[RFC] Tunable elision patch for siddhesh/tunables

[Siddhesh Poyarekar]

On Tuesday 06 October 2015 07:39 PM, Paul E. Murphy wrote:
> My interpretation is that COMPAT_TUNABLE_* macros should only be used
> for porting existing tunables with an existing env var.

Ugh, sorry, I misread the code and thought you were introducing an
environment variable.  I see now that you're only introducing a new envp
parameter to tunable_register.  The WITH_ENV is fine for now.  If we go
the environment variable route, we can use that by default since we
always want envp to be available by some means or the other.

> I agree. The intent of these patches is to validate your existing
> work with tunables by using something clearly demanded by the glibc
> users.
> 
> What work remains on this branch before you open it for debate?

I need to split the patches into two sets, one that simply adds an
abstract framework without doing any work and another that opens up the
discussion about what the tunable interface should look like, i.e. one
envvar per variable, one envvar for all, some file, auxval passed by the
kernel, etc.

Siddhesh