[PATCH v1] mgmtops: Handle Start Discovery Complete in case of failure

[PATCH v1] mgmtops: Handle Start Discovery Complete in case of failure

[Hemant Gupta]

This patch adds handling of start discovery complete event in cases
when discovery fails to get started.
---
 plugins/mgmtops.c |   33 ++++++++++++++++++++++++++++++++-
 1 files changed, 32 insertions(+), 1 deletions(-)

diff --git a/plugins/mgmtops.c b/plugins/mgmtops.c
index c08117f..9de490d 100644
--- a/plugins/mgmtops.c
+++ b/plugins/mgmtops.c
@@ -1255,6 +1255,37 @@ static void read_local_oob_data_complete(int sk, uint16_t index, void *buf,
 		oob_read_local_data_complete(adapter, rp->hash, rp->randomizer);
 }
 
+static void start_discovery_complete(int sk, uint16_t index, uint8_t status,
+				     void *buf, size_t len)
+{
+	uint8_t *type = buf;
+	struct btd_adapter *adapter;
+
+	if (len != sizeof(*type)) {
+		error("start_discovery_complete event size mismatch "
+			"(%zu != %zu)", len, sizeof(*type));
+		return;
+	}
+
+	DBG("type %u status %u", *type, status);
+
+	if (index > max_index) {
+		error("Unexpected index %u in start_discovery_complete",
+			index);
+		return;
+	}
+
+	if (!status)
+		return;
+
+	DBG("hci%u", index);
+
+	adapter = manager_find_adapter_by_id(index);
+	if (adapter)
+		/* Start discovery failed, inform upper layers. */
+		adapter_set_discovering(adapter, FALSE);
+}
+
 static void read_local_oob_data_failed(int sk, uint16_t index)
 {
 	struct btd_adapter *adapter;
@@ -1421,7 +1452,7 @@ static void mgmt_cmd_complete(int sk, uint16_t index, void *buf, size_t len)
 		DBG("set_fast_connectable complete");
 		break;
 	case MGMT_OP_START_DISCOVERY:
-		DBG("start_discovery complete");
+		start_discovery_complete(sk, index, ev->status, ev->data, len);
 		break;
 	case MGMT_OP_STOP_DISCOVERY:
 		DBG("stop_discovery complete");
-- 
1.7.0.4

Re: [PATCH v1] mgmtops: Handle Start Discovery Complete in case of failure

[Johan Hedberg]

Hi Hemant,

On Tue, Apr 10, 2012, Hemant Gupta wrote:
> This patch adds handling of start discovery complete event in cases
> when discovery fails to get started.
> ---
>  plugins/mgmtops.c |   33 ++++++++++++++++++++++++++++++++-
>  1 files changed, 32 insertions(+), 1 deletions(-)

The patch is now applied but I still had to fix it up manually myself:

> +static void start_discovery_complete(int sk, uint16_t index, uint8_t status,
> +				     void *buf, size_t len)

Incorrect indentation above: you should indent with tabs as much as
possible while remaining under the 80 character boundary.

> +	if (len != sizeof(*type)) {
> +		error("start_discovery_complete event size mismatch "
> +			"(%zu != %zu)", len, sizeof(*type));

This should also be indented more.

> +	DBG("type %u status %u", *type, status);

Where did the hci%u go? I meant move the entire DBG statement here and
not just part of it (it's perfectly fine to print the index value even
though it's > max_index).

> +	if (index > max_index) {
> +		error("Unexpected index %u in start_discovery_complete",
> +			index);

Incorrect indentation again (should be indented more)

> +	DBG("hci%u", index);

And the above should go away assuming that you move the entire statement
to the new location.

I think we'll need to reexamine the user space coding style now that the
kernel doesn't use a strict tabs-only approach anymore. But for now
we're still following the style we've always used in user space.

Johan

Re: [PATCH v1] mgmtops: Handle Start Discovery Complete in case of failure

[Hemant Gupta]

Hi Johan,

On Tue, Apr 10, 2012 at 4:03 PM, Johan Hedberg <johan.hedberg@gmail.com> wrote:
> Hi Hemant,
>
> On Tue, Apr 10, 2012, Hemant Gupta wrote:
>> This patch adds handling of start discovery complete event in cases
>> when discovery fails to get started.
>> ---
>>  plugins/mgmtops.c |   33 ++++++++++++++++++++++++++++++++-
>>  1 files changed, 32 insertions(+), 1 deletions(-)
>
> The patch is now applied but I still had to fix it up manually myself:
>
>> +static void start_discovery_complete(int sk, uint16_t index, uint8_t status,
>> +                                  void *buf, size_t len)
>
> Incorrect indentation above: you should indent with tabs as much as
> possible while remaining under the 80 character boundary.
>
>> +     if (len != sizeof(*type)) {
>> +             error("start_discovery_complete event size mismatch "
>> +                     "(%zu != %zu)", len, sizeof(*type));
>
> This should also be indented more.
>
>> +     DBG("type %u status %u", *type, status);
>
> Where did the hci%u go? I meant move the entire DBG statement here and
> not just part of it (it's perfectly fine to print the index value even
> though it's > max_index).
>
>> +     if (index > max_index) {
>> +             error("Unexpected index %u in start_discovery_complete",
>> +                     index);
>
> Incorrect indentation again (should be indented more)
>
>> +     DBG("hci%u", index);
>
> And the above should go away assuming that you move the entire statement
> to the new location.
>
Thanks for the comments, I will upload a new patch soon.

> I think we'll need to reexamine the user space coding style now that the
> kernel doesn't use a strict tabs-only approach anymore. But for now
> we're still following the style we've always used in user space.
>
> Johan
> --
> To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html



-- 
Best Regards
Hemant Gupta
ST-Ericsson India

Re: [PATCH v1] mgmtops: Handle Start Discovery Complete in case of failure

[Johan Hedberg]

Hi Hemant,

On Tue, Apr 10, 2012, Hemant Gupta wrote:
> On Tue, Apr 10, 2012 at 4:03 PM, Johan Hedberg <johan.hedberg@gmail.com> wrote:
> > Hi Hemant,
> >
> > On Tue, Apr 10, 2012, Hemant Gupta wrote:
> >> This patch adds handling of start discovery complete event in cases
> >> when discovery fails to get started.
> >> ---
> >>  plugins/mgmtops.c |   33 ++++++++++++++++++++++++++++++++-
> >>  1 files changed, 32 insertions(+), 1 deletions(-)
> >
> > The patch is now applied but I still had to fix it up manually myself:
<snip>
> Thanks for the comments, I will upload a new patch soon.

I think you missed my comment above that I already fixed this myself and
applied the patch :)

Johan