I've just noticed an ssh mux process using close to 100% cpu, ktrace
shows this repeating in a tight loop,
...
5687 ssh CALL select(0xb,0x91210235500,0x91210235980,0,0)
5687 ssh STRU struct fd_set { 3-7 }
5687 ssh STRU struct fd_set { }
5687 ssh STRU struct fd_set { 7 }
5687 ssh STRU struct fd_set { }
5687 ssh RET select 1
5687 ssh CALL gettimeofday(0x7f7ffffeb910,0)
5687 ssh STRU struct timeval { 1360110940.345853 }
5687 ssh RET gettimeofday 0
5687 ssh CALL read(0x7,0x7f7ffffe78d0,0x4)
5687 ssh RET read 0
5687 ssh CALL gettimeofday(0x7f7ffffeb820,0)
5687 ssh STRU struct timeval { 1360110940.345876 }
5687 ssh RET gettimeofday 0
5687 ssh CALL select(0xb,0x91210235500,0x91210235980,0,0)
5687 ssh STRU struct fd_set { 3-7 }
...
5687 ?? Rs/1 438:06.37 ssh: /tmp/ssh-multi-XXX-22-sthen [mux] (ssh)
I'm not sure of the circumstances triggering this though it wouldn't
be a big surprise if the connection to the server had dropped out for a
few minutes.
Breaking into it with gdb, a backtrace looks like this:
#0 0x000009120f924f1a in read () at <stdin>:2
#1 0x000009100c43befa in channel_input_data () from /usr/bin/ssh
#2 0x000009100c43c09b in channel_input_data () from /usr/bin/ssh
#3 0x000009100c4402c6 in channel_input_close_confirmation () from /usr/bin/ssh
#4 0x000009100c412acf in client_simple_escape_filter () from /usr/bin/ssh
#5 0x000009100c40c1fb in ?? () from /usr/bin/ssh
#6 0x000009100c40a061 in ?? () from /usr/bin/ssh
#7 0x0000000000000002 in ?? ()
#8 0x00007f7ffffeebf0 in ?? ()
#9 0x00007f7ffffeebf4 in ?? ()
#10 0x0000000000000000 in ?? ()
