On Wed, Apr 09, 2025 at 12:30:49PM -0700, Darrick J. Wong wrote:
> On Wed, Apr 09, 2025 at 09:55:45AM +0200, Christoph Hellwig wrote:
> > Signed-off-by: Christoph Hellwig <hch@xxxxxx>
> > ---
> > scrub/phase1.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/scrub/phase1.c b/scrub/phase1.c
> > index d03a9099a217..e71cab7b7d90 100644
> > --- a/scrub/phase1.c
> > +++ b/scrub/phase1.c
> > @@ -341,7 +341,8 @@ _("Kernel metadata repair facility is not available. Use -n to scrub."));
> > _("Unable to find log device path."));
> > return ECANCELED;
> > }
> > - if (ctx->mnt.fsgeom.rtblocks && ctx->fsinfo.fs_rt == NULL) {
> > + if (ctx->mnt.fsgeom.rtblocks && ctx->fsinfo.fs_rt == NULL &&
> > + !(ctx->mnt.fsgeom.flags & XFS_FSOP_GEOM_FLAGS_ZONED)) {
>
> Shouldn't this be gated on ctx->mnt.fsgeom.rtstart == 0 instead of
> !ZONED? I think we still want to be able to do media scans of zoned
> external rt devices.
It checks for ctx->fsinfo.fs_rt above, which is non-NULL for
external devices. But looking at rtstart might still be better, I'll
look into it.
