[Bug 230273] linux_select CALL is very slow (100 ms)
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230273 --- Comment #15 from Dmitry Chagin --- (In reply to Tatsuki Makino from comment #13) Does it take time to detect if the listening socket is readable, or were there simply no clients? select return 1, so 1 readfds is ready to accept. we should check timeout. from the first messgae: 0.09030 linux_select(0x1,0x7fffd310,0x0,0x0,0x7fffd390) = 1 (0x1) ^^ 0.09219 read(0,"f\0\0\0",4) = 4 (0x4) 0.104589092 linux_select(0x1,0x7fffd290,0x0,0x0,0x7fffd310) = 1 (0x1) ^^ supplied timevals differs, I need to know its values -- You are receiving this mail because: You are the assignee for the bug.
[Bug 230273] linux_select CALL is very slow (100 ms)
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230273 --- Comment #16 from Dmitry Chagin --- (In reply to Tatsuki Makino from comment #14) AFAIR, i fixed this some years ago -- You are receiving this mail because: You are the assignee for the bug.
[Bug 230273] linux_select CALL is very slow (100 ms)
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230273 --- Comment #17 from Tatsuki Makino --- (In reply to Dmitry Chagin from comment #15) comment #0 log is cut off below that select, so we don't know what it is. readfds are also different (on stack?) so it may be a different job with different functions that should not be compared in terms of time taken. (In reply to Dmitry Chagin from comment #16) If there is no problem there, we may need to limit the number of clients served by this program and see how it goes. This is to eliminate the possibility that the number of file descriptors has increased to a number that cannot be handled by linux_select(). -- You are receiving this mail because: You are the assignee for the bug.
[Bug 230273] linux_select CALL is very slow (100 ms)
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230273 --- Comment #18 from Dmitry Chagin --- (In reply to Tatsuki Makino from comment #17) As far as I see 1 descriptor is used everywhere, so the log has nothing to do I think -- You are receiving this mail because: You are the assignee for the bug.
