Re: [GIT PULL v2] bkl tracepoints + filter regex support

[Frederic Weisbecker <fweisbec@gmail.com>]

On Sat, Sep 26, 2009 at 11:47:06AM -0400, Frank Ch. Eigler wrote:
> 
> Hi -
> 
> 
> Frederic Weisbecker <fweisbec@gmail.com> writes:
> 
> >> > That said, the future plans have evolved, and I'm fine if you have
> >> > changed your opinion and think about a better way to develop this.
> 
> >> No, but the thing is, IF we're going to freeze this into ABI, then
> >> there's no second chances.
> 
> > Right. Once it becomes an ioctl, it becomes an ABI :-/
> 
> Are y'all convinced that a little ascii language parser/interpreter is
> the right thing to put into the kernel, as opposed to bytecode (with a
> userspace compiler)?
> 
> - FChE


We need something directly understandable from the kernel interfaces.
As an example we don't want the ftrace users to bother about compiling
a string before applying it on an event filter through ftrace debugfs
interface, we want it directly usable without middle-state in the worklow.

That's also the same for perf syscalls users.

Moreover it's a really small language, based on predicates. Fetching
strings against bytecodes won't make that much differences in the
amount of kernel code to maintain.