diff mbox

[1/2] af9033: fix signal strength value not correct issue

Message ID 1411980225.1747.10.camel@ite-desktop (mailing list archive)
State Superseded
Headers show

Commit Message

Bimow Chen Sept. 29, 2014, 8:43 a.m. UTC 
Register 0x800048 is not dB measure but relative scale. Fix it and conform to NorDig specifications.

Comments

Antti Palosaari Sept. 29, 2014, 12:02 p.m. UTC | #1 
On 09/29/2014 11:43 AM, Bimow Chen wrote:
> Register 0x800048 is not dB measure but relative scale. Fix it and conform to NorDig specifications.

eh, 0x800048 register returned strength normalized to 0-100 %. But that 
was earlier when older firmwares used. I have seen it does not return 
anything anymore, so I am very fine it is replaced with something 
meaningful.

But the issues is that this patches changes current DVBv5 signal 
reporting from dBm to relative. I indeed implemented it is as a dBm and 
I checked it using modulator RF strength it really is dBm. Now you add 
some glue which converts dBm to relative value between 0-0xffff.

I encourage you to use modulator yourself to generate signals. Then use 
dvbv5-zap to see values DVBv5 API reports.

If you really want return 0-0xffff values, then do it for old DVBv3 
read_signal_strength(), but do not change new DVBv5 statistics to 
relative. dBm, as a clearly defined unit, is always preferred over 
relative. Relative was added to API for cases we cannot report well 
known units.

Could you tell which is unit NorDig specification defines for signal 
strength?

