Re: [cmake] Problem with dependencies for configure checks

From: Peter Dimov (pdimov_at_[hidden])
Date: 2021-06-04 15:32:07

Niall Douglas wrote:
> On 04/06/2021 15:54, Peter Dimov via Boost wrote:
> > A kernel call is infinitely more expensive than checking a number
> > against ENOSYS. It's absolutely impossible for the test to ever be a
> > performance issue.
> 
> Well, that depends.
> 
> If you're on a kernel that has no statx syscall, then the userspace syscall
> dispatcher will spot the unknown syscall number, and you'll get back ENOSYS
> very quickly, as you point out.
> 
> But if you're on a kernel with statx syscall, and instead it is the filing system
> which doesn't support statx, then that's a full fat syscall just to get back
> ENOSYS.

Sure, but the alternative is to do that same fat syscall just to fail, as Filesystem
currently does. The performance concerns in our case are about the scenario
where the syscall doesn't fail, because that's the case that works. The
performance of something that doesn't work is irrelevant.