Re: [GIT PULL] tracing: make signal tracepoints more useful

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 2012-01-17 at 11:02 +0100, Ingo Molnar wrote:
>  
> > The kernel does this all the time. We have syscalls that may 
> > extend the data structure. [...]
> 
> That is not true *AT ALL* in such an unqualified manner. Steve, 
> stop being stupid.

OK, Ingo I think we had a little miscommunication here.

I forgot what I originally wrote, and you cut off too much in your
reply. What I originally said:


"The kernel does this all the time. We have syscalls that may extend the
data structure. This is a common practice. Any app that depends on a
data structure remaining the same size for no good reason is broken by
design."

If you took the "Any app that depends on a data structure remaining the
same size for no good reason is broken by design" was suppose to mean,
any app that depends on syscall data structures is broken by design, I
would agree, that statement is stupid. I didn't write that well, as my
thought process switched back to reading tracepoints and wasn't meant to
be about syscalls there. I didn't articulate that properly.

As you can tell, I was a little upset at being called stupid this
morning. If you took that statement to mean about syscalls, I understand
your calling me stupid, and I apologize for being a bit snippy. I like
to keep LKML discussions technical and not something for name calling.

-- Steve