[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <361921e4-0f73-3dce-302a-dfe76a9b0d77@suse.com>
Date: Fri, 14 Sep 2018 11:01:14 +0200
From: Juergen Gross <jgross@...e.com>
To: Dongli Zhang <dongli.zhang@...cle.com>,
xen-devel@...ts.xenproject.org, linux-kernel@...r.kernel.org
Cc: boris.ostrovsky@...cle.com, paul.durrant@...rix.com,
wei.liu2@...rix.com, konrad.wilk@...cle.com, roger.pau@...rix.com,
srinivas.eeda@...cle.com
Subject: Re: [PATCH 3/6] xenbus: dispatch per-domU watch event to per-domU
xenwatch thread
On 14/09/18 09:34, Dongli Zhang wrote:
> This is the 3rd patch of a (6-patch) patch set.
>
> This patch dispatches the watch event to per-domU xenwatch thread when the
> event meets all of below conditions:
>
> 1. The watch is registered to use xenwatch multithreading feature and the
> get_domid() method returns valid domid.
> 2. There is xenwatch thread (mtwatch domain) available for the domid
> obtained from get_domid() method.
> 3. The xenwatch thread is forked successfully by kthread_run() during
> initialization and therefore its state should be MTWATCH_DOMAIN_UP.
>
> Signed-off-by: Dongli Zhang <dongli.zhang@...cle.com>
> ---
> drivers/xen/xenbus/xenbus_xs.c | 53 ++++++++++++++++++++++++++++++++++++++----
> 1 file changed, 49 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/xen/xenbus/xenbus_xs.c b/drivers/xen/xenbus/xenbus_xs.c
> index 741dc54..7335e19 100644
> --- a/drivers/xen/xenbus/xenbus_xs.c
> +++ b/drivers/xen/xenbus/xenbus_xs.c
> @@ -896,6 +896,32 @@ static struct xenbus_watch *find_watch(const char *token)
> return NULL;
> }
>
> +static int xs_watch_insert_event(struct xs_watch_event *event, domid_t domid)
> +{
> + struct mtwatch_domain *domain;
> + int ret = -1;
> +
> + rcu_read_lock();
> +
> + domain = mtwatch_find_domain(domid);
> + if (!domain) {
> + rcu_read_unlock();
> + return ret;
> + }
> +
> + spin_lock(&domain->events_lock);
> + if (domain->state == MTWATCH_DOMAIN_UP) {
> + list_add_tail(&event->list, &domain->events);
> + wake_up(&domain->events_wq);
> + ret = 0;
> + }
> + spin_unlock(&domain->events_lock);
> +
> + rcu_read_unlock();
> +
> + return ret;
> +}
> +
> int xs_watch_msg(struct xs_watch_event *event)
> {
> if (count_strings(event->body, event->len) != 2) {
> @@ -908,10 +934,29 @@ int xs_watch_msg(struct xs_watch_event *event)
> spin_lock(&watches_lock);
> event->handle = find_watch(event->token);
> if (event->handle != NULL) {
> - spin_lock(&watch_events_lock);
> - list_add_tail(&event->list, &watch_events);
> - wake_up(&watch_events_waitq);
> - spin_unlock(&watch_events_lock);
> + domid_t domid = 0;
I won't repeat it again (including following patches): use another
special value like DOMID_SELF.
Juergen
Powered by blists - more mailing lists