of the Fore ATM API files with 2.1 and 75for his help in making the tar files and such more user-friendly 76 77Michael Shuldman <michaels@inet.no> 78Improvements to the makefile and additional checks for OpenBSD 79Changes to make netperf more profiling friendly 80 81Kris Corwin 82discovery of a debug statement outside of if (debug) that may have 83been the cause of all the nasty connection refused errors in random 84UDP_RR tests... 85 86Charles Harris 87Initial prototype of the TCP_SENDFILE test support 88 89Philip Pishioneri of Cornel 90Conversion of the netperf.ps manul to PDF format. 91 92The Hewlett-Packard OpenVMS folks 93Assistance with the port to OpenVMS 94 95Munechika SUMIKAWA @ KAME Project / FreeBSD.org 96IPv6 fixes 97 98Jan Iven of CERN 99initial mods for sendfile() under Linux 100 101Fabrice Bacchella 102for pointing-out that Solaris 9 has a copy of Linux sendfile() 103 104Andrew Gallatin 105for assistance with the FreeBSD sysctl() stuff, and later making it 106calibration-free 107fixes to configure to recognize Solaris 11 108fixes to netcpu_procstat.c for later linux kernels 109workarounds to get Linux to report ENOBUFS on TX queue overflows 110 111Mark Cooper 112pointing-out the need for -lresolv when compiling -DDO_DNS on RedHat 1137.1 114 115Carl Mascott 116finding some cut-and-paste erors in create_data_socket error logging 117 118Fabrice Bacchella 119Fixes for -DHISTOGRAM and -DUNIX on Mac OS X, updates to usage strings 120 121Spencer Frink 122Fixes and Cleanup for WIN32. Many over many years. 123 124Nicholas Thomas 125Fixes for DLPI on SVR4 Streams under Linux 126 127Dave Craig 128Fixes for getaddrinfo error returns 129 130David Mosberger of HP 131Workaround for the Linux getsockopt() bug that returns more than that 132for which one asked. 133 134Stephen Burger of HP 135Code to implement the netserver CPU binding. 136 137Vladislav "Vlad" Yasevich of HP 138Initial SCTP tests. Enhancements to the configure.ac sources to show 139the way to make many of the LIBS="foo" before ./configure unnecessary. 140 141Padmanabhan "Paddu" S N of HP 142Patches for /proc/stat CPU util and recv_tcp_rr. 143 144Cary Coutant and other hp-mac-users of HP 145Access to Mac OS X systems for porting netperf 2.4.0 146 147Chris Bertin of HP 148Access to AIX for initial porting of netperf 2.4.0 149 150James Carlson 151Assistance finding the right magic to compile SCTP on Solaris 10. 152 153Gavin 154Pointers on Solaris 10 Microstate accounting. 155 156Brent Draney 157Getting netcpu_perfstat.c in running order on AIX and other misc 158fixups in places such as BSD. 159 160Samuel Ying 161Change struct sockaddr to struct sockaddr storage in netserver.c 162 163Rodolpho Boer 164Fix for default message size in UDP_STREAM when defualt SO_SNDBUF size 165is > max UDP datagram size. 166 167Michael Dorff 168Getting netperf/netserver to compile under Windows with MS Visual 169Studio 2003 170 171George Davis 172Changes to deal with different floating-point formats. 173 174Anonymous 175Changes to retrieve CPU util on MacOS X. 176 177Dickon Reed 178Patches to attend to some windows in TCP_CRR and TCP_CC under Windows 179 180Bret McKee <bret@hp.com> 181Fixes to get netcpu_looper compiling and working after the "netcpu" 182split 183 184Hans Blom 185Improvements to closing/redirecting stdin/stdout/stderr in netserver 186 187Martin Brown 188RPM support in the form of netperf.spec.in and related configure.ac 189etc changes 190 191Shilpi Agarwal 192Changes to allow UDP_STREAM to use connected sockets on both sides. 193 194Steve Reinhardt <stever@eecs.umich.edu> 195Fixes for buffer filling. 