lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180704234244.32d20f6b@coco.lan>
Date:   Wed, 4 Jul 2018 23:42:44 -0300
From:   Mauro Carvalho Chehab <mchehab+samsung@...nel.org>
To:     "Katsuhiro Suzuki" <suzuki.katsuhiro@...ionext.com>
Cc:     <linux-media@...r.kernel.org>,
        "Masami Hiramatsu" <masami.hiramatsu@...aro.org>,
        "Jassi Brar" <jaswinder.singh@...aro.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3] media: dvb-frontends: add Socionext SC1501A ISDB-S/T
 demodulator driver

Em Thu, 5 Jul 2018 10:58:42 +0900
"Katsuhiro Suzuki" <suzuki.katsuhiro@...ionext.com> escreveu:

> Hi Mauro,
> 
> > -----Original Message-----
> > From: Mauro Carvalho Chehab <mchehab+samsung@...nel.org>
> > Sent: Thursday, July 5, 2018 1:58 AM
> > To: Suzuki, Katsuhiro/鈴木 勝博 <suzuki.katsuhiro@...ionext.com>
> > Cc: linux-media@...r.kernel.org; Masami Hiramatsu  
> <masami.hiramatsu@...aro.org>;
> > Jassi Brar <jaswinder.singh@...aro.org>; linux-arm-kernel@...ts.infradead.org;
> > linux-kernel@...r.kernel.org
> > Subject: Re: [PATCH v3] media: dvb-frontends: add Socionext SC1501A ISDB-S/T
> > demodulator driver
> > 
> > Hi Katsuhiro-san,
> > 
> > Em Thu, 21 Jun 2018 12:17:48 +0900
> > Katsuhiro Suzuki <suzuki.katsuhiro@...ionext.com> escreveu:
> >   
> > > This patch adds a frontend driver for the Socionext SC1501A series
> > > and Socionext MN88443x ISDB-S/T demodulators.  
> > 
> > Sorry for taking so long to review it. We're missing a sub-maintainer
> > for DVB, with would otherwise speed up reviews of DVB patches.  
> 
> No problem, thank you for reviewing! I appreciate it.
> 
> 
> > >
> > > The maximum and minimum frequency of Socionext SC1501A comes from
> > > ISDB-S and ISDB-T so frequency range is the following:
> > >   - ISDB-S (BS/CS110 IF frequency in kHz, Local freq 10.678GHz)
> > >     - Min: BS-1: 1032000 => 1032.23MHz
> > >     - Max: ND24: 2701000 => 2070.25MHz
> > >   - ISDB-T (in Hz)
> > >     - Min: ch13: 470000000 => 470.357857MHz
> > >     - Max: ch62: 770000000 => 769.927857MHz  
> > 
> > There is actually an error on that part of the driver. Right now,
> > the DVB core expects Satellite frequencies (DVB-S, ISDB-S, ...)
> > in kHz. For all other delivery systems, it is in Hz.
> > 
> > It is this way due to historic reasons. While it won't be hard to
> > change the core, that would require to touch all Satellite drivers.
> > 
> > As there are very few frontend drivers that accept both Satellite
> > and Terrestrial standards, what we do, instead, is to setup
> > two frontends. See, for example, drivers/media/dvb-frontends/helene.c.
> >   
> 
> Thank you for describing it. I understand our device is rare case, and 
> the reason why Helene has Terrestrial and Satellite structures.
> 
> I'm using MN884434 device that has 2 cores. I want to setup DVB adapter 
> devices (/dev/dvb/adapter0/*) for our frontend system as the following:
> 
>   - adapter0: for core 0, ISDB-T, ISDB-S
>   - adapter1: for core 1, ISDB-T, ISDB-S

Yeah, that is what it was supposed to work, if the core was ready for it.

> But it seems one DVB adapter device support only ISDB-T or only ISDB-S 
> if I divide structures. So I define the adapters as the following:
> 
>   - adapter0: for core 0, ISDB-T
>   - adapter1: for core 0, ISDB-S
>   - adapter2: for core 1, ISDB-T
>   - adapter3: for core 1, ISDB-S
> 
> Is this correct?

That's the way the current driver with uses helene does.

> 
> 
> > ...  
> > > +static const struct dvb_frontend_ops sc1501a_ops = {
> > > +	.delsys = { SYS_ISDBS, SYS_ISDBT },
> > > +	.info = {
> > > +		.name          = "Socionext SC1501A",
> > > +		.frequency_min = 1032000,
> > > +		.frequency_max = 770000000,
> > > +		.caps = FE_CAN_INVERSION_AUTO | FE_CAN_FEC_AUTO |
> > > +			FE_CAN_QAM_AUTO | FE_CAN_TRANSMISSION_MODE_AUTO |
> > > +			FE_CAN_GUARD_INTERVAL_AUTO | FE_CAN_HIERARCHY_AUTO,
> > > +	},
> > > +
> > > +	.sleep                   = sc1501a_sleep,
> > > +	.set_frontend            = sc1501a_set_frontend,
> > > +	.get_tune_settings       = sc1501a_get_tune_settings,
> > > +	.read_status             = sc1501a_read_status,
> > > +};  
> > 
> > In other words, you'll need to declare two structs here, one for ISDB-T
> > and another one for ISDB-S.
> >   
> 
> OK, I'm going to divide this structure for Terrestrial and Satellite. And
> add attach functions same as Helene driver.
> 
> I'll send v4 patch.

I ended by writing two patches that should be solving the issues
inside the core. With them[1], it will work the way you want.

There is a catch: you'll need to convert Helene to have a single
entry and be sure that the driver that currently uses it (netup_unidvb)
will keep working. I guess I have one such hardware here for testing.

[1] after tested/reviewed - I didn't test them yet. Feel free to test.

So, please look at the two patches I sent today to the mailing list.

(not sure why, they're taking a long time to arrive there - perhaps
vger has some issues).

I added them on this tree:
	https://git.linuxtv.org/mchehab/experimental.git/log/?h=dvb_freq_hz

it is the last two patches there:
	- https://git.linuxtv.org/mchehab/experimental.git/commit/?h=dvb_freq_hz&id=b3d63a8f038d136b26692bc3a14554960e767f4a
	- https://git.linuxtv.org/mchehab/experimental.git/commit/?h=dvb_freq_hz&id=2a369e8faf3b277baff4026371f298e95c84fbb2

I'm not sure if all applications will do the right thing, though, as
it will depend  if they query the capabilities before or after switching
to a different delivery system. If it get caps before and store them
in Hz, apps will work, but tests are required.

> 
> 
> > Yeah, I know that this sucks. If you are in the mood of touching the
> > DVB core, I'm willing to consider a patch that would fix this, provided
> > that it won't break backward compatibility with other drivers (or would
> > convert the other satellite drivers to use the new way).
> > 
> > Thanks,
> > Mauro  
> 
> Hmm, I don't know the details of DVB core, I try to investigate it.
> 
> 
> Regards,
> --
> Katsuhiro Suzuki
> 
> 
> 



Thanks,
Mauro

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