[PATCH] AVerTV MCE 116 Plus radio

[PATCH] AVerTV MCE 116 Plus radio

[Aleksandr V. Piskunov]

Added FM radio support to Avermedia AVerTV MCE 116 Plus card

Signed-off-by: Aleksandr V. Piskunov <alexandr.v.piskunov@gmail.com>

diff --git a/linux/drivers/media/video/ivtv/ivtv-cards.c b/linux/drivers/media/video/ivtv/ivtv-cards.c
--- a/linux/drivers/media/video/ivtv/ivtv-cards.c
+++ b/linux/drivers/media/video/ivtv/ivtv-cards.c
@@ -965,6 +965,7 @@
                { IVTV_CARD_INPUT_AUD_TUNER,  CX25840_AUDIO5       },
                { IVTV_CARD_INPUT_LINE_IN1,   CX25840_AUDIO_SERIAL, 1 },
        },
+       .radio_input = { IVTV_CARD_INPUT_AUD_TUNER,  CX25840_AUDIO5 },
        /* enable line-in */
        .gpio_init = { .direction = 0xe000, .initial_value = 0x4000 },
        .xceive_pin = 10,

Re: [PATCH] AVerTV MCE 116 Plus radio

[Aleksandr V. Piskunov]

On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> 

What leaves me puzzled, radio only works ok with ivtv newi2c=1

With default newi2c audio is tinny, metallic, with some strange static.
Similar problem with pvr-150 was reported years ago, guess issue is still
unresolved, perhaps something with cx25840..

Re: [PATCH] AVerTV MCE 116 Plus radio

[Aleksandr V. Piskunov]

On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
> On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> > Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> > 
> 
> What leaves me puzzled, radio only works ok with ivtv newi2c=1
> 
> With default newi2c audio is tinny, metallic, with some strange static.
> Similar problem with pvr-150 was reported years ago, guess issue is still
> unresolved, perhaps something with cx25840..

This particular "tinny" audio problem is definitely I2C speed related, to be
more precise, audio only goes bad if i2c-algo-bit is being run with udelay
less than 15, i.e. i2c bus frequency is higher than 30 KHz.

So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
reciever on AVerTV M116.

I2C reads/writes to cx25840 themself are ok, verified using register readback
after each write/write4. Problem seems to be that with cx25840 register writes
coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
audio detection routine in inconsistent state.

The only solution I found is to do standard routine (assert_reset + write +
deassert_reset) followed by 50ms delay and another reset.

Following patch works_for_me, can be improved to only delay/doublereset when
really needed, etc. Andy, could you comment/review?

diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
--- a/linux/drivers/media/video/cx25840/cx25840-core.c
+++ b/linux/drivers/media/video/cx25840/cx25840-core.c
@@ -626,7 +642,13 @@
 	if (state->radio) {
 		cx25840_write(client, 0x808, 0xf9);
 		cx25840_write(client, 0x80b, 0x00);
-	}
+		/* Double reset cx2384x after setting FM radio mode, helps to
+		   avoid "tinny" audio when ivtv I2C bus is being run on
+		   frequency higher than 30 KHz */
+		cx25840_and_or(client, 0x810, ~0x01, 0);
+		msleep(50);
+		cx25840_and_or(client, 0x810, ~0x01, 1);
+	}	
 	else if (std & V4L2_STD_525_60) {
 		/* Certain Hauppauge PVR150 models have a hardware bug
 		   that causes audio to drop out. For these models the

Re: [PATCH] AVerTV MCE 116 Plus radio

[Andy Walls]

