Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[vdr] Re: Improved CAM support in VDR



Reinhard Walter Buchner wrote:
> 
> Hi Klaus,
> 
> YIIIPPEEE, It works!!!! with the MM.

Glad to hear that!

> I changed to Pentacrypt 1.09 and low and behold it
> works just GREAT. T*H*A*N*K**Y*O*U* !! Of course,
> everything that worked before (see my last mail)
> also still works. I reach you a virtual beer (others
> just won't fit through my phone lines ;o))
> 
> When the first CAM connection is established it takes
> about 4-5 seconds before it start decrypting.

That's the time it takes to collect the CA descriptors.

> After
> the first connection channel zapping and decrypt time
> is much much less than my Nokia 9902. As a matter of
> fact it is almost (really less than 0.5 seconds)
> unnoticable. Also much quicker than with the non
> experimental firmware. The old delay and screen goes
> black behavior is gone. Its just as quick as zapping to
> a FTA channel.

That's because VDR keeps a list of CA descriptors, so when you
switch channels it doesn't first have to scan the tables, but
can rather set up the CAM right away. Works really fast :-).

> Back to Globecrypt: Would it help yor studies if I
> reinstalled this and send you the debug messages?

Just give it a try.

> BTW: I am still having massiv problems with RTLers.
> I can record, but there is block noise in a lot of
> the pictures. Doing an instant replay makes VDR react
> VERY sluggish and neat things like FF or FR are not
> possible. So the problem is still there.

This "skuggishness" is a problem I noted in my initial
release note. It has to do with the new "Link Layer"
interface probably interfering with the replay control
data. Hopefully Holger can fix this once the actual CAM
stuff works.

> As I wrote
> to you in an Email during XMAS, your CI implementation
> is also triggered by RTL (just testing RTL II now)
> The debug that appears in /log/messages is
> 
> ERROR: CI MMI: unknown tag 9F8800

This is just a request from the CAM to close the MMI (Man Machine
Interface - the "menu"). I'll implement that.

> The VDR window regularly spits out:
> --> 00 01 A0 01 01

That's normal. VDR needs to query the CAM regularly to see if
it has anything to "say".

> The same goes for a replay. I CAN replay it, but VDR also
> becoems very sluggish and FF, etc is usually not possible.
> 
> Hmm, I just did some more testing while writing this mail,
> and found out that other (non RTL) channels exhibt the
> same behavior during a record & replay or just a replay
> alone (with two exceptions: they don't have block noise in
> the picture and the unknown tag no longer appears in the
> messages log).
> 
> Could this be your CAM implentation? (I tried ARD, SAT1,
> RTL, PRO7, RTL-II).Yes, I did set both debug lines to false
> ;o)) Might it be a good idea to turn the CAM implementaion
> into a seperatly running thread? Just as you did with the
> tuning routine?

See above for the problem with replaying. Putting CAM control
into a separate thread won't help here. The problem needs to be fixed
in the driver.

> Mind you I am NOT complaining ;o)) I think the CAM support
> just about "erases" the above problems ;o)) I hope for the others
> that you can get the rest of the CAMS up & running. I will test
> the stability of the system during the next few days and let you
> know if any other quirks turn up.
> 
> And one last thing: I have been meaning to write this, but kept
> on forgetting: Every since VDR 1.1.16, (my mods or vanilla
> makes no difference) I have had a small problem in the channels
> menu:
> 
> If I scroll through the pages, sometimes the last two or three
> entries are not shown. If I continue scrolling (left right) OR move
> the cursor bar (up down) to where they are (the entries seem to
> be there, but they are in clrTransparent ;o)), they appear. I just
> noticed this behaviour in 1.1.21, which is why I am now certain
> that it is not due to one of my mods ;o)) I also don't have any plugins
> installed. If I setup the OSD to scroll pages = yes it gets a lot worse
> (i.e. occurs very often). Setting this to no helps quite a bit, but (very)
> rarely it still randomly happens. Not a big problem for me, I just
> thought I would mention it ;o)

I believe this also has to do with the "Link Layer" interface interfering
with data transfers. Let's hope this get's cured together with the replay
problem...

Klaus
-- 
_______________________________________________________________

Klaus Schmidinger                       Phone: +49-8635-6989-10
CadSoft Computer GmbH                   Fax:   +49-8635-6989-40
Hofmark 2                               Email:   kls@cadsoft.de
D-84568 Pleiskirchen, Germany           URL:     www.cadsoft.de
_______________________________________________________________


-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe vdr" as subject.



Home | Main Index | Thread Index