[PATCH 1/2] staging: media: av7110: replace msleep with usleep_range

[PATCH 1/2] staging: media: av7110: replace msleep with usleep_range

[André Moreira]

The msleep() function is not precise for short delays under 20ms.
Replace it with usleep_range() to provide more accurate timing
and avoid unnecessary scheduler overhead.

Signed-off-by: André Moreira <andrem.33333@gmail.com>
---
 drivers/staging/media/av7110/av7110.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/media/av7110/av7110.c b/drivers/staging/media/av7110/av7110.c
index 862aee993889..e77be16f442c 100644
--- a/drivers/staging/media/av7110/av7110.c
+++ b/drivers/staging/media/av7110/av7110.c
@@ -1818,7 +1818,7 @@ static int nexusca_stv0297_tuner_set_params(struct dvb_frontend *fe)
 		if (i2c_transfer(&av7110->i2c_adap, &readmsg, 1) == 1)
 			if (data[0] & 0x40)
 				break;
-		msleep(10);
+		usleep_range(10000, 11000);
 	}
 
 	return 0;
-- 
2.43.0

[PATCH 2/2] staging: media: av7110: refactor av7110_start_feed to reduce indentation

[André Moreira]

The av7110_start_feed function contains heavily nested if-statements,
causing excessive indentation and violating code style guidelines.
Refactor the logic inside the DMX_MEMORY_FE case by inverting the
conditional check to break early, reducing the indentation level.

Signed-off-by: André Moreira <andrem.33333@gmail.com>
---
 drivers/staging/media/av7110/av7110.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/staging/media/av7110/av7110.c b/drivers/staging/media/av7110/av7110.c
index e77be16f442c..2e32acf897ff 100644
--- a/drivers/staging/media/av7110/av7110.c
+++ b/drivers/staging/media/av7110/av7110.c
@@ -942,16 +942,16 @@ static int av7110_start_feed(struct dvb_demux_feed *feed)
 		    (feed->pes_type <= DMX_PES_PCR)) {
 			switch (demux->dmx.frontend->source) {
 			case DMX_MEMORY_FE:
-				if (feed->ts_type & TS_DECODER)
-					if (feed->pes_type < 2 &&
-					    !(demux->pids[0] & 0x8000) &&
-					    !(demux->pids[1] & 0x8000)) {
-						dvb_ringbuffer_flush_spinlock_wakeup(&av7110->avout);
-						dvb_ringbuffer_flush_spinlock_wakeup(&av7110->aout);
-						ret = av7110_av_start_play(av7110, RP_AV);
-						if (!ret)
-							demux->playing = 1;
-					}
+				if (feed->pes_type >= 2 ||
+				    (demux->pids[0] & 0x8000) ||
+				    (demux->pids[1] & 0x8000))
+					break;
+
+				dvb_ringbuffer_flush_spinlock_wakeup(&av7110->avout);
+				dvb_ringbuffer_flush_spinlock_wakeup(&av7110->aout);
+				ret = av7110_av_start_play(av7110, RP_AV);
+				if (!ret)
+					demux->playing = 1;
 				break;
 			default:
 				ret = dvb_feed_start_pid(feed);
-- 
2.43.0

Re: [PATCH 2/2] staging: media: av7110: refactor av7110_start_feed to reduce indentation

[Dan Carpenter]

On Fri, Jun 19, 2026 at 07:15:24PM -0300, André Moreira wrote:
> The av7110_start_feed function contains heavily nested if-statements,
> causing excessive indentation and violating code style guidelines.
> Refactor the logic inside the DMX_MEMORY_FE case by inverting the
> conditional check to break early, reducing the indentation level.
> 
> Signed-off-by: André Moreira <andrem.33333@gmail.com>

Imagine I read this commit message

> ---
>  drivers/staging/media/av7110/av7110.c | 20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/staging/media/av7110/av7110.c b/drivers/staging/media/av7110/av7110.c
> index e77be16f442c..2e32acf897ff 100644
> --- a/drivers/staging/media/av7110/av7110.c
> +++ b/drivers/staging/media/av7110/av7110.c
> @@ -942,16 +942,16 @@ static int av7110_start_feed(struct dvb_demux_feed *feed)
>  		    (feed->pes_type <= DMX_PES_PCR)) {
>  			switch (demux->dmx.frontend->source) {
>  			case DMX_MEMORY_FE:
> -				if (feed->ts_type & TS_DECODER)

And the first thing that I saw is that you removed this check.  There
is no information in the email why this is okay.  I have to look at
the code itself to see that.

The patch is fine, but I don't want to have to read the code to find my
answers.  Re-write the commit message and resend.

regards,
dan carpenter

> -					if (feed->pes_type < 2 &&
> -					    !(demux->pids[0] & 0x8000) &&
> -					    !(demux->pids