On Sun, 2009-10-11 at 04:01 +0300, Aleksandr V. Piskunov wrote:
> On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
> > On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> > > Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> > > 
> > 
> > What leaves me puzzled, radio only works ok with ivtv newi2c=1
> > 
> > With default newi2c audio is tinny, metallic, with some strange static.
> > Similar problem with pvr-150 was reported years ago, guess issue is still
> > unresolved, perhaps something with cx25840..
> 
> This particular "tinny" audio problem is definitely I2C speed related, to be
> more precise, audio only goes bad if i2c-algo-bit is being run with udelay
> less than 15, i.e. i2c bus frequency is higher than 30 KHz.
> 
> So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
> radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
> reciever on AVerTV M116.
> 
> I2C reads/writes to cx25840 themself are ok, verified using register readback
> after each write/write4. Problem seems to be that with cx25840 register writes
> coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
> TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
> audio detection routine in inconsistent state.
> 
> The only solution I found is to do standard routine (assert_reset + write +
> deassert_reset) followed by 50ms delay and another reset.
> 
> Following patch works_for_me, can be improved to only delay/doublereset when
> really needed, etc. Andy, could you comment/review?

Aleksandr,

I will when I get time.  This past week and next few weeks are very busy
for me for personal (non-linux) reasons.  I'll try to get caught up with
the patches I still have to rework and then look at this.

Obviously, your patch is fairly straightforward and looks OK.  I just
haven't checked for any implications.  The "general" tinny audio problem
with the CX25840 on ivtv boards is *always* resolved with an audio
microcontroller reset.

The problem is the microcontroller may restart its detection loop and
tinny audio may return.  Can you run FM radio for a long time (a day ?),
and see if it ever goes back to tinny audio?

Regards,
Andy

> 
> diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
> --- a/linux/drivers/media/video/cx25840/cx25840-core.c
> +++ b/linux/drivers/media/video/cx25840/cx25840-core.c
> @@ -626,7 +642,13 @@
>  	if (state->radio) {
>  		cx25840_write(client, 0x808, 0xf9);
>  		cx25840_write(client, 0x80b, 0x00);
> -	}
> +		/* Double reset cx2384x after setting FM radio mode, helps to
> +		   avoid "tinny" audio when ivtv I2C bus is being run on
> +		   frequency higher than 30 KHz */
> +		cx25840_and_or(client, 0x810, ~0x01, 0);
> +		msleep(50);
> +		cx25840_and_or(client, 0x810, ~0x01, 1);
> +	}	
>  	else if (std & V4L2_STD_525_60) {
>  		/* Certain Hauppauge PVR150 models have a hardware bug
>  		   that causes audio to drop out. For these models the
> 

Re: [PATCH] AVerTV MCE 116 Plus radio

[Andy Walls]

On Sun, 2009-10-11 at 04:01 +0300, Aleksandr V. Piskunov wrote:
> On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
> > On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> > > Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> > > 
> > 
> > What leaves me puzzled, radio only works ok with ivtv newi2c=1
> > 
> > With default newi2c audio is tinny, metallic, with some strange static.
> > Similar problem with pvr-150 was reported years ago, guess issue is still
> > unresolved, perhaps something with cx25840..
> 
> This particular "tinny" audio problem is definitely I2C speed related, to be
> more precise, audio only goes bad if i2c-algo-bit is being run with udelay
> less than 15, i.e. i2c bus frequency is higher than 30 KHz.
> 
> So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
> radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
> reciever on AVerTV M116.
> 
> I2C reads/writes to cx25840 themself are ok, verified using register readback
> after each write/write4. Problem seems to be that with cx25840 register writes
> coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
> TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
> audio detection routine in inconsistent state.
> 
> The only solution I found is to do standard routine (assert_reset + write +
> deassert_reset) followed by 50ms delay and another reset.
> 
> Following patch works_for_me, can be improved to only delay/doublereset when
> really needed, etc. Andy, could you comment/review?

Aleksandr,

Could you provide your Signed-off-by for this patch?  I'm going to
commit it as is.

