PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Manu Abraham]

Hi,

As discussed prior, the following changes help to advertise a
frontend's delivery system capabilities.

Sending out the patches as they are being worked out.

The following patch series are applied against media_tree.git
after the following commit

commit e9eb0dadba932940f721f9d27544a7818b2fa1c5
Author: Hans Verkuil <hans.verkuil@cisco.com>
Date:   Tue Nov 8 11:02:34 2011 -0300

    [media] V4L menu: add submenu for platform devices


Regards,
Manu

Re: PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Michael Krufky]

On Mon, Nov 21, 2011 at 4:05 PM, Manu Abraham <abraham.manu@gmail.com> wrote:
> Hi,
>
> As discussed prior, the following changes help to advertise a
> frontend's delivery system capabilities.
>
> Sending out the patches as they are being worked out.
>
> The following patch series are applied against media_tree.git
> after the following commit
>
> commit e9eb0dadba932940f721f9d27544a7818b2fa1c5
> Author: Hans Verkuil <hans.verkuil@cisco.com>
> Date:   Tue Nov 8 11:02:34 2011 -0300
>
>    [media] V4L menu: add submenu for platform devices
>
>
> Regards,
> Manu

I am on board with this change -- it is a positive move in the right
direction.  I believe that after this is merged, we may be able to
obsolete and remove the set_params callback.  In fact, we can even
obsolete the set_analog_params callback as well, using set_state as
the single entry point for setting the tuner.  Of course, one step at
a time -- this is great for now.  We should consider the other
optimizations after this has been merged and tested. :-)

Reviewed-by: Michael Krufky <mkrufky@linuxtv.org>

Re: PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Andreas Oberritter]

On 21.11.2011 22:05, Manu Abraham wrote:
> Hi,
> 
> As discussed prior, the following changes help to advertise a
> frontend's delivery system capabilities.
> 
> Sending out the patches as they are being worked out.
> 
> The following patch series are applied against media_tree.git
> after the following commit

Patches 7, 9 and 10 semm to be unneeded, because they just set the defaults.

Regarding the patches adding SYS_DSS: If I remember correctly, DSS
doesn't use MPEG-2 TS packets. Do we have a way to deliver DSS payload
to userspace?

Regards,
Andreas

Re: PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Michael Krufky]

[-- Attachment #1: Type: text/plain, Size: 772 bytes --]

On Mon, Nov 21, 2011 at 7:22 PM, Andreas Oberritter <obi@linuxtv.org> wrote:
> On 21.11.2011 22:05, Manu Abraham wrote:
>> Hi,
>>
>> As discussed prior, the following changes help to advertise a
>> frontend's delivery system capabilities.
>>
>> Sending out the patches as they are being worked out.
>>
>> The following patch series are applied against media_tree.git
>> after the following commit
>
> Patches 7, 9 and 10 semm to be unneeded, because they just set the defaults.
>
> Regarding the patches adding SYS_DSS: If I remember correctly, DSS
> doesn't use MPEG-2 TS packets. Do we have a way to deliver DSS payload
> to userspace?
>
> Regards,
> Andreas

I need this as well for delivering ATSC-MH raw payload to userspace.
I propose the following patch (attached)

[-- Attachment #2: dvb-demux-add-functionality-to-send-raw-payload-to-t.patch --]
[-- Type: application/octet-stream, Size: 1964 bytes --]

From 0320a7e10ebb343d0f9612035dd099273d91fc15 Mon Sep 17 00:00:00 2001
From: Michael Krufky <mkrufky@kernellabs.com>
Date: Sat, 27 Aug 2011 17:46:37 -0400
Subject: [PATCH 05/13] dvb-demux: add functionality to send raw payload to the dvr device

If your driver needs to deliver the raw payload to userspace without
passing through the kernel demux, use function: dvb_dmx_swfilter_raw

Signed-off-by: Michael Krufky <mkrufky@kernellabs.com>
---
 drivers/media/dvb/dvb-core/dvb_demux.c |   10 ++++++++++
 drivers/media/dvb/dvb-core/dvb_demux.h |    2 ++
 2 files changed, 12 insertions(+), 0 deletions(-)

diff --git a/drivers/media/dvb/dvb-core/dvb_demux.c b/drivers/media/dvb/dvb-core/dvb_demux.c
index faa3671..d82469f 100644
--- a/drivers/media/dvb/dvb-core/dvb_demux.c
+++ b/drivers/media/dvb/dvb-core/dvb_demux.c
@@ -568,6 +568,16 @@ void dvb_dmx_swfilter_204(struct dvb_demux *demux, const u8 *buf, size_t count)
 }
 EXPORT_SYMBOL(dvb_dmx_swfilter_204);
 
