A Signal Profiler

As I wrote last week, recently I'm looking into SIGNAL handling of Linux, especially SIGSEGV.

I have some crash dump files to see, but I thought it better if we have a new but simple mechanism giving us statistics of SIGNALS sent in a running system.

At first, I thought 'sar -B' could be used to see how many segmentation faults were generated in a system. But, that didn't work well for me, and I wanted to see more direct information that how many SIGNALs were sent in total at each sampling time.
Thus, I wrote one. :)

There are some more TODOs to make it public, and it's a homework of this weekend. :)

Here is an example of the profiler.

# head -15 /proc/sigprof
SIG   COUNT
[ 0]          0
[ 1]          4
[ 2]         52
[ 3]          0
[ 4]          0
[ 5]          0
[ 6]          0
[ 7]          0
[ 8]          0
[ 9]          0
[10]          4
[11]   70384086  <== SIGSEGV : 70M SEGVs!? Because of the SEGV storm generator. :)
[12]         12
[13]          0