[Linux-kernel-mentees] [PATCH v2 1/2] cec-follower: add tuner step increment/decrement

On 9/25/19 8:54 AM, Hans Verkuil wrote:

Hi Jiunn,

On 9/24/19 9:24 PM, Jiunn Chang wrote:

quoted

Tuner step increment/decrement will select the next highest or next
lowest service frequency.  There are a total of three possible
frequencies from analog_freqs_khz given a broadcast type and system.

Opcodes implemented:
  - <Tuner Step Increment>
  - <Tuner Step Decrement>

Signed-off-by: Jiunn Chang <c0d1n61at3 at gmail.com>
---
 utils/cec-follower/cec-tuner.cpp | 47 ++++++++++++++++++++++++++++++--
 1 file changed, 44 insertions(+), 3 deletions(-)

diff --git a/utils/cec-follower/cec-tuner.cpp b/utils/cec-follower/cec-tuner.cpp
index acc3fd00..1a9b9d90 100644
--- a/utils/cec-follower/cec-tuner.cpp
+++ b/utils/cec-follower/cec-tuner.cpp

@@ -135,6 +135,22 @@ static bool analog_set_tuner_dev_info(struct node *node, struct cec_msg *msg)
 	return false;
 }
 
+static int analog_find_freq_index(struct cec_op_tuner_device_info *info)
+{
+	if (info->analog.ana_freq == 0)
+		return -1;
+
+	int ana_freq_khz = (info->analog.ana_freq * 625) / 10;
+
+	for (int i = 0; i < NUM_ANALOG_FREQS; i++) {
+		int freq = analog_freqs_khz[info->analog.ana_bcast_type][info->analog.bcast_system][i];
+
+		if (ana_freq_khz == freq)
+			return i;
+	}
+	return -1;
+}

Hmm, I don't like this code. I think that it would be cleaner to just add:

static unsigned int analog_freq_idx[3][9];

Which contains the index of the current frequency for each bcast_type/system
combination.

This ensures that 1) there always is a current frequency and 2) tuner stepping
is now trivial since all you need to do is to update the corresponding index.

I've been reading up on this a bit more, and section CEC 13.8.2 states that the
Tuner Step commands cycle through a set of services defined by the tuner device.
But normally it will cycle through all services.

That means that rather than cycling through the frequencies of just the current
bcast_type/system, the follower should cycle through ALL channels, both analog
and digital.

So for analog we really have 3 * 9 * 3 channels, and this will be more once the
digital channels are added.

In any case, this means that there is just a single 'static unsigned int freq_idx;'
and not an array.

Regards,

	Hans

Now, that said, I was thinking that it would be nice to have one bcast_type/system
combo with no channels at all, since that's a nice test as well.

I think that setting secam-lq for satellite to { 0, 0, 0 } to indicate no
defined channels would make a nice test case.

Something for a follow-up patch, though.

quoted

+
 void process_tuner_record_timer_msgs(struct node *node, struct cec_msg &msg, unsigned me)
 {
 	bool is_bcast = cec_msg_is_broadcast(&msg);

@@ -178,12 +194,37 @@ void process_tuner_record_timer_msgs(struct node *node, struct cec_msg &msg, uns
 		return;
 
 	case CEC_MSG_SELECT_DIGITAL_SERVICE:
-	case CEC_MSG_TUNER_STEP_DECREMENT:
-	case CEC_MSG_TUNER_STEP_INCREMENT:
-		if (!cec_has_tuner(1 << me))
+	case CEC_MSG_TUNER_STEP_DECREMENT: {
+		if (!cec_has_tuner(1 << me) && !cec_has_tv(1 << me))
 			break;
+
+		struct cec_op_tuner_device_info *info = &node->state.tuner_dev_info;
+		int freq_idx = analog_find_freq_index(info);
+		int idx = (freq_idx == 0) ? NUM_ANALOG_FREQS : freq_idx;
+
+		if (info->is_analog) {
+			int freq = analog_freqs_khz[info->analog.ana_bcast_type][info->analog.bcast_system][--idx];
+
+			info->analog.ana_freq = (freq * 10) / 625;
+		}
 		return;
+	}
+
+	case CEC_MSG_TUNER_STEP_INCREMENT: {
+		if (!cec_has_tuner(1 << me) && !cec_has_tv(1 << me))
+			break;
 
+		struct cec_op_tuner_device_info *info = &node->state.tuner_dev_info;
+		int freq_idx = analog_find_freq_index(info);
+		int idx = (freq_idx == NUM_ANALOG_FREQS - 1) ? -1 : freq_idx;
+
+		if (info->is_analog) {
+			int freq = analog_freqs_khz[info->analog.ana_bcast_type][info->analog.bcast_system][++idx];
+
+			info->analog.ana_freq = (freq * 10) / 625;
+		}
+		return;
+	}
 
 		/*
 		  One Touch Record

Regards,

	Hans