regards
Antti



>
>
> 0001-af9033-fix-signal-strength-value-not-correct-issue.patch
>
>
>>From b85ad9df69884b80cce62877039aa9130243ef3a Mon Sep 17 00:00:00 2001
> From: Bimow Chen<Bimow.Chen@ite.com.tw>
> Date: Mon, 29 Sep 2014 13:57:07 +0800
> Subject: [PATCH 1/2] af9033: fix signal strength value not correct issue
>
> Register 0x800048 is not dB measure but relative scale. Fix it and conform to NorDig specifications.
>
> Signed-off-by: Bimow Chen<Bimow.Chen@ite.com.tw>
> ---
>   drivers/media/dvb-frontends/af9033.c      |   54 ++++++++++++++++++-----------
>   drivers/media/dvb-frontends/af9033_priv.h |    6 +++
>   2 files changed, 40 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/media/dvb-frontends/af9033.c b/drivers/media/dvb-frontends/af9033.c
> index 63a89c1..e191bd5 100644
> --- a/drivers/media/dvb-frontends/af9033.c
> +++ b/drivers/media/dvb-frontends/af9033.c
> @@ -862,23 +862,14 @@ static int af9033_read_snr(struct dvb_frontend *fe, u16 *snr)
>   static int af9033_read_signal_strength(struct dvb_frontend *fe, u16 *strength)
>   {
>   	struct af9033_dev *dev = fe->demodulator_priv;
> -	int ret;
> -	u8 strength2;
> -
> -	/* read signal strength of 0-100 scale */
> -	ret = af9033_rd_reg(dev, 0x800048, &strength2);
> -	if (ret < 0)
> -		goto err;
> +	struct dtv_frontend_properties *c = &dev->fe.dtv_property_cache;
>
> -	/* scale value to 0x0000-0xffff */
> -	*strength = strength2 * 0xffff / 100;
> +	if (c->strength.stat[0].scale == FE_SCALE_RELATIVE)
> +		*strength = c->strength.stat[0].uvalue;
> +	else
> +		*strength = 0;
>
>   	return 0;
> -
> -err:
> -	dev_dbg(&dev->client->dev, "failed=%d\n", ret);
> -
> -	return ret;
>   }
>
>   static int af9033_read_ber(struct dvb_frontend *fe, u32 *ber)
> @@ -974,8 +965,8 @@ static void af9033_stat_work(struct work_struct *work)
>   {
>   	struct af9033_dev *dev = container_of(work, struct af9033_dev, stat_work.work);
>   	struct dtv_frontend_properties *c = &dev->fe.dtv_property_cache;
> -	int ret, tmp, i, len;
> -	u8 u8tmp, buf[7];
> +	int ret, tmp, i, len, power_real;
> +	u8 u8tmp, gain_offset, buf[7];
>
>   	dev_dbg(&dev->client->dev, "\n");
>
> @@ -983,17 +974,40 @@ static void af9033_stat_work(struct work_struct *work)
>   	if (dev->fe_status & FE_HAS_SIGNAL) {
>   		if (dev->is_af9035) {
>   			ret = af9033_rd_reg(dev, 0x80004a, &u8tmp);
> -			tmp = -u8tmp * 1000;
> +			/* scale value to 0x0000-0xffff */
> +			tmp = u8tmp * 0xffff / 100;
>   		} else {
>   			ret = af9033_rd_reg(dev, 0x8000f7, &u8tmp);
> -			tmp = (u8tmp - 100) * 1000;
> +			ret |= af9033_rd_regs(dev, 0x80f900, buf, 7);
> +
> +			if (c->frequency <= 300000000)
> +				gain_offset = 7; /* VHF */
> +			else
> +				gain_offset = 4; /* UHF */
> +
> +			power_real = (u8tmp - 100 - gain_offset) -
> +				power_reference[((buf[3] >> 0) & 3)][((buf[6] >> 0) & 7)];
> +
> +			if (power_real < -15)
> +				tmp = 0;
> +			else if ((power_real >= -15) && (power_real < 0))
> +				tmp = (2 * (power_real + 15)) / 3;
> +			else if ((power_real >= 0) && (power_real < 20))
> +				tmp = 4 * power_real + 10;
> +			else if ((power_real >= 20) && (power_real < 35))
> +				tmp = (2 * (power_real - 20)) / 3 + 90;
> +			else
> +				tmp = 100;
> +
> +			/* scale value to 0x0000-0xffff */
> +			tmp = tmp * 0xffff / 100;
>   		}
>   		if (ret)
>   			goto err;
>
>   		c->strength.len = 1;
> -		c->strength.stat[0].scale = FE_SCALE_DECIBEL;
> -		c->strength.stat[0].svalue = tmp;
> +		c->strength.stat[0].scale = FE_SCALE_RELATIVE;
> +		c->strength.stat[0].uvalue = tmp;
>   	} else {
>   		c->strength.len = 1;
>   		c->strength.stat[0].scale = FE_SCALE_NOT_AVAILABLE;
> diff --git a/drivers/media/dvb-frontends/af9033_priv.h b/drivers/media/dvb-frontends/af9033_priv.h
> index c12c92c..c9c8798 100644
> --- a/drivers/media/dvb-frontends/af9033_priv.h
> +++ b/drivers/media/dvb-frontends/af9033_priv.h
> @@ -2051,4 +2051,10 @@ static const struct reg_val tuner_init_it9135_62[] = {
>   	{ 0x80fd8b, 0x00 },
>   };
>
> +/* NorDig power reference table */
> +static const int power_reference[][5] = {
> +	{-93, -91, -90, -89, -88}, /* QPSK 1/2 ~ 7/8 */
> +	{-87, -85, -84, -83, -82}, /* 16QAM 1/2 ~ 7/8 */
> +	{-82, -80, -78, -77, -76}, /* 64QAM 1/2 ~ 7/8 */
> +};
>   #endif /* AF9033_PRIV_H */
> -- 1.7.0.4
>
Antti Palosaari Sept. 29, 2014, 2:12 p.m. UTC | #2 
On 09/29/2014 03:02 PM, Antti Palosaari wrote:
> On 09/29/2014 11:43 AM, Bimow Chen wrote:
>> Register 0x800048 is not dB measure but relative scale. Fix it and
>> conform to NorDig specifications.
>
> eh, 0x800048 register returned strength normalized to 0-100 %. But that
> was earlier when older firmwares used. I have seen it does not return
> anything anymore, so I am very fine it is replaced with something
> meaningful.
>
> But the issues is that this patches changes current DVBv5 signal
> reporting from dBm to relative. I indeed implemented it is as a dBm and
> I checked it using modulator RF strength it really is dBm. Now you add
> some glue which converts dBm to relative value between 0-0xffff.
>
> I encourage you to use modulator yourself to generate signals. Then use
> dvbv5-zap to see values DVBv5 API reports.
>
> If you really want return 0-0xffff values, then do it for old DVBv3
> read_signal_strength(), but do not change new DVBv5 statistics to
> relative. dBm, as a clearly defined unit, is always preferred over
> relative. Relative was added to API for cases we cannot report well
> known units.
>
> Could you tell which is unit NorDig specification defines for signal
> strength?

According to latest NorDig specification, page 39
http://www.nordig.org/pdf/NorDig-Unified_ver_2.5.1.pdf
there is two kind of reports, basic and advanced. Normal report is (%) 
which suits well for relative scale.

The basic status check shall include:
* channel id, according to Annex B.2
* centre frequency
* Signal Strength Indicator, SSI (%), according to section 3.4.4.6
* Signal Quality Indicator, SQI (%), according to section 3.4.4.7


The terrestrial NorDig IRD should provide an advanced status check 
function (accessible through the Navigator) that presents the following 
information:
* channel id, according to Annex B.2
* centre frequency
* signal strength (dBm or dB?V)
* signal strength indicator, SSI (%), according to section 3.4.4.6
* signal quality indicator, SQI (%), according to section 3.4.4.7
* C/N (dB)
* BER before Reed Solomon decoding (DVB-T) or BCH decoding (DVB-T2)
* Uncorrected packets


regards
Antti
diff mbox

Patch

From b85ad9df69884b80cce62877039aa9130243ef3a Mon Sep 17 00:00:00 2001
From: Bimow Chen <Bimow.Chen@ite.com.tw>
Date: Mon, 29 Sep 2014 13:57:07 +0800
Subject: [PATCH 1/2] af9033: fix signal strength value not correct issue

Register 0x800048 is not dB measure but relative scale. Fix it and conform to NorDig specifications.