+void dvb_dmx_swfilter_raw(struct dvb_demux *demux, const u8 *buf, size_t count)
+{
+	spin_lock(&demux->lock);
+
+	demux->feed->cb.ts(buf, count, NULL, 0, &demux->feed->feed.ts, DMX_OK);
+
+	spin_unlock(&demux->lock);
+}
+EXPORT_SYMBOL(dvb_dmx_swfilter_raw);
+
 static struct dvb_demux_filter *dvb_dmx_filter_alloc(struct dvb_demux *demux)
 {
 	int i;
diff --git a/drivers/media/dvb/dvb-core/dvb_demux.h b/drivers/media/dvb/dvb-core/dvb_demux.h
index a7d876f..fa7188a 100644
--- a/drivers/media/dvb/dvb-core/dvb_demux.h
+++ b/drivers/media/dvb/dvb-core/dvb_demux.h
@@ -145,5 +145,7 @@ void dvb_dmx_swfilter_packets(struct dvb_demux *dvbdmx, const u8 *buf,
 void dvb_dmx_swfilter(struct dvb_demux *demux, const u8 *buf, size_t count);
 void dvb_dmx_swfilter_204(struct dvb_demux *demux, const u8 *buf,
 			  size_t count);
+void dvb_dmx_swfilter_raw(struct dvb_demux *demux, const u8 *buf,
+			  size_t count);
 
 #endif /* _DVB_DEMUX_H_ */
-- 
1.7.0.4

Re: PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Manu Abraham]

On Tue, Nov 22, 2011 at 5:52 AM, Andreas Oberritter <obi@linuxtv.org> wrote:
> On 21.11.2011 22:05, Manu Abraham wrote:
>> Hi,
>>
>> As discussed prior, the following changes help to advertise a
>> frontend's delivery system capabilities.
>>
>> Sending out the patches as they are being worked out.
>>
>> The following patch series are applied against media_tree.git
>> after the following commit
>
> Patches 7, 9 and 10 semm to be unneeded, because they just set the defaults.
>

cx24116 AFAIR handles dss, but no code exists in the driver to handle the same.
ds3000, I have no clue
tda10071, is a derivative of the CX demods and likely supports DSS,
just like the ST ones.
but that said no code exists in the driver.

As you state, yes, it is pointless to have these 3 patches, but then I
thought maybe if someone adds in DSS it would be okay. DSS hasn't been
added to most frontends for the reasons

- earlier there was no support from the frontend API
- need to handle it with demux as well.

Well, anyway patch exists, so anyone can add it in later if they need
when adding support to the driver.

> Regarding the patches adding SYS_DSS: If I remember correctly, DSS
> doesn't use MPEG-2 TS packets. Do we have a way to deliver DSS payload
> to userspace?

Yes, even the SYNC is different, length is different too. Somebody
wrote a parser a while back. But nothing happened on that front due to
the mentioned 2 reasons.

there are 2 options,

- send the captured packets as it is to userspace
- based on delivery system (dvb-s2 has generic streams as well, other
than MPEG2-TS, didn't look at dvb-t2, most likely the same option
exists there as well), look for sync and length, send packets to
userspace.

Regards,
Manu

Re: PATCH 00/13: Enumerate DVB frontend Delivery System capabilities to identify devices correctly.

[Mauro Carvalho Chehab]

Em 21-11-2011 22:43, Manu Abraham escreveu:
> On Tue, Nov 22, 2011 at 5:52 AM, Andreas Oberritter <obi@linuxtv.org> wrote:
>> On 21.11.2011 22:05, Manu Abraham wrote:
>>> Hi,
>>>
>>> As discussed prior, the following changes help to advertise a
>>> frontend's delivery system capabilities.
>>>
>>> Sending out the patches as they are being worked out.
>>>
>>> The following patch series are applied against media_tree.git
>>> after the following commit
>>
>> Patches 7, 9 and 10 semm to be unneeded, because they just set the defaults.
>>
> 
> cx24116 AFAIR handles dss, but no code exists in the driver to handle the same.
> ds3000, I have no clue
> tda10071, is a derivative of the CX demods and likely supports DSS,
> just like the ST ones.

cx24123 also handles dss, but the driver doesn't support it, AFAICT.

> but that said no code exists in the driver.
> 
> As you state, yes, it is pointless to have these 3 patches, 

Yes, with the current way.

> but then I
> thought maybe if someone adds in DSS it would be okay. DSS hasn't been
> added to most frontends for the reasons
> 
> - earlier there was no support from the frontend API
> - need to handle it with demux as well.
> 
> Well, anyway patch exists, so anyone can add it in later if they need
> when adding support to the driver.
> 
>> Regarding the patches adding SYS_DSS: If I remember correctly, DSS
>> doesn't use MPEG-2 TS packets. Do we have a way to deliver DSS payload
>> to userspace?
> 
> Yes, even the SYNC is different, length is different too. Somebody
> wrote a parser a while back. But nothing happened on that front due to
> the mentioned 2 reasons.
> 
> there are 2 options,
> 
> - send the captured packets as it is to userspace
> - based on delivery system (dvb-s2 has generic streams as well, other
> than MPEG2-TS, didn't look at dvb-t2, most likely the same option
> exists there as well), look for sync and length, send packets to
> userspace.

Just flagging that the driver supports DSS won't work, as there's currently
no way for userspace to get it.

My suggestion is to delay those patches to be added after we have a solution
that works for DSS, or if we end by getting rid of the compat code inside
the drivers (e. g. use struct dvb_frontend_parameters only at the part of the
DVB core that provides backward support for v3 calls).

> 
> Regards,
> Manu
> --
> 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

Regards,
Mauro