[RFC][PATCH] rtl2832: PID filter support for slave demod

[RFC][PATCH] rtl2832: PID filter support for slave demod

[Benjamin Larsson]

[-- Attachment #1: Type: text/plain, Size: 261 bytes --]

Is this structure ok for the slave pid implementation? Or should there 
be only one filters parameter? Will the overlaying pid filter framework 
properly "flush" the set pid filters ?

Note that this code currently is only compile tested.

MvH
Benjamin Larsson

[-- Attachment #2: 0001-rtl2832-PID-filter-support-for-slave-demod.patch --]
[-- Type: text/x-patch, Size: 3834 bytes --]

>From 8efb26c18b4f9416bd516195c6a82853c9cccc24 Mon Sep 17 00:00:00 2001
From: Benjamin Larsson <benjamin@southpole.se>
Date: Mon, 16 Mar 2015 22:59:50 +0100
Subject: [PATCH] rtl2832: PID filter support for slave demod
Cc: Linux Media Mailing List <linux-media@vger.kernel.org>

RTL2832p supports a slave configuration with a demodulator connected
to a ts input on the chip. This makes it possible to receive DVB-T2
muxes that are of larger size then what the rtl2832p usb-bridge is
capable of transfering.

Signed-off-by: Benjamin Larsson <benjamin@southpole.se>
---
 drivers/media/dvb-frontends/rtl2832.c      | 40 ++++++++++++++++++++++--------
 drivers/media/dvb-frontends/rtl2832_priv.h |  2 ++
 2 files changed, 32 insertions(+), 10 deletions(-)

diff --git a/drivers/media/dvb-frontends/rtl2832.c b/drivers/media/dvb-frontends/rtl2832.c
index 5d2d8f4..3725211 100644
--- a/drivers/media/dvb-frontends/rtl2832.c
+++ b/drivers/media/dvb-frontends/rtl2832.c
@@ -488,6 +488,8 @@ static int rtl2832_set_frontend(struct dvb_frontend *fe)
 	if (ret)
 		goto err;
 
+	dev->slave_demod_active = 0;
+
 	/* If the frontend has get_if_frequency(), use it */
 	if (fe->ops.tuner_ops.get_if_frequency) {
 		u32 if_freq;
@@ -1114,6 +1116,8 @@ static int rtl2832_enable_slave_ts(struct i2c_client *client)
 	if (ret)
 		goto err;
 
+	dev->slave_demod_active = 1;
+
 	return 0;
 err:
 	dev_dbg(&client->dev, "failed=%d\n", ret);
@@ -1135,7 +1139,10 @@ static int rtl2832_pid_filter_ctrl(struct dvb_frontend *fe, int onoff)
 	else
 		u8tmp = 0x00;
 
-	ret = rtl2832_update_bits(client, 0x061, 0xc0, u8tmp);
+	if (dev->slave_demod_active)
+		ret = rtl2832_update_bits(client, 0x021, 0xc0, u8tmp);
+	else
+		ret = rtl2832_update_bits(client, 0x061, 0xc0, u8tmp);
 	if (ret)
 		goto err;
 
@@ -1152,6 +1159,7 @@ static int rtl2832_pid_filter(struct dvb_frontend *fe, u8 index, u16 pid,
 	struct i2c_client *client = dev->client;
 	int ret;
 	u8 buf[4];
+	unsigned long* filters;
 
 	dev_dbg(&client->dev, "index=%d pid=%04x onoff=%d\n",
 		index, pid, onoff);
@@ -1160,24 +1168,36 @@ static int rtl2832_pid_filter(struct dvb_frontend *fe, u8 index, u16 pid,
 	if (pid > 0x1fff || index > 32)
 		return 0;
 
-	if (onoff)
-		set_bit(index, &dev->filters);
+	if (dev->slave_demod_active)
+		filters = &dev->filters_slave;
 	else
-		clear_bit(index, &dev->filters);
+		filters = &dev->filters;
+
+	if (onoff) {
+		set_bit(index, filters);
+	} else {
+		clear_bit(index, filters);
+	}
 
 	/* enable / disable PIDs */
-	buf[0] = (dev->filters >>  0) & 0xff;
-	buf[1] = (dev->filters >>  8) & 0xff;
-	buf[2] = (dev->filters >> 16) & 0xff;
-	buf[3] = (dev->filters >> 24) & 0xff;
-	ret = rtl2832_bulk_write(client, 0x062, buf, 4);
+	buf[0] = (*filters >>  0) & 0xff;
+	buf[1] = (*filters >>  8) & 0xff;
+	buf[2] = (*filters >> 16) & 0xff;
+	buf[3] = (*filters >> 24) & 0xff;
+	if (dev->slave_demod_active)
+		ret = rtl2832_bulk_write(client, 0x022, buf, 4);
+	else
+		ret = rtl2832_bulk_write(client, 0x062, buf, 4);
 	if (ret)
 		goto err;
 
