FileBench fans, I have uploaded a new webrev for your reviewing pleasure at:
http://cr.opensolaris.org/~dreww/rawfilebug/ This fixes a collection of recently discovered bugs, including: 6673349 seg fault when variables defined after define stmt - caused by checking for raw files too early. 6549500 Filebench error reporting requires traditionalization 6673391 go_filebench exit code 1 Fixed by now exiting with code 0 on normal termination, and sending error messages to stderr rather than stdout. 6673928 creating large filesets is slow - fixed by modifying ipc.c to remember the index of the last allocated resource so it doesn't have to start over from index 0 each time. Makes a dramatic difference when a fileset with many entries is populated. (For example, 43 seconds becomes 3 seconds on my test machine with 100,000 entries in a fileset.) 6672924 The fileserver.f FileBench workload needs to create files too. - Two more workloads needed to be set to "mode quit alldone". And while fileserver.f could have been made to work that way, it was decided to have it create a file for every file deleted, so it will not run out of files to operate on. Also, made all 6 that quit early consistent in their "usage" messages. Enjoy! _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
