A approach for 2.0.5 bugs is to try iperf 2.0.8 as many bugs were addressed there
https://sourceforge.net/projects/iperf2/?source=directory Another option is iperf3 https://github.com/esnet/iperf I'd prefer not to analyze 2.0.5 bugs unless 2.0.8 inherited them. Thanks, Bob -----Original Message----- From: Uffe Vedenbrant [mailto:[email protected]] Sent: Thursday, February 05, 2015 11:41 AM To: [email protected] Subject: Re: [Iperf-users] iperf2.0.9 plans Hi! As you asked for bug reports..:-) When testing older versions on different platforms i ran into a lockup bug that caused iperf2 to hang and a need to kill the process and start it again. ( Around 2.05+ Jan 2013 ) Interesting it seemed that most error reports that i could find for a lot of the older releases talked about the same issue. Just the way people described the problem differed.. Most of them talked about time out when testing.. And the reports regarding the issue ( similiar reports ) that i could trace was several years holder than the release i was using... Basically it was possible to have a single path tested but when doing duplex path (more advanced) the server needed to be restarted.. I tested with Windows, linux and macosx clients towards a remote linux client. Then i compiled a macosx client and a local linux client/server with more or less the same result. I.e. server hang. I'm not sure if this problem is fixed.. Below is what i sent to [email protected] at the time when i ran into the problem.. Googling for iperf and FUTEX_WAKE_PRIVATE will give results for people running into the same issue with older versions.. ---------------- Some more stuff when doing some strace -f -p <processid> [pid 24811] recvfrom(7, <unfinished ...> [pid 24810] <... write resumed> ) = 131072 [pid 24810] write(10, "01234567890123456789012345678901"..., 131072) = 131072 [pid 24810] write(10, "01234567890123456789012345678901"..., 131072) = 131072 [pid 24810] futex(0x610490, FUTEX_WAKE_PRIVATE, 1) = 0 [pid 24810] futex(0x610464, FUTEX_WAIT_PRIVATE, 2, NULL <unfinished ...> [pid 24811] <... recvfrom resumed> "89012345678901234567890123456789"..., 131072, 0, NULL, NULL) = 4344 This is the last i see from the write thread.. Ill see if there is a newer version or possible patches on @ to download and compile as when googling on "FUTEX_WAIT_PRIVATE iperf" more people seems to have the exact same problem..;-) It seems to be the server side that has a problem as well and not the client side.. And it might be that this only occur when doing -d and not -r ----------------------------- Regards/Uffe On 2015-02-05 18:10, Bob (Robert) McMahon wrote: > Hi, > > Now that iperf 2.0.8 is out I've been getting requests for new > features and bug fixes. Here are my current plans: > > o) Support write error count to UDP client reports > > o) Add support for TTL information into server reports > > o) Add support for ToS information into server reports > > o) Fix -o (redirect to file) option > > o) Integrate simple IGMP/MLD querier into client to support testing > v4/v6 multicast > > o) Thoroughly test v6 and fix bugs found > > o) Support new traffic option giving packet burst length, interburst > gap, and interpacket gap > > o) Come up with an analysis technique on the server to report how well > the above is able to be reconstructed by the server app > > Please feel free to contact me if other things are needed. Also, if > you have ideas on analyzing the traffic pattern per the above in a > simple way, please share. > > Bob McMahon > > Broadcom Senior Principal SW Engineer, Wi-Fi Division > > > > ---------------------------------------------------------------------- > -------- Dive into the World of Parallel Programming. The Go Parallel > Website, sponsored by Intel and developed in partnership with Slashdot > Media, is your hub for all things parallel software development, from > weekly thought leadership blogs to news, videos, case studies, > tutorials and more. Take a look and join the conversation now. > http://goparallel.sourceforge.net/ > > > > _______________________________________________ > Iperf-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/iperf-users > ------------------------------------------------------------------------------ Dive into the World of Parallel Programming. The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Iperf-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/iperf-users ------------------------------------------------------------------------------ Dive into the World of Parallel Programming. The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Iperf-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/iperf-users
