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

RE: [Xen-devel] [PATCH][RFC] permit domU userspace to watch xenstore


  • To: "Michael LeMay" <mdlemay@xxxxxxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Cihula, Joseph" <joseph.cihula@xxxxxxxxx>
  • Date: Fri, 4 Aug 2006 13:48:32 -0700
  • Delivery-date: Fri, 04 Aug 2006 13:49:25 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Aca2LlE/vMi31VguRK+KKhBbyVw0sgBxen6Q
  • Thread-topic: [Xen-devel] [PATCH][RFC] permit domU userspace to watch xenstore

Michael,

There was some discussion on this in January
(http://lists.xensource.com/archives/cgi-bin/extract-mesg.cgi?a=xen-deve
l&m=2006-01&i=5901fecb838e65c42d197d246ec28e47%40cl.cam.ac.uk).  At that
time I looked at both this approach and pushing the watches to xenstore
(which currently only knows about per-domain watches and not
intra-domain, except for dom0).  My thinking was that doing the
multiplexing in the xenstore kernel module would add complexity and
fragility (you need to worry about apps DoS'ing the kernel by adding too
many watches, etc.).  So I proposed using the connection id as an
intra-domain id that xenstore would recognize and manage like it managed
domain watches.  This would have the side-effect of allowing the removal
of the current kernel module multiplexing, since they could also use
this field.

I realize that working code is worth more than a proposal, but perhaps
you could see if the advantages of this approach might be worth a little
re-coding.

Joe

On Wednesday, August 02, 2006 5:18 AM,  Michael LeMay <> wrote:

