[PATCH libnvme 2/1] nvme: Add generic connect parameter support detection

[Daniel Wagner]

On Mon, Sep 12, 2022 at 03:32:01PM +0300, Sagi Grimberg wrote:
> > > +int nvmf_check_param_support(const char *param, bool *supported)
> > > +{
> > 
> > I was wondering if t would be better to return all supported features in
> > one go. Not really sure about it though.
> 
> Why? it is already held in the kernel, why hold it again?

True, I am probably making it more complex than needed.

> > This triggered my question above why we don't return all supported
> > features up and have the caller added something like this?
> 
> Why? the caller doesn't need the entire array, it is just looking
> to check if a param is supported. We can move this out if someone
> actually needs the entire string or set of caps...

Right, let's address this when there is actually a user for it.