Thanks,
Andy

> diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
> --- a/linux/drivers/media/video/cx25840/cx25840-core.c
> +++ b/linux/drivers/media/video/cx25840/cx25840-core.c
> @@ -626,7 +642,13 @@
>  	if (state->radio) {
>  		cx25840_write(client, 0x808, 0xf9);
>  		cx25840_write(client, 0x80b, 0x00);
> -	}
> +		/* Double reset cx2384x after setting FM radio mode, helps to
> +		   avoid "tinny" audio when ivtv I2C bus is being run on
> +		   frequency higher than 30 KHz */
> +		cx25840_and_or(client, 0x810, ~0x01, 0);
> +		msleep(50);
> +		cx25840_and_or(client, 0x810, ~0x01, 1);
> +	}	
>  	else if (std & V4L2_STD_525_60) {
>  		/* Certain Hauppauge PVR150 models have a hardware bug
>  		   that causes audio to drop out. For these models the
> 
> --
> 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
> 

Re: [PATCH] AVerTV MCE 116 Plus radio

[Mauro Carvalho Chehab]

Hi Andy,

This patch has never been applied or nacked. From your last comment, it
seems that you're waiting for Aleksandr Signed-of-by:.

If this is still the case, I suggest you to wait for a couple days. If he doesn't
send it, it is safe to add it without his SOB, since it is really a trivial change.

Cheers,
Mauro.

Andy Walls wrote:
> On Sun, 2009-10-11 at 04:01 +0300, Aleksandr V. Piskunov wrote:
>> On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
>>> On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
>>>> Added FM radio support to Avermedia AVerTV MCE 116 Plus card
>>>>
>>> What leaves me puzzled, radio only works ok with ivtv newi2c=1
>>>
>>> With default newi2c audio is tinny, metallic, with some strange static.
>>> Similar problem with pvr-150 was reported years ago, guess issue is still
>>> unresolved, perhaps something with cx25840..
>> This particular "tinny" audio problem is definitely I2C speed related, to be
>> more precise, audio only goes bad if i2c-algo-bit is being run with udelay
>> less than 15, i.e. i2c bus frequency is higher than 30 KHz.
>>
>> So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
>> radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
>> reciever on AVerTV M116.
>>
>> I2C reads/writes to cx25840 themself are ok, verified using register readback
>> after each write/write4. Problem seems to be that with cx25840 register writes
>> coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
>> TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
>> audio detection routine in inconsistent state.
>>
>> The only solution I found is to do standard routine (assert_reset + write +
>> deassert_reset) followed by 50ms delay and another reset.
>>
>> Following patch works_for_me, can be improved to only delay/doublereset when
>> really needed, etc. Andy, could you comment/review?
> 
> Aleksandr,
> 
> Could you provide your Signed-off-by for this patch?  I'm going to
> commit it as is.
> 
> Thanks,
> Andy
> 
>> diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
>> --- a/linux/drivers/media/video/cx25840/cx25840-core.c
>> +++ b/linux/drivers/media/video/cx25840/cx25840-core.c
>> @@ -626,7 +642,13 @@
>>  	if (state->radio) {
>>  		cx25840_write(client, 0x808, 0xf9);
>>  		cx25840_write(client, 0x80b, 0x00);
>> -	}
>> +		/* Double reset cx2384x after setting FM radio mode, helps to
>> +		   avoid "tinny" audio when ivtv I2C bus is being run on
>> +		   frequency higher than 30 KHz */
>> +		cx25840_and_or(client, 0x810, ~0x01, 0);
>> +		msleep(50);
>> +		cx25840_and_or(client, 0x810, ~0x01, 1);
>> +	}	
>>  	else if (std & V4L2_STD_525_60) {
>>  		/* Certain Hauppauge PVR150 models have a hardware bug
>>  		   that causes audio to drop out. For these models the
>>
>> --
>> 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
>>
> 
> --
> 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


-- 

Cheers,
Mauro

Re: [PATCH] AVerTV MCE 116 Plus radio

[Andy Walls]

On Tue, 2010-02-02 at 13:29 -0200, Mauro Carvalho Chehab wrote:
> Hi Andy,
> 
> This patch has never been applied or nacked. From your last comment, it
> seems that you're waiting for Aleksandr Signed-of-by:.
> 
> If this is still the case, I suggest you to wait for a couple days. If he doesn't
> send it, it is safe to add it without his SOB, since it is really a trivial change.

