how to explain the results?

[Ming Z. <mi...@el...>]

i tried to profile the iscsi target and see where is the bottleneck, or
cpu spend most time in it.

so i do

opcontrol --vmlinux=/usr/src/linux/vmlinux
opcontrol -s

run iscsi workload from initiator dump 4GB data.

opcontrol -d
opcontrol -t

then do 
opreport -l -p ~/b.0.4.12/kernel/ iscsi_trgt.ko --sort=sample -o olog

iscsi_trgt.ko is the iscsi target module.

this result (attached) show lots of "samples and %". but i donot know
what they represent for. and they wrap horribly and could not read and
compare at all.

how to have a simple cleaner result to read?

also if i want to check whether most time are spend on kmalloc or
page_alloc, what should i do?

thanks!


Ming