On Mon, Nov 07, 2016 at 04:12:18PM +0800, tang.wenjun3@xxxxxxxxxx wrote:
> From: 10144149 <tang.wenjun3@xxxxxxxxxx>
>
> uev_update_path locked &vecs->lock, and it will get lock twice in uev_add_path;
> and i think it better to retigger add path uevent directly in check_path.
Good catch on the deadlock, but I'm not sure if triggering multiple
"add" events is a good idea. Generally, there's supposed to only be one
add event, and different actions can get run on them then on the change
events. Unless someone knows that this is o.k. I'd rather just break out
the reinitialization code from uev_add_path to its own function and call
that directly both uev_add_path and uev_update_path
-Ben
> ---
> multipathd/main.c | 10 ++++------
> 1 file changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/multipathd/main.c b/multipathd/main.c
> index d6f081f..3b1bea8 100644
> --- a/multipathd/main.c
> +++ b/multipathd/main.c
> @@ -994,9 +994,7 @@ uev_update_path (struct uevent *uev, struct vectors * vecs)
> pp->wwid_changed = 0;
> }
>
> - if (pp->initialized == INIT_REQUESTED_UDEV)
> - retval = uev_add_path(uev, vecs);
> - else if (mpp && ro >= 0) {
> + if (mpp && ro >= 0) {
> condlog(2, "%s: update path write_protect to '%d' (uevent)", uev->kernel, ro);
>
> if (mpp->wait_for_udev)
> @@ -1504,12 +1502,12 @@ check_path (struct vectors * vecs, struct path * pp, int ticks)
> put_multipath_config(conf);
> if (!pp->mpp && pp->initialized == INIT_MISSING_UDEV &&
> pp->retriggers < retrigger_tries) {
> - condlog(2, "%s: triggering change event to reinitialize",
> + condlog(2, "%s: triggering add event to reinitialize",
> pp->dev);
> pp->initialized = INIT_REQUESTED_UDEV;
> pp->retriggers++;
> - sysfs_attr_set_value(pp->udev, "uevent", "change",
> - strlen("change"));
> + sysfs_attr_set_value(pp->udev, "uevent", "add",
> + strlen("add"));
> return 0;
> }
>
> --
> 2.8.1.windows.1
>
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
