Re: 8ee579abe09e - main - zfs: fall back if block_cloning feature is disabled

In message <CAM5tNy7Oh=ftiM6BToo0LRitF-w3EsX-WznjMqZptR1gWs9jvw@mail.gmail.c
om>
, Rick Macklem writes:
> On Tue, Apr 4, 2023 at 12:31=E2=80=AFPM Cy Schubert <Cy.Schubert@cschubert.=
> com> wrote:
> >
> > CAUTION: This email originated from outside of the University of Guelph. =
> Do not click links or open attachments unless you recognize the sender and =
> know the content is safe. If in doubt, forward suspicious emails to IThelp@=
> uoguelph.ca
> >
> >
> > On Tue, 4 Apr 2023 17:54:28 +0000
> > "Pokala, Ravi" <rpokala@panasas.com> wrote:
> >
> > > Cy,
> > >
> > > The patch adds 'bool done_outvp', unconditionally sets it to 'true', an=
> d then later has a check for 'if (!done_outvp)'. Since there is no interven=
> ing place where 'done_outvp' could be set to 'false', that check will never=
>  succeed and that branch is unreachable.
> >
> > It's set to false at line 6454, in the loop locking vnodes.
> >
> > >
> > > Or am I mis-reading something?
> >
> > Maybe Rick can explain but all we're doing is ensuring that the first
> > part of the loop is executed only first time through. We could invert it
> > and save setting it to false every loop.
> Yep. All I did was copy the first lines in the loop up to before the loop
> so that it would be done once to lock outvp before the check on it.
> The bool just avoids doing those lines for the first loop iteration.
>
> And, yes, you could reorganize the loop to avoid using the bool
> to skip the lines on the first iteration, but I think the code would be
> more confusing with the loop reorganized.
> However, if someone wants to re-write it, I have no problem with that.

Agreed. Having looked at refactoring it, what did result was more confusing 
with no meaningful saving of cycles spent.


-- 
Cheers,
Cy Schubert <Cy.Schubert@cschubert.com>
FreeBSD UNIX:  <cy@FreeBSD.org>   Web:  https://FreeBSD.org
NTP:           <cy@nwtime.org>    Web:  https://nwtime.org

			e^(i*pi)+1=0