> Hello,
> 
> This patch allows userspace tools on a domU to setup a watch on
> the xenstore.  It does this by intercepting XS_WATCH requests written
> to /proc/xen/xenbus and then re-submitting the request to the
> in-kernel 
> xenstore interface, in
> linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_xs.c.  When a callback
> occurs, an in-kernel function is invoked, which then reconstructs a
> response in the format expected by userspace, and sends this response
> through /proc/xen/xenbus.
> 
> It was necessary to add some supporting infrastructure to
> linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c, such as an
> additional mutex to protect the response queue and a list of active
> watches associated with each connection.
> 
> 
> Signed-off-by: Michael LeMay <mdlemay@xxxxxxxxxxxxxx>
> 
> ---
> 
>  linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c |  114
>  +++++++++++++++++++ 1 file changed, 114 insertions(+)
> 
> diff -r eb8083d63198 -r 0c8a22ad7e46
> linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c ---
> a/linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c        Tue Aug
01
> 10:52:02 2006 -0400 +++
>       b/linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_dev.c  Wed Aug
02
>       08:20:06 2006 -0400 @@ -58,6 +58,9 @@ struct xenbus_dev_data {
/*
> In-progress transaction. */ struct list_head transactions; 
> 
> +     /* Active watches. */
> +     struct list_head watches;
> +
>       /* Partial request. */
>       unsigned int len;
>       union {
> @@ -70,6 +73,8 @@ struct xenbus_dev_data {
>       char read_buffer[PAGE_SIZE];
>       unsigned int read_cons, read_prod;
>       wait_queue_head_t read_waitq;
> +
> +     struct mutex reply_mutex;
>  };
> 
>  static struct proc_dir_entry *xenbus_dev_intf;
> @@ -100,13 +105,59 @@ static void queue_reply(struct xenbus_de
>  {
>       int i;
> 
> +     mutex_lock(&u->reply_mutex);
> +
>       for (i = 0; i < len; i++, u->read_prod++)
>               u->read_buffer[MASK_READ_IDX(u->read_prod)] = data[i];
> 
>       BUG_ON((u->read_prod - u->read_cons) > sizeof(u->read_buffer));
> 
> +     mutex_unlock(&u->reply_mutex);
> +
>       wake_up(&u->read_waitq);
>  }
> +
> +struct watch_adapter
> +{
> +     struct list_head list;
> +     struct xenbus_watch watch;
> +     struct xenbus_dev_data *dev_data;
> +     char *token;
> +};
> +
> +static void free_watch_adapter (struct watch_adapter *watch)
> +{
> +     kfree(watch->watch.node);
> +     kfree(watch->token);
> +     kfree(watch);
> +}
> +
> +static void watch_fired(struct xenbus_watch *watch,
> +                     const char **vec,
> +                     unsigned int len)
> +{
> +     struct watch_adapter *adap =
> +            container_of(watch, struct watch_adapter, watch);
> +     struct xsd_sockmsg hdr;
> +     const char *path, *token;
> +     int path_len, tok_len, body_len;
> +
> +     path = vec[XS_WATCH_PATH];
> +     token = adap->token;
> +
> +     path_len = strlen(path) + 1;
> +     tok_len = strlen(token) + 1;
> +     body_len = path_len + tok_len;
> +
> +     hdr.type = XS_WATCH_EVENT;
> +     hdr.len = body_len;
> +
> +     queue_reply(adap->dev_data, (char *)&hdr, sizeof(hdr));
> +     queue_reply(adap->dev_data, (char *)path, path_len);
> +     queue_reply(adap->dev_data, (char *)token, tok_len);
> +}
> +
> +static LIST_HEAD(watch_list);
> 
>  static ssize_t xenbus_dev_write(struct file *filp,
>                               const char __user *ubuf,
> @@ -116,6 +167,9 @@ static ssize_t xenbus_dev_write(struct f
>       struct xenbus_dev_transaction *trans = NULL;
>       uint32_t msg_type;
>       void *reply;
> +     char *path, *token;
> +     struct watch_adapter *watch, *tmp_watch;
> +     int err;
> 
>       if ((len + u->len) > sizeof(u->u.buffer))
>               return -EINVAL;
> @@ -169,6 +223,56 @@ static ssize_t xenbus_dev_write(struct f
>               kfree(reply);
>               break;
> 
> +     case XS_WATCH:
> +     case XS_UNWATCH:
> +             path = u->u.buffer + sizeof(u->u.msg);
> +             token = memchr(path, 0, u->u.msg.len);
> +             if (token == NULL)
> +                     return -EILSEQ;
> +             token++;
> +
> +             if (msg_type == XS_WATCH) {
> +                     static const char * XS_WATCH_RESP = "OK";
> +                     struct xsd_sockmsg hdr;
> +
> +                     watch = kmalloc(sizeof(*watch), GFP_KERNEL);
> +                     watch->watch.node = kmalloc(strlen(path)+1,
> +                                                    GFP_KERNEL);
> +                     strcpy((char *)watch->watch.node, path);
> +                     watch->watch.callback = watch_fired;
> +                     watch->token = kmalloc(strlen(token)+1,
GFP_KERNEL);
> +                     strcpy(watch->token, token);
> +                     watch->dev_data = u;
> +
> +                     err = register_xenbus_watch(&watch->watch);
> +                     if (err) {
> +                             free_watch_adapter(watch);
> +                             return err;
> +                     }
> +
> +                     list_add(&watch->list, &u->watches);
> +
> +                     hdr.type = XS_WATCH;
> +                     hdr.len = strlen(XS_WATCH_RESP) + 1;
> +                     queue_reply(u, (char *)&hdr, sizeof(hdr));
> +                     queue_reply(u, (char *)XS_WATCH_RESP, hdr.len);
> +             } else {
> +                     list_for_each_entry_safe(watch, tmp_watch,
> +                                                 &u->watches, list) {
> +                             if (!strcmp(watch->token, token) &&
> +                                 !strcmp(watch->watch.node, path))
> +                                     break;
> +                             {
> +
unregister_xenbus_watch(&watch->watch);
> +                                     list_del(&watch->list);
> +                                     free_watch_adapter(watch);
> +                                     break;
> +                             }
> +                     }
> +             }
> +
> +             break;
> +
>       default:
>               return -EINVAL;
>       }
> @@ -191,7 +295,10 @@ static int xenbus_dev_open(struct inode
>               return -ENOMEM;
> 
>       INIT_LIST_HEAD(&u->transactions);
> +     INIT_LIST_HEAD(&u->watches);
>       init_waitqueue_head(&u->read_waitq);
> +
> +     mutex_init(&u->reply_mutex);
> 
>       filp->private_data = u;
> 
> @@ -202,11 +309,18 @@ static int xenbus_dev_release(struct ino
>  {
>       struct xenbus_dev_data *u = filp->private_data;
>       struct xenbus_dev_transaction *trans, *tmp;
> +     struct watch_adapter *watch, *tmp_watch;
> 
>       list_for_each_entry_safe(trans, tmp, &u->transactions, list) {
>               xenbus_transaction_end(trans->handle, 1);
>               list_del(&trans->list);
>               kfree(trans);
> +     }
> +
> +     list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) {
> +             unregister_xenbus_watch(&watch->watch);
> +             list_del(&watch->list);
> +             free_watch_adapter(watch);
>       }
> 
>       kfree(u);
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.