I'd like to look at this one once more.  The extra 50 ms and another
reset may be avoidable.

cx25840-core.c:set_input() gets called for s_frequency so I'd like not
to add 50 ms if not needed.

Regards,
Andy

> Cheers,
> Mauro.
> 
> Andy Walls wrote:
> > On Sun, 2009-10-11 at 04:01 +0300, Aleksandr V. Piskunov wrote:
> >> On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
> >>> On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> >>>> Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> >>>>
> >>> What leaves me puzzled, radio only works ok with ivtv newi2c=1
> >>>
> >>> With default newi2c audio is tinny, metallic, with some strange static.
> >>> Similar problem with pvr-150 was reported years ago, guess issue is still
> >>> unresolved, perhaps something with cx25840..
> >> This particular "tinny" audio problem is definitely I2C speed related, to be
> >> more precise, audio only goes bad if i2c-algo-bit is being run with udelay
> >> less than 15, i.e. i2c bus frequency is higher than 30 KHz.
> >>
> >> So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
> >> radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
> >> reciever on AVerTV M116.
> >>
> >> I2C reads/writes to cx25840 themself are ok, verified using register readback
> >> after each write/write4. Problem seems to be that with cx25840 register writes
> >> coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
> >> TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
> >> audio detection routine in inconsistent state.
> >>
> >> The only solution I found is to do standard routine (assert_reset + write +
> >> deassert_reset) followed by 50ms delay and another reset.
> >>
> >> Following patch works_for_me, can be improved to only delay/doublereset when
> >> really needed, etc. Andy, could you comment/review?
> > 
> > Aleksandr,
> > 
> > Could you provide your Signed-off-by for this patch?  I'm going to
> > commit it as is.
> > 
> > Thanks,
> > Andy
> > 
> >> diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
> >> --- a/linux/drivers/media/video/cx25840/cx25840-core.c
> >> +++ b/linux/drivers/media/video/cx25840/cx25840-core.c
> >> @@ -626,7 +642,13 @@
> >>  	if (state->radio) {
> >>  		cx25840_write(client, 0x808, 0xf9);
> >>  		cx25840_write(client, 0x80b, 0x00);
> >> -	}
> >> +		/* Double reset cx2384x after setting FM radio mode, helps to
> >> +		   avoid "tinny" audio when ivtv I2C bus is being run on
> >> +		   frequency higher than 30 KHz */
> >> +		cx25840_and_or(client, 0x810, ~0x01, 0);
> >> +		msleep(50);
> >> +		cx25840_and_or(client, 0x810, ~0x01, 1);
> >> +	}	
> >>  	else if (std & V4L2_STD_525_60) {
> >>  		/* Certain Hauppauge PVR150 models have a hardware bug
> >>  		   that causes audio to drop out. For these models the
> >>
> >> --
> >> 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
> >>
> > 
> > --
> > 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
> 
> 

Re: [PATCH] AVerTV MCE 116 Plus radio

[Aleksandr V. Piskunov]

On Wed, Feb 03, 2010 at 07:33:20PM -0500, Andy Walls wrote:
> On Tue, 2010-02-02 at 13:29 -0200, Mauro Carvalho Chehab wrote:
> > Hi Andy,
> > 
> > This patch has never been applied or nacked. From your last comment, it
> > seems that you're waiting for Aleksandr Signed-of-by:.
> > 
> > If this is still the case, I suggest you to wait for a couple days. If he doesn't
> > send it, it is safe to add it without his SOB, since it is really a trivial change.
> 
> I'd like to look at this one once more.  The extra 50 ms and another
> reset may be avoidable.
> 
> cx25840-core.c:set_input() gets called for s_frequency so I'd like not
> to add 50 ms if not needed.

This patch is ugly, thats why no SOB was provided by me :)
Delaying every frequency change by 50 ms is no-go.

