1
1
iperf3/ChangeLog
Jon Dugan bd8c92d8f6 merged Gerrit's nanosleep patch
tweaked type of currLen as fixup for Nathan's patch
2008-05-10 03:18:35 +00:00

74 строки
2.4 KiB
Plaintext

2008-05-09 Gerrit Renker <grenker@users.sourceforge.net>
* replace costly gettimeofday() with nanonsleep() See:
https://sourceforge.net/tracker/index.php?func=detail&aid=1940009&group_id=128336&atid=711373
2008-05-09 Jon Dugan <jdugan@x1024.net>
* change currLen to unsigned to squelch warning generated by Nathan's patch
2008-05-09 Nathan Jones <nmjones@users.sourceforge.net>
* prevent underflow when the amount of data to be transmitted (-n) is not a
multiple of the buffer size (-l) Patch:
https://sourceforge.net/tracker/index.php?func=detail&aid=1943432&group_id=128336&atid=711373
2008-04-08 Jon Dugan <jdugan@x1024.net>
* print report headers only once
* use appropriate report header for UDP tests
2008-04-07 Jon Dugan <jdugan@x1024.net>
* Add man page to autoconf goo
* Bump the version in autoconf
* Add missing LICENSE file
2008-04-04 Jon Dugan <jdugan@x1024.net>
* Incorporate help string fixes from Claus Klein
* Incorporate patch to restore header line in stats output from Claus Klein
* Reworked help string code to remove redundancy
* Add initial man page
2008-03-19 Stephen Hemminger <shemminger@linux-foundation.org>
* Add new option to allow selecting the sender congestion control
algorithm. This is a Linux specific extension useful for testing
other congestion control algorithms.
2008-03-19 Stephen Hemminger <shemminger@linux-foundation.org>
* Use sched_yield rather than usleep(0) as it behaves better on Linux
2008-03-12 Jon Dugan <jdugan@x1024.net>
* Rework Client::RunTCP so to update the stats for every packet if -i is set.
If -i is not set then the stats are only updated at the end of the run. This
was fallout of Andrew's patch.
* Added -Wall to CXXFLAGS and CFLAGS. This generated a lot of noise in the
autotools generated files.
* Added initial AUTHORS file. This is incomplete, but it's a start.
2008-03-12 Andrew Gallatin <gallatin@gmail.com>
* Rework threading. This is done by having a condition variable for the reporting
thread rather than using sched_yield and/or usleep.
* Address performance issues on non Linux systems by avoiding system calls
which are expensive on platforms other than Linux.
* Fixes from FreeBSD ports: max_size_t is now 64 bits and not Iperf does not
exit on ENOBUFS.
2007-10-01 JSE <jestabro@ncsa.uiuc.edu>
* ChangeLog file added
2007-10-05 JSE <jestabro@ncsa.uiuc.edu>
* Thread.c (thread_stop): "pthread_cancel (oldTID )" should be
"pthread_cancel (thread->mTID)"