On Tue, May 03, 2022 at 03:50:09PM -0700, Darrick J. Wong wrote:
> On Wed, May 04, 2022 at 08:17:28AM +1000, Dave Chinner wrote:
> > @@ -985,6 +993,13 @@ xlog_cil_build_lv_chain(
> >
> > item = list_first_entry(&cil->xc_cil,
> > struct xfs_log_item, li_cil);
> > +
> > + if (test_bit(XFS_LI_WHITEOUT, &item->li_flags)) {
> > + list_move(&item->li_cil, whiteouts);
> > + trace_xfs_cil_whiteout_skip(item);
> > + continue;
> > + }
> > +
> > list_del_init(&item->li_cil);
> > if (!ctx->lv_chain)
> > ctx->lv_chain = item->li_lv;
> > @@ -1000,6 +1015,19 @@ xlog_cil_build_lv_chain(
> > }
> > }
> >
> > +static void
> > +xlog_cil_push_cleanup_whiteouts(
>
> Pushing cleanup whiteouts?
>
> Oh, clean up whiteouts as part of pushing CIL.
>
> I almost want to ask for a comment here:
>
> /* Remove log items from the CIL that have been elided from the checkpoint. */
> static void
> xlog_cil_push_cleanup_whiteouts(
>
> But fmeh, aside from my own momentary confusion this isn't that big of a
> deal.
Oh, fair comment - it's not consistent with other helpers that
are just named xlog_cil_<thing>, like xlog_cil_build_lv_chain().
I dropped the "push" out of the name.
> Reviewed-by: Darrick J. Wong <djwong@xxxxxxxxxx>
Thanks!
-Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