One possible solution is to mess with audio controller registers only when its really
required: when standard has to be changed, when TV<->Radio switch occurs, etc.

I have a patch that does something like this, will mail when get to PC with that card
and retest it.

As far as I remember there was another codepath that 100% triggers tinny audio, perhaps
it was a change of sampling rate.

> 
> Regards,
> Andy
> 
> > Cheers,
> > Mauro.
> > 
> > Andy Walls wrote:
> > > On Sun, 2009-10-11 at 04:01 +0300, Aleksandr V. Piskunov wrote:
> > >> On Tue, Oct 06, 2009 at 11:11:59AM +0300, Aleksandr V. Piskunov wrote:
> > >>> On Tue, Oct 06, 2009 at 11:04:06AM +0300, Aleksandr V. Piskunov wrote:
> > >>>> Added FM radio support to Avermedia AVerTV MCE 116 Plus card
> > >>>>
> > >>> What leaves me puzzled, radio only works ok with ivtv newi2c=1
> > >>>
> > >>> With default newi2c audio is tinny, metallic, with some strange static.
> > >>> Similar problem with pvr-150 was reported years ago, guess issue is still
> > >>> unresolved, perhaps something with cx25840..
> > >> This particular "tinny" audio problem is definitely I2C speed related, to be
> > >> more precise, audio only goes bad if i2c-algo-bit is being run with udelay
> > >> less than 15, i.e. i2c bus frequency is higher than 30 KHz.
> > >>
> > >> So with default udelay=10 or udelay=5 (optimal for IR reciever on that board)
> > >> radio goes bad. Running with newi2c=1 is ok, but again it isn't optimal for IR
> > >> reciever on AVerTV M116.
> > >>
> > >> I2C reads/writes to cx25840 themself are ok, verified using register readback
> > >> after each write/write4. Problem seems to be that with cx25840 register writes
> > >> coming too fast on higher i2c bus speed, switching register 0x808 _from_ 
> > >> TV standard autodetection mode (0xff) _to_ FM radio mode (0xf9) leaves chip 
> > >> audio detection routine in inconsistent state.
> > >>
> > >> The only solution I found is to do standard routine (assert_reset + write +
> > >> deassert_reset) followed by 50ms delay and another reset.
> > >>
> > >> Following patch works_for_me, can be improved to only delay/doublereset when
> > >> really needed, etc. Andy, could you comment/review?
> > > 
> > > Aleksandr,
> > > 
> > > Could you provide your Signed-off-by for this patch?  I'm going to
> > > commit it as is.
> > > 
> > > Thanks,
> > > Andy
> > > 
> > >> diff --git a/linux/drivers/media/video/cx25840/cx25840-core.c b/linux/drivers/media/video/cx25840/cx25840-core.c
> > >> --- a/linux/drivers/media/video/cx25840/cx25840-core.c
> > >> +++ b/linux/drivers/media/video/cx25840/cx25840-core.c
> > >> @@ -626,7 +642,13 @@
> > >>  	if (state->radio) {
> > >>  		cx25840_write(client, 0x808, 0xf9);
> > >>  		cx25840_write(client, 0x80b, 0x00);
> > >> -	}
> > >> +		/* Double reset cx2384x after setting FM radio mode, helps to
> > >> +		   avoid "tinny" audio when ivtv I2C bus is being run on
> > >> +		   frequency higher than 30 KHz */
> > >> +		cx25840_and_or(client, 0x810, ~0x01, 0);
> > >> +		msleep(50);
> > >> +		cx25840_and_or(client, 0x810, ~0x01, 1);
> > >> +	}	
> > >>  	else if (std & V4L2_STD_525_60) {
> > >>  		/* Certain Hauppauge PVR150 models have a hardware bug
> > >>  		   that causes audio to drop out. For these models the
> > >>
> > >> --
> > >> 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
> > >>
> > > 
> > > --
> > > 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
> > 
> > 
>