On Wed, Aug 3, 2022 at 1:15 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
>
> On Wed, 2022-08-03 at 09:13 +0200, Ilya Dryomov wrote:
> > On Thu, Jun 30, 2022 at 10:21 PM Jeff Layton <jlayton@xxxxxxxxxx> wrote:
> > >
> > > This function always returns 0, and ignores the nofail boolean. Drop the
> > > nofail argument, make the function void return and fix up the callers.
> > >
> > > Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > ---
> > > drivers/block/rbd.c | 6 +++---
> > > fs/ceph/addr.c | 32 ++++++++++++--------------------
> > > fs/ceph/file.c | 32 +++++++++++++-------------------
> > > include/linux/ceph/osd_client.h | 5 ++---
> > > net/ceph/osd_client.c | 15 ++++++---------
> > > 5 files changed, 36 insertions(+), 54 deletions(-)
> > >
> > > diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
> > > index 91e541aa1f64..a8af0329ab77 100644
> > > --- a/drivers/block/rbd.c
> > > +++ b/drivers/block/rbd.c
> > > @@ -1297,7 +1297,7 @@ static void rbd_osd_submit(struct ceph_osd_request *osd_req)
> > > dout("%s osd_req %p for obj_req %p objno %llu %llu~%llu\n",
> > > __func__, osd_req, obj_req, obj_req->ex.oe_objno,
> > > obj_req->ex.oe_off, obj_req->ex.oe_len);
> > > - ceph_osdc_start_request(osd_req->r_osdc, osd_req, false);
> > > + ceph_osdc_start_request(osd_req->r_osdc, osd_req);
> > > }
> > >
> > > /*
> > > @@ -2081,7 +2081,7 @@ static int rbd_object_map_update(struct rbd_obj_request *obj_req, u64 snap_id,
> > > if (ret)
> > > return ret;
> > >
> > > - ceph_osdc_start_request(osdc, req, false);
> > > + ceph_osdc_start_request(osdc, req);
> > > return 0;
> > > }
> > >
> > > @@ -4768,7 +4768,7 @@ static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
> > > if (ret)
> > > goto out_req;
> > >
> > > - ceph_osdc_start_request(osdc, req, false);
> > > + ceph_osdc_start_request(osdc, req);
> > > ret = ceph_osdc_wait_request(osdc, req);
> > > if (ret >= 0)
> > > ceph_copy_from_page_vector(pages, buf, 0, ret);
> > > diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
> > > index fe6147f20dee..66dc7844fcc6 100644
> > > --- a/fs/ceph/addr.c
> > > +++ b/fs/ceph/addr.c
> > > @@ -357,9 +357,7 @@ static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq)
> > > req->r_inode = inode;
> > > ihold(inode);
> > >
> > > - err = ceph_osdc_start_request(req->r_osdc, req, false);
> > > - if (err)
> > > - iput(inode);
> > > + ceph_osdc_start_request(req->r_osdc, req);
> > > out:
> > > ceph_osdc_put_request(req);
> > > if (err)
> >
> > Hi Jeff,
> >
> > I'm confused by this err != 0 check. Previously err was set to 0
> > by ceph_osdc_start_request() and netfs_subreq_terminated() was never
> > called after an OSD request submission. Now it is called, but only if
> > len != 0?
> >
> > I see that netfs_subreq_terminated() accepts either the amount of data
> > transferred or an error code but it also has some transferred_or_error
> > == 0 handling which this check effectively disables. And do we really
> > want to account for transferred data before the transfer occurs?
> >
>
> No we don't. I think you're correct. What I'm not sure of is why this
> doesn't cause test failures all over the place.
This is due to "libceph: add new iov_iter-based ceph_msg_data_type and
ceph_osd_data_type" and "ceph: use osd_req_op_extent_osd_iter for netfs
reads" end up fixing it later by removing iov_iter_get_pages_alloc()
call. I think because these commits were backed out of testing at one
point and re-added later, the order got messed up by accident.
Thanks,
Ilya
