Re: [linus:master] [NFSD] 39d432fc76: fsmark.files_per_sec -100.0% regression

[Jeff Layton]

On Thu, 2023-08-24 at 10:12 -0400, Benjamin Coddington wrote:
> On 24 Aug 2023, at 9:23, Chuck Lever wrote:
> 
> > On Thu, Aug 24, 2023 at 01:59:06PM +0800, kernel test robot wrote:
> > > 
> > > 
> > > hi, Chuck Lever,
> > > 
> > > Fengwei (CCed) helped us review this astonishing finding by fsmark tests,
> > > and doubt below part:
> > > -			nfsd4_end_grace(nn);
> > > +			trace_nfsd_end_grace(netns(file));
> > > 
> > > and confirmed if adding back:
> > > 			nfsd4_end_grace(nn);
> > > 
> > > the regression is gone and files_per_sec restore to 61.93.
> > 
> > As always, thanks for the report. However, this result is not
> > plausible. "end_grace" happens only once after a server reboot.
> > 
> > Can you confirm that the NFS server kernel is not crashing
> > during the test?
> 
> Does the removal of nfsd4_end_grace() here disable the ability of nfsdcltrack to
> terminate the grace period early on a first start of the server?
> 
> 

Yes. That should be the only effect (which is why we didn't notice it).
-- 
Jeff Layton <jlayton@xxxxxxxxxx>