Message-ID: <20180521171415.00c56487@vento.lan>
Date:   Mon, 21 May 2018 17:14:15 -0300
From:   Mauro Carvalho Chehab <mchehab+samsung@...nel.org>
To:     "Gustavo A. R. Silva" <gustavo@...eddedor.com>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
        Devin Heitmueller <dheitmueller@...uxtv.org>
Subject: Re: [media] duplicate code in media drivers

Em Mon, 21 May 2018 14:39:51 -0500
"Gustavo A. R. Silva" <gustavo@...eddedor.com> escreveu:

> Hi Mauro,
> 
> I found some duplicate code with the help of Coccinelle and Coverity. Notice that these are not code patches, they only point out the duplicate code in some media drivers:
> 
> diff -u -p drivers/media/pci/bt8xx/dvb-bt8xx.c /tmp/nothing/media/pci/bt8xx/dvb-bt8xx.c
> --- drivers/media/pci/bt8xx/dvb-bt8xx.c
> +++ /tmp/nothing/media/pci/bt8xx/dvb-bt8xx.c
> @@ -389,9 +389,7 @@ static int advbt771_samsung_tdtc9251dh0_
>         else if (c->frequency < 600000000)
>                 bs = 0x08;
>         else if (c->frequency < 730000000)
> -               bs = 0x08;
>         else
> -               bs = 0x08;
> 
>         pllbuf[0] = 0x61;
>         pllbuf[1] = div >> 8;


Hmm... I *suspect* that "bs" here controls the frequency range for the
tuner. Analog tuners have separate frequency regions that are controlled
via a register, into a 4 or 5 bytes I2C sequence. They're all somewhat
a clone of an old Philips design.

It should be safe to convert the "BS" sequence on something like:

	if (c->frequency < 173000000)
                bs = 0x01;
        else if (c->frequency < 470000000)
                bs = 0x02;
        else 
                bs = 0x08;



> diff -u -p drivers/media/usb/dvb-usb/dib0700_devices.c /tmp/nothing/media/usb/dvb-usb/dib0700_devices.c
> --- drivers/media/usb/dvb-usb/dib0700_devices.c
> +++ /tmp/nothing/media/usb/dvb-usb/dib0700_devices.c
> @@ -1741,13 +1741,6 @@ static int dib809x_tuner_attach(struct d
>         struct dib0700_adapter_state *st = adap->priv;
>         struct i2c_adapter *tun_i2c = st->dib8000_ops.get_i2c_master(adap->fe_adap[0].fe, DIBX000_I2C_INTERFACE_TUNER, 1);
> 
> -       if (adap->id == 0) {
> -               if (dvb_attach(dib0090_register, adap->fe_adap[0].fe, tun_i2c, &dib809x_dib0090_config) == NULL)
> -                       return -ENODEV;
> -       } else {
> -               if (dvb_attach(dib0090_register, adap->fe_adap[0].fe, tun_i2c, &dib809x_dib0090_config) == NULL)
> -                       return -ENODEV;
> -       }

I'm almost sure that, on the second if, it should be adap->fe_adap[1].fe.
I tried in the past to check this, but didn't got an answer from the one
that wrote the code.

Maybe we could add a /* FIXME: check if it is fe_adap[1] */ on the
second clause.

> 
>         st->set_param_save = adap->fe_adap[0].fe->ops.tuner_ops.set_params;
>         adap->fe_adap[0].fe->ops.tuner_ops.set_params = dib8096_set_param_override;
> diff -u -p drivers/media/dvb-frontends/mb86a16.c /tmp/nothing/media/dvb-frontends/mb86a16.c
> --- drivers/media/dvb-frontends/mb86a16.c
> +++ /tmp/nothing/media/dvb-frontends/mb86a16.c
> @@ -1466,9 +1466,7 @@ static int mb86a16_set_fe(struct mb86a16
>                                                         wait_t = (1572864 + state->srate / 2) / state->srate;
>                                                 if (state->srate < 5000)
>                                                         /* FIXME ! , should be a long wait ! */
> -                                                       msleep_interruptible(wait_t);
>                                                 else
> -                                                       msleep_interruptible(wait_t);

I suspect that the goal here is to point that sleeping for
(1572864 + state->srate / 2) / state->srate when srate is low will mean
that it will take a lot of time to converge (probably causing timeout at
userspace).

Basically, if srate is < 5000, the sleep time will be between
314 and 1575364 ms. The worse case scenario - although not realistic,
in practice - is to wait up to 26 seconds. This is a very long time!

Probably, the right fix here would be to check if wait_t is bigger than
a certain amount of time. If so, return an error.

I'm not against removing the if, but, if so, better to add a /* FIXME */
block explaining that.

That's said, this is an old device. I doubt anyone would fix it.


> 
>                                                 if (sync_chk(state, &junk) == 0) {
>                                                         iq_vt_set(state, 1);
> diff -u -p drivers/media/dvb-frontends/au8522_decoder.c /tmp/nothing/media/dvb-frontends/au8522_decoder.c
> --- drivers/media/dvb-frontends/au8522_decoder.c
> +++ /tmp/nothing/media/dvb-frontends/au8522_decoder.c
> @@ -280,14 +280,9 @@ static void setup_decoder_defaults(struc
>                         AU8522_TOREGAAGC_REG0E5H_CVBS);
>         au8522_writereg(state, AU8522_REG016H, AU8522_REG016H_CVBS);
> 
> -       if (is_svideo) {
>                 /* Despite what the table says, for the HVR-950q we still need
>                    to be in CVBS mode for the S-Video input (reason unknown). */
>                 /* filter_coef_type = 3; */
> -               filter_coef_type = 5;
> -       } else {
> -               filter_coef_type = 5;
> -       }

Better ask Devin about this (c/c).

> 
>         /* Load the Video Decoder Filter Coefficients */
>         for (i = 0; i < NUM_FILTER_COEF; i++) {
> 
> 
> I wonder if some of the cases above were intentionally coded that way or some code needs to be removed.
> 
> Thanks
> --
> Gustavo



Thanks,
Mauro

Powered by blists - more mailing lists