Re: [PATCH] xsplice: Don't perform multiple operations on same payload once work is scheduled.

From: "Jan Beulich" <JBeulich@suse.com>
To: Konrad Rzeszutek Wilk <konrad@kernel.org>
Cc: Stefano Stabellini <sstabellini@kernel.org>,
	Wei Liu <wei.liu2@citrix.com>,
	George Dunlap <George.Dunlap@eu.citrix.com>,
	Andrew Cooper <andrew.cooper3@citrix.com>,
	Ian Jackson <ian.jackson@eu.citrix.com>, Tim Deegan <tim@xen.org>,
	Ross Lagerwall <ross.lagerwall@citrix.com>,
	xen-devel@lists.xenproject.org, roger.pau@citrix.com
Subject: Re: [PATCH] xsplice: Don't perform multiple operations on same payload once work is scheduled.
Date: Fri, 29 Apr 2016 01:35:08 -0600	[thread overview]
Message-ID: <57232ACC02000078000E7157@prv-mh.provo.novell.com> (raw)
In-Reply-To: <20160429015212.GA12857@localhost.localdomain>

>>> On 29.04.16 at 03:52, <konrad@kernel.org> wrote:
> --- a/xen/common/xsplice.c
> +++ b/xen/common/xsplice.c
> @@ -1363,6 +1363,9 @@ static int xsplice_action(xen_sysctl_xsplice_action_t *action)
>          return PTR_ERR(data);
>      }
>  
> +    if ( data->rc == -EAGAIN ) /* schedule_work has been called. */
> +        goto out;

And nothing else can set data->rc to this value, now or in the
future? Because if that's possible, you'd deny any further actions
on that payload.

> @@ -1423,6 +1426,7 @@ static int xsplice_action(xen_sysctl_xsplice_action_t *action)
>          break;
>      }
>  
> + out:
>      spin_unlock(&payload_lock);
>  
>      return rc;

Taking both together, and looking at patch 2 of the original series, I'm
getting the impression you return success in that case rather than e.g.
-EBUSY or indeed -EAGAIN.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel