On Mon, 2020-03-30 at 19:56 +0800, Yan, Zheng wrote:
> write can stuck at waiting for larger max_size in following sequence of
> events:
>
> - client opens a file and writes to position 'A' (larger than unit of
> max size increment)
> - client closes the file handle and updates wanted caps (not wanting
> file write caps)
> - client opens and truncates the file, writes to position 'A' again.
>
> At the 1st event, client set inode's requested_max_size to 'A'. At the
> 2nd event, mds removes client's writable range, but client does not reset
> requested_max_size. At the 3rd event, client does not request max size
> because requested_max_size is already larger than 'A'.
>
> Signed-off-by: "Yan, Zheng" <zyan@xxxxxxxxxx>
> ---
> fs/ceph/caps.c | 29 +++++++++++++++++++----------
> 1 file changed, 19 insertions(+), 10 deletions(-)
>
> diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
> index f8b51d0c8184..61808793e0c0 100644
> --- a/fs/ceph/caps.c
> +++ b/fs/ceph/caps.c
> @@ -1363,8 +1363,12 @@ static void __prep_cap(struct cap_msg_args *arg, struct ceph_cap *cap,
> arg->size = inode->i_size;
> ci->i_reported_size = arg->size;
> arg->max_size = ci->i_wanted_max_size;
> - if (cap == ci->i_auth_cap)
> - ci->i_requested_max_size = arg->max_size;
> + if (cap == ci->i_auth_cap) {
> + if (want & CEPH_CAP_ANY_FILE_WR)
> + ci->i_requested_max_size = arg->max_size;
> + else
> + ci->i_requested_max_size = 0;
> + }
>
> if (flushing & CEPH_CAP_XATTR_EXCL) {
> arg->old_xattr_buf = __ceph_build_xattrs_blob(ci);
> @@ -3343,10 +3347,6 @@ static void handle_cap_grant(struct inode *inode,
> ci->i_requested_max_size = 0;
> }
> wake = true;
> - } else if (ci->i_wanted_max_size > ci->i_max_size &&
> - ci->i_wanted_max_size > ci->i_requested_max_size) {
> - /* CEPH_CAP_OP_IMPORT */
> - wake = true;
> }
> }
>
> @@ -3422,9 +3422,18 @@ static void handle_cap_grant(struct inode *inode,
> fill_inline = true;
> }
>
> - if (le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) {
> + if (ci->i_auth_cap == cap &&
> + le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) {
> if (newcaps & ~extra_info->issued)
> wake = true;
> +
> + if (ci->i_requested_max_size > max_size ||
> + !(le32_to_cpu(grant->wanted) & CEPH_CAP_ANY_FILE_WR)) {
> + /* re-request max_size if necessary */
> + ci->i_requested_max_size = 0;
> + wake = true;
> + }
> +
> ceph_kick_flushing_inode_caps(session, ci);
> spin_unlock(&ci->i_ceph_lock);
> up_read(&session->s_mdsc->snap_rwsem);
> @@ -3882,9 +3891,6 @@ static void handle_cap_import(struct ceph_mds_client *mdsc,
> __ceph_remove_cap(ocap, (ph->flags & CEPH_CAP_FLAG_RELEASE));
> }
>
> - /* make sure we re-request max_size, if necessary */
> - ci->i_requested_max_size = 0;
> -
> *old_issued = issued;
> *target_cap = cap;
> }
> @@ -4318,6 +4324,9 @@ int ceph_encode_inode_release(void **p, struct inode *inode,
> cap->issued &= ~drop;
> cap->implemented &= ~drop;
> cap->mds_wanted = wanted;
> + if (cap == ci->i_auth_cap &&
> + !(wanted & CEPH_CAP_ANY_FILE_WR))
> + ci->i_requested_max_size = 0;
> } else {
> dout("encode_inode_release %p cap %p %s"
> " (force)\n", inode, cap,
Thanks Zheng. I assume this is a regression in the "don't request caps
for idle open files" series? If so, is there a commit that definitively
broke it? It'd be good to add a Fixes: tag for that if we can to help
backporters.
Thanks,
--
Jeff Layton <jlayton@xxxxxxxxxx>