 	/* add PID */
 	buf[0] = (pid >> 8) & 0xff;
 	buf[1] = (pid >> 0) & 0xff;
-	ret = rtl2832_bulk_write(client, 0x066 + 2 * index, buf, 2);
+	if (dev->slave_demod_active)
+		ret = rtl2832_bulk_write(client, 0x026 + 2 * index, buf, 2);
+	else
+		ret = rtl2832_bulk_write(client, 0x066 + 2 * index, buf, 2);
 	if (ret)
 		goto err;
 
diff --git a/drivers/media/dvb-frontends/rtl2832_priv.h b/drivers/media/dvb-frontends/rtl2832_priv.h
index c3a922c..b95a7b7 100644
--- a/drivers/media/dvb-frontends/rtl2832_priv.h
+++ b/drivers/media/dvb-frontends/rtl2832_priv.h
@@ -46,6 +46,8 @@ struct rtl2832_dev {
 	bool sleeping;
 	struct delayed_work i2c_gate_work;
 	unsigned long filters; /* PID filter */
+	unsigned long filters_slave; /* PID filter */
+	int slave_demod_active;
 };
 
 struct rtl2832_reg_entry {
-- 
2.1.0

Re: [RFC][PATCH] rtl2832: PID filter support for slave demod

[Antti Palosaari]

On 03/17/2015 12:12 AM, Benjamin Larsson wrote:
> Is this structure ok for the slave pid implementation? Or should there
> be only one filters parameter? Will the overlaying pid filter framework
> properly "flush" the set pid filters ?
>
> Note that this code currently is only compile tested.

I am fine with it.

byw. Have you tested if your QAM256 (DVB-C or DVB-T2) stream is valid 
even without a PID filtering? IIRC mine stream is correct and PID 
filtering is not required (but surely it could be implemented if you wish).

regards
Antti

-- 
http://palosaari.fi/

Re: [RFC][PATCH] rtl2832: PID filter support for slave demod

[Benjamin Larsson]

On 03/16/2015 11:57 PM, Antti Palosaari wrote:
> On 03/17/2015 12:12 AM, Benjamin Larsson wrote:
>> Is this structure ok for the slave pid implementation? Or should there
>> be only one filters parameter? Will the overlaying pid filter framework
>> properly "flush" the set pid filters ?
>>
>> Note that this code currently is only compile tested.
>
> I am fine with it.
>
> byw. Have you tested if your QAM256 (DVB-C or DVB-T2) stream is valid
> even without a PID filtering? IIRC mine stream is correct and PID
> filtering is not required (but surely it could be implemented if you wish).
>
> regards
> Antti
>

DVB-C seems fine and one of my DVB-T2 muxes is fine also. But one other 
DVB-T2 mux completely fails. It could be the reception but it might be 
that it needs pid filtering. I do get small disturbances on my DVB-C 
muxes. Will report back if pid filtering makes things better or not.

MvH
Benjamin Larsson

Re: [RFC][PATCH] rtl2832: PID filter support for slave demod

[Mauro Carvalho Chehab]

Em Tue, 17 Mar 2015 00:16:06 +0100
Benjamin Larsson <benjamin@southpole.se> escreveu:

> On 03/16/2015 11:57 PM, Antti Palosaari wrote:
> > On 03/17/2015 12:12 AM, Benjamin Larsson wrote:
> >> Is this structure ok for the slave pid implementation? Or should there
> >> be only one filters parameter? Will the overlaying pid filter framework
> >> properly "flush" the set pid filters ?
> >>
> >> Note that this code currently is only compile tested.
> >
> > I am fine with it.
> >
> > byw. Have you tested if your QAM256 (DVB-C or DVB-T2) stream is valid
> > even without a PID filtering? IIRC mine stream is correct and PID
> > filtering is not required (but surely it could be implemented if you wish).
> >
> > regards
> > Antti
> >
> 
> DVB-C seems fine and one of my DVB-T2 muxes is fine also. But one other 
> DVB-T2 mux completely fails. It could be the reception but it might be 
> that it needs pid filtering. I do get small disturbances on my DVB-C 
> muxes. Will report back if pid filtering makes things better or not.

What's the status of this patch?

Btw, checkpatch.pl complains about a few things there:

WARNING: 'transfering' may be misspelled - perhaps 'transferring'?
#31: capable of transfering.

ERROR: "foo* bar" should be "foo *bar"
#77: FILE: drivers/media/dvb-frontends/rtl2832.c:1162:
+	unsigned long* filters;

WARNING: braces {} are not necessary for any arm of this statement
#93: FILE: drivers/media/dvb-frontends/rtl2832.c:1176:
+	if (onoff) {
[...]
+	} else {
[...]

total: 1 errors, 2 warnings, 87 lines checked

For now, as it was sent as RFC, I'll tag as such at patchwork.

If this is patch is pertinent, please re-send it (with the coding style
issues pointed by checkpatch.pl fixed).

Regards,
Mauro

> 
> MvH
> Benjamin Larsson
> --
> To unsubscribe from this list: send the line "unsubscribe linux-media" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html