Signed-off-by: Bimow Chen <Bimow.Chen@ite.com.tw>
---
 drivers/media/dvb-frontends/af9033.c      |   54 ++++++++++++++++++-----------
 drivers/media/dvb-frontends/af9033_priv.h |    6 +++
 2 files changed, 40 insertions(+), 20 deletions(-)

diff --git a/drivers/media/dvb-frontends/af9033.c b/drivers/media/dvb-frontends/af9033.c
index 63a89c1..e191bd5 100644
--- a/drivers/media/dvb-frontends/af9033.c
+++ b/drivers/media/dvb-frontends/af9033.c
@@ -862,23 +862,14 @@  static int af9033_read_snr(struct dvb_frontend *fe, u16 *snr)
 static int af9033_read_signal_strength(struct dvb_frontend *fe, u16 *strength)
 {
 	struct af9033_dev *dev = fe->demodulator_priv;
-	int ret;
-	u8 strength2;
-
-	/* read signal strength of 0-100 scale */
-	ret = af9033_rd_reg(dev, 0x800048, &strength2);
-	if (ret < 0)
-		goto err;
+	struct dtv_frontend_properties *c = &dev->fe.dtv_property_cache;
 
-	/* scale value to 0x0000-0xffff */
-	*strength = strength2 * 0xffff / 100;
+	if (c->strength.stat[0].scale == FE_SCALE_RELATIVE)
+		*strength = c->strength.stat[0].uvalue;
+	else
+		*strength = 0;
 
 	return 0;
-
-err:
-	dev_dbg(&dev->client->dev, "failed=%d\n", ret);
-
-	return ret;
 }
 
 static int af9033_read_ber(struct dvb_frontend *fe, u32 *ber)
@@ -974,8 +965,8 @@  static void af9033_stat_work(struct work_struct *work)
 {
 	struct af9033_dev *dev = container_of(work, struct af9033_dev, stat_work.work);
 	struct dtv_frontend_properties *c = &dev->fe.dtv_property_cache;
-	int ret, tmp, i, len;
-	u8 u8tmp, buf[7];
+	int ret, tmp, i, len, power_real;
+	u8 u8tmp, gain_offset, buf[7];
 
 	dev_dbg(&dev->client->dev, "\n");
 
@@ -983,17 +974,40 @@  static void af9033_stat_work(struct work_struct *work)
 	if (dev->fe_status & FE_HAS_SIGNAL) {
 		if (dev->is_af9035) {
 			ret = af9033_rd_reg(dev, 0x80004a, &u8tmp);
-			tmp = -u8tmp * 1000;
+			/* scale value to 0x0000-0xffff */
+			tmp = u8tmp * 0xffff / 100;
 		} else {
 			ret = af9033_rd_reg(dev, 0x8000f7, &u8tmp);
-			tmp = (u8tmp - 100) * 1000;
+			ret |= af9033_rd_regs(dev, 0x80f900, buf, 7);
+
+			if (c->frequency <= 300000000)
+				gain_offset = 7; /* VHF */
+			else
+				gain_offset = 4; /* UHF */
+
+			power_real = (u8tmp - 100 - gain_offset) -
+				power_reference[((buf[3] >> 0) & 3)][((buf[6] >> 0) & 7)];
+
+			if (power_real < -15)
+				tmp = 0;
+			else if ((power_real >= -15) && (power_real < 0))
+				tmp = (2 * (power_real + 15)) / 3;
+			else if ((power_real >= 0) && (power_real < 20))
+				tmp = 4 * power_real + 10;
+			else if ((power_real >= 20) && (power_real < 35))
+				tmp = (2 * (power_real - 20)) / 3 + 90;
+			else
+				tmp = 100;
+
+			/* scale value to 0x0000-0xffff */
+			tmp = tmp * 0xffff / 100;
 		}
 		if (ret)
 			goto err;
 
 		c->strength.len = 1;
-		c->strength.stat[0].scale = FE_SCALE_DECIBEL;
-		c->strength.stat[0].svalue = tmp;
+		c->strength.stat[0].scale = FE_SCALE_RELATIVE;
+		c->strength.stat[0].uvalue = tmp;
 	} else {
 		c->strength.len = 1;
 		c->strength.stat[0].scale = FE_SCALE_NOT_AVAILABLE;
diff --git a/drivers/media/dvb-frontends/af9033_priv.h b/drivers/media/dvb-frontends/af9033_priv.h
index c12c92c..c9c8798 100644
--- a/drivers/media/dvb-frontends/af9033_priv.h
+++ b/drivers/media/dvb-frontends/af9033_priv.h
@@ -2051,4 +2051,10 @@  static const struct reg_val tuner_init_it9135_62[] = {
 	{ 0x80fd8b, 0x00 },
 };
 
+/* NorDig power reference table */
+static const int power_reference[][5] = {
+	{-93, -91, -90, -89, -88}, /* QPSK 1/2 ~ 7/8 */
+	{-87, -85, -84, -83, -82}, /* 16QAM 1/2 ~ 7/8 */
+	{-82, -80, -78, -77, -76}, /* 64QAM 1/2 ~ 7/8 */
+};
 #endif /* AF9033_PRIV_H */
-- 
1.7.0.4