Re: [PATCH] libceph: fix misleading ceph_osdc_cancel_request() comment

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 2022-05-17 at 11:55 +0200, Ilya Dryomov wrote:
> cancel_request() never guaranteed that after its return the OSD
> client would be completely done with the OSD request.  The callback
> (if specified) can still be invoked and a ref can still be held.
> 
> Signed-off-by: Ilya Dryomov <idryomov@xxxxxxxxx>
> ---
>  net/ceph/osd_client.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c
> index 4b88f2a4a6e2..9d82bb42e958 100644
> --- a/net/ceph/osd_client.c
> +++ b/net/ceph/osd_client.c
> @@ -4591,8 +4591,13 @@ int ceph_osdc_start_request(struct ceph_osd_client *osdc,
>  EXPORT_SYMBOL(ceph_osdc_start_request);
>  
>  /*
> - * Unregister a registered request.  The request is not completed:
> - * ->r_result isn't set and __complete_request() isn't called.
> + * Unregister request.  If @req was registered, it isn't completed:
> + * r_result isn't set and __complete_request() isn't invoked.
> + *
> + * If @req wasn't registered, this call may have raced with
> + * handle_reply(), in which case r_result would already be set and
> + * __complete_request() would be getting invoked, possibly even
> + * concurrently with this call.
>   */
>  void ceph_osdc_cancel_request(struct ceph_osd_request *req)
>  {

Reviewed-by: Jeff Layton <jlayton@xxxxxxxxxx>




[Index of Archives]     [CEPH Users]     [Ceph Large]     [Ceph Dev]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux