On Wed, Dec 21, 2011 at 2:06 PM, Eyal Shapira <eyal@xxxxxxxxxx> wrote:
> mac80211 requires that drv_sched_scan_stop
> will always succeed. We have a few possible errors
> such as OOM, failure to ELP wakeup, fail in the FW command.
> Instead of no-op , trigger a recovery which would
> mark sched scan as stopped and clear up any bad FW state.
>
> Signed-off-by: Eyal Shapira <eyal@xxxxxxxxxx>
> ---
this patch seems a bit redundant.
> ret = wl1271_ps_elp_wakeup(wl);
> - if (ret < 0)
> + if (ret < 0) {
> + wl12xx_queue_recovery_work(wl);
> goto out;
> + }
>
we enqueue recovery in wl1271_ps_elp_wakeup() in most error paths.
> - /* FIXME: what to do if alloc'ing to stop fails? */
> stop = kzalloc(sizeof(*stop), GFP_KERNEL);
> if (!stop) {
> wl1271_error("failed to alloc memory to send sched scan stop");
> - return;
> + goto recovery;
> }
not sure if initiating recovery here will really help. it'll probably
fail as well, and good chances we are going to panic soon anyway :)
> ret = wl1271_cmd_send(wl, CMD_STOP_PERIODIC_SCAN, stop,
> sizeof(*stop), 0);
> +
> + kfree(stop);
> +
> if (ret < 0) {
> wl1271_error("failed to send sched scan stop command");
> - goto out_free;
> + goto recovery;
> }
>
wl1271_cmd_send enqueues recovery work on error as well.
anyway, the major thing i don't like here is handling the sched_stop
in a different manner than the rest of the commands.
Eliad.
--
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