196 197Gisle Vanem 198Fixes for Windows compilation under MingW/gcc. 199 200Scott Weitzenkamp 201Patches to enable demo mode in the UDP_STREAM test 202 203Emir Halepovic 204Feedback on the manual 205 206Kouhei Sutou 207Generate netperf_version.h and netperf.spec via configure rather than 208makefile, include limits to get PathMAX on FreeBSD. 209 210Dan Yost 211Fix to fflush() each interim result in demo mode to make things 212happier for folks redirecting same to a file. 213 214Alexander Duyck 215Fixes to replace struct sockaddr_in with struct sockaddr_storage 216Fixes to UDP_RR to preclude hangs on Windows 217Fizes to UDP_RR to honour -f and -B options 218 219Anonymous 220Support for sendfile() on OSX 221 222Matt Waddel 223Fix to use vfork() instead of fork() on platforms without fork() 224 225Simon Burge <simonb@NetBSD.org> 226Fixes for *BSD CPU util. 227 228Adam Bidema 229Fixes for launching netserver children when the path to netserver.exe 230is very long. 231 232Gisle Vanem 233MingW cnd MSDOS (djgpp) cleanups. 234 235Bruno Cornec 236Fixes to supply some missing fprintf format statements. 237 238Jose Pedro Oliveira 239Linux compilation fixes when SCTP and DCCP are enabled. Changes to 240configure.ac to allow it to inherit compiler flags from environment 241variables. 242 243Pal Baranyai 244Typo in nettest_dlpi.c 245 246Brian Haley 247Changes to allow something like netperf -H ::1 to work 248 249Frank Schuster 250Discovering that FreeBSD 8.X also does not need an SCTP library 251 252Josselin Costanzi 253DEBUG_LOG_FILE patch for Android 254 255Dongsheng Song 256Assorted Windows fixes 257 258Hushan Jia 259Fix to actually run SCTP_RR_MANY when requested 260 261David MacMahon davidm@astro.berkeley.edu 262Fix to allow configure in a build directory. 263Remove duplicate htonl in recv_udp_stream(). 264 265Masanari Iida 266Cleanups for the manual. 267 268Amir Vidai 269Patch from which changes to set SO_PRIORITY were based. 270 271Alexander Duyck 272Uncovering an out of bounds access of the netperf_output_source array. 273 274Bjoern A. Zeeb - a patch to enable netperf to correctly manipulate 275IPv6 addresses in the control message. 276 277Shachar Raindel - a patch to bring the "get linux to emit ENOBUFS" 278functionality from classic netperf tests to the omni tests. 279 280Chema Gonzalez - fixes for some assorted warnings. 281 282Dave Taht - changes for symbolic handling of TOS values 283 284David Ahern - fix for an incorrect pointer check 285 286Jim Nell - Fix for Solaris 11.1 compilation problem in dscp.c 287 288Steven Noonan - found a variable name bug for the ever unlikely case 289of a platform not defining SO_SNDBUF 290 291Amos Kong - a fix to report the correct send size in a migrated 292TCP_MAERTS test. Bug reported by Wenli Quan 293 294Eric Dumazet - improve the performance of launching many netperfs at 295once by obtaining the 16 bytes of randomness for the UUID via slightly 296different means. Dump TCP_INFO results to stderr when either debug > 2971 or DUMP_TCP_INFO envvar is set. Bump the listen backlog for the 298control endpoint to 1024. 299 300David Decotigny - a series of cleanups and enhancements in the area of 301CPU utilization reporting and making certain that sanity checks are 302made on the size of things being passed in control messages. 303 304Yaogong Wang - enable support for the -N test-specific option to have 305the UDP socket connected at the netserver side. 306 307Weijia Song - a fix inspiration for a NULL pointer problem. 308 309Gisle Vanem - some Windows compilation fixes