On Sun, 2011-07-24 at 16:10 +0300, Victor Goldenshtein wrote:
> Mostly from an energy consumption and cpu load perspective,
> it can be useful to stop a triggered scan earlier, when one
> finds the desired scan results.
> Adding the ability to stop a scan from user space by exposing
> new NL80211_CMD_SCAN_CANCEL API.
As I said to you privately, I don't think having this API before a
partial results notification is useful. Therefore, I don't think this
should be merged before a partial results notification patch.
A few comments on the code:
> +++ b/include/linux/nl80211.h
> @@ -491,6 +491,11 @@
> * this command may also be sent by the driver as an MLME event to
> * inform userspace of the new replay counter.
> *
> + * @NL80211_CMD_SCAN_CANCEL: Stop currently running scan (both sw and hw).
> + * This will eventually call ieee80211_scan_cancel() in mac80211 which
> + * will clean up scan_req resources. If we try to stop a not running
> + * scan it will return -ENOENT.
You must not mention mac80211 here. You're writing a comment about the
implementation, while a comment about the API/semantics is required.
> @@ -615,6 +620,7 @@ enum nl80211_commands {
>
> NL80211_CMD_SET_REKEY_OFFLOAD,
>
> + NL80211_CMD_SCAN_CANCEL,
> /* add new commands above here */
Please keep the blank line.
> /* used to define NL80211_CMD_MAX below */
> diff --git a/include/net/cfg80211.h b/include/net/cfg80211.h
> index 4bf101b..0d619b9 100644
> --- a/include/net/cfg80211.h
> +++ b/include/net/cfg80211.h
> @@ -1334,6 +1334,7 @@ struct cfg80211_gtk_rekey_data {
> * @set_ringparam: Set tx and rx ring sizes.
> *
> * @get_ringparam: Get tx and rx ring current and maximum sizes.
> + * @scan_cancel: Stop triggered scan, return 0 if successful.
> */
> struct cfg80211_ops {
> int (*suspend)(struct wiphy *wiphy, struct cfg80211_wowlan *wow);
> @@ -1516,6 +1517,7 @@ struct cfg80211_ops {
>
> int (*set_rekey_data)(struct wiphy *wiphy, struct net_device *dev,
> struct cfg80211_gtk_rekey_data *data);
> + void (*scan_cancel)(struct wiphy *wiphy, struct net_device *dev);
Since this struct isn't ABI it'd be more useful to put the scan_cancel
with the scan call.
> + .internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
> + NL80211_FLAG_NEED_RTNL,
whitespace?
> + if (!rdev->ops->scan_cancel)
> + return -EOPNOTSUPP;
> + if (!rdev->scan_req)
> + return -ENOENT;
> +
> + dev = rdev->scan_req->dev;
> + rdev->ops->scan_cancel(&rdev->wiphy, dev);
> + return 0;
> +}
So ... does this mean the driver is still expected to send a "scan
aborted" event? That would be useful to document ... not the mac80211
internals.
johannes
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
