Mailing List archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[linux-dvb] Re: [PATCH] dvb-ttpci+budgetpatch integrated v22
On Tue, Jan 04, 2005 at 10:46:10PM +0100, emard@softhome.net wrote:
> On Tue, Jan 04, 2005 at 03:39:02PM +0100, Johannes Stezenbach wrote:
> > emard@softhome.net wrote:
> > > Naming of dvbnet devices is fixed, they are
> > > now called dvb000 for net0 and dvb010 for net1
> > > ("_" in the middle of dvb0_0 is used to indicate
> > > id of net* interface)
> >
> > Please don't do that. There is not need to, is there?
> > I guess that would break various scripts.
>
> HM! Something is a need, beacuse it got me oops when I first
> want to register dvb0_0 from net0 and then again I
> tried to register dvb0_0 from net1.
I would prefer dvbA_N where A is adapter number (like it is now)
and N is simply incremented regardless of whether it's net0 or net1.
> A solution for dvbX_Y would be to code adapter+net number
> into the first digit, so adapter0-3/net0 range would be
> dvb0-3_X as before, and adapter0-3/net1 would create dvb4-7 etc.
>
> Theoretically, we would have to allow for HEX digits, since a
> range of 4 adpaters x 4 nets gives 16, so some exotic
> configurations could give something like dvbA_0 as well.
>
> Is that acceptable?
What I want is that for unmodified cards (no demux1/net1)
the network devices are named dvb0_0, dvb0_1 etc. as
before. We could name network devices differently
for net1, but that's inconsistent and ugly.
Is it important to encode the netX number in the network
device name? Why?
> > > demux1 exists but receives no data.
> >
> > IMHO that's not acceptable. If you can't probe for the presence
> > of the hardware support for demux1 you need to have a
> > module parameter to enable it.
>
> I *can* probe but prior to reliable detection, it seems that
> I have to reset saa7146 :(. Better to go for module parameter.
It should be no problem to reset the saa7146, the comment
in saa7146_init_one() only applies to a few non-av7110 cards.
MiHu agreed that it would be better to enable the master
reset in saa7146_init_one() and add a black list for the
few cards which must not be reset.
Johannes
Home |
Main Index |
Thread Index