Re: [PATCH] [media] drxk: Switch the delivery system on FE_SET_PROPERTY

[Mauro Carvalho Chehab <mchehab@redhat.com>]

On 09-12-2011 17:08, Antti Palosaari wrote:
> On 12/09/2011 08:58 PM, Mauro Carvalho Chehab wrote:
>> On 09-12-2011 16:26, Antti Palosaari wrote:
>>> On 12/09/2011 08:20 PM, Mauro Carvalho Chehab wrote:
>>>> The DRX-K doesn't change the delivery system at set_properties,
>>>> but do it at frontend init. This causes problems on programs like
>>>> w_scan that, by default, opens both frontends.
>>>>
>>>> Instead, explicitly set the format when set_parameters callback is
>>>> called.
>>>
>>> May I ask why you don't use mfe_shared flag instead?
>>
>> Tested with it. Works. It takes a little more time to switch, but the
>> solution will be cleaner. version 2 will follow.
>
> Yes, there is kind of loop timer which tries to take FE and swithing takes second or two
>because it waits FE is freed. I looked it when I did MFE and I did not understood why
>it was done like that.
>
> cxd2820r has earlier simple lock (inside of demod driver) that just
> returned error immediately when busy. It is a little bit mystery for
> me why mfe_shared has that kind of waiting mechanism.

Still, it doesn't make much sense, at least on w_scan, as the only thing that is
called with each adapter is FE_GET_INFO:


open("/dev/dvb/adapter0/frontend0", O_RDWR|O_NONBLOCK) = 3
ioctl(3, FE_GET_INFO, 0x635120)         = 0
write(2, "\t/dev/dvb/adapter0/frontend0 -> "..., 92	/dev/dvb/adapter0/frontend0 -> DVB-C "DRXK DVB-C": specified was DVB-T -> SEARCH NEXT ONE.
) = 92
close(3)                                = 0
open("/dev/dvb/adapter0/frontend1", O_RDWR|O_NONBLOCK) = 3
ioctl(3, FE_GET_INFO, 0x635120)         = 0
write(2, "\t/dev/dvb/adapter0/frontend1 -> "..., 52	/dev/dvb/adapter0/frontend1 -> DVB-T "DRXK DVB-T": ) = 52
write(2, "good :-)\n", 9good :-)
)               = 9
close(3)                                = 0

Still, the second open takes about 4 seconds to complete, _even_ with O_NONBLOCK.

There's something bad there, as it is violating POSIX.

>Could someone explain reason for that?

I dunno, but I think this needs to be fixed, at least when the frontend
is opened with O_NONBLOCK.

Regards,
Mauro.