2018-06-01 16:10:40

by Ivan Bornyakov

Subject: [PATCH] media: stv090x: fix if-else order

There is this code:

if (v >= 0x20) {
...
} else if (v < 0x20) {
...
} else if (v > 0x30) {
/* this branch is impossible */
}

It would be sensibly for last branch to be on the top.

Signed-off-by: Ivan Bornyakov <[email protected]>
---
drivers/media/dvb-frontends/stv090x.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/media/dvb-frontends/stv090x.c b/drivers/media/dvb-frontends/stv090x.c
index 9133f65d4623..d70eb311ebaf 100644
--- a/drivers/media/dvb-frontends/stv090x.c
+++ b/drivers/media/dvb-frontends/stv090x.c
@@ -4841,7 +4841,11 @@ static int stv090x_setup(struct dvb_frontend *fe)
}

state->internal->dev_ver = stv090x_read_reg(state, STV090x_MID);
- if (state->internal->dev_ver >= 0x20) {
+ if (state->internal->dev_ver > 0x30) {
+ /* we shouldn't bail out from here */
+ dprintk(FE_ERROR, 1, "INFO: Cut: 0x%02x probably incomplete support!",
+ state->internal->dev_ver);
+ } else if (state->internal->dev_ver >= 0x20) {
if (stv090x_write_reg(state, STV090x_TSGENERAL, 0x0c) < 0)
goto err;

@@ -4857,10 +4861,6 @@ static int stv090x_setup(struct dvb_frontend *fe)
state->internal->dev_ver);

goto err;
- } else if (state->internal->dev_ver > 0x30) {
- /* we shouldn't bail out from here */
- dprintk(FE_ERROR, 1, "INFO: Cut: 0x%02x probably incomplete support!",
- state->internal->dev_ver);
}

/* ADC1 range */
--
2.16.4

2018-07-26 19:27:48

by Mauro Carvalho Chehab

Subject: Re: [PATCH] media: stv090x: fix if-else order

Em Fri, 1 Jun 2018 19:12:21 +0300
Ivan Bornyakov <[email protected]> escreveu:

> There is this code:
>
> if (v >= 0x20) {
> ...
> } else if (v < 0x20) {
> ...
> } else if (v > 0x30) {
> /* this branch is impossible */
> }
>
> It would be sensibly for last branch to be on the top.

Have you tested it and check at the datasheets if dev_ver > 0x30 makes
sense?

If not, I would prefer, instead, to remove the dead code, as this
patch may cause regressions (adding a FIXME comment about this
special case).

>
> Signed-off-by: Ivan Bornyakov <[email protected]>
> ---
> drivers/media/dvb-frontends/stv090x.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/media/dvb-frontends/stv090x.c b/drivers/media/dvb-frontends/stv090x.c
> index 9133f65d4623..d70eb311ebaf 100644
> --- a/drivers/media/dvb-frontends/stv090x.c
> +++ b/drivers/media/dvb-frontends/stv090x.c
> @@ -4841,7 +4841,11 @@ static int stv090x_setup(struct dvb_frontend *fe)
> }
>
> state->internal->dev_ver = stv090x_read_reg(state, STV090x_MID);
> - if (state->internal->dev_ver >= 0x20) {
> + if (state->internal->dev_ver > 0x30) {
> + /* we shouldn't bail out from here */
> + dprintk(FE_ERROR, 1, "INFO: Cut: 0x%02x probably incomplete support!",
> + state->internal->dev_ver);
> + } else if (state->internal->dev_ver >= 0x20) {
> if (stv090x_write_reg(state, STV090x_TSGENERAL, 0x0c) < 0)
> goto err;
>
> @@ -4857,10 +4861,6 @@ static int stv090x_setup(struct dvb_frontend *fe)
> state->internal->dev_ver);
>
> goto err;
> - } else if (state->internal->dev_ver > 0x30) {
> - /* we shouldn't bail out from here */
> - dprintk(FE_ERROR, 1, "INFO: Cut: 0x%02x probably incomplete support!",
> - state->internal->dev_ver);
> }
>
> /* ADC1 range */



Thanks,
Mauro

2018-07-27 11:11:47

by Ivan Bornyakov

Subject: Re: [PATCH] media: stv090x: fix if-else order

On Thu, Jul 26, 2018 at 04:26:07PM -0300, Mauro Carvalho Chehab wrote:
> Em Fri, 1 Jun 2018 19:12:21 +0300
> Ivan Bornyakov <[email protected]> escreveu:
>
> > There is this code:
> >
> > if (v >= 0x20) {
> > ...
> > } else if (v < 0x20) {
> > ...
> > } else if (v > 0x30) {
> > /* this branch is impossible */
> > }
> >
> > It would be sensibly for last branch to be on the top.
>
> Have you tested it and check at the datasheets if dev_ver > 0x30 makes
> sense?
>
> If not, I would prefer, instead, to remove the dead code, as this
> patch may cause regressions (adding a FIXME comment about this
> special case).
>

Hi, Mauro!

Actually, I did not test it, because I did not have the hardware.
But, in the other places of the code, the same if cases are used.

Here are a couple examples:

static int stv090x_dvbs_track_crl(struct stv090x_state *state)
{
if (state->internal->dev_ver >= 0x30) {
...
} else {
...
}

...
}

static u32 stv090x_srate_srch_coarse(struct stv090x_state *state)
{
...

if (state->internal->dev_ver >= 0x30) {
...
} else if (state->internal->dev_ver >= 0x20) {
...
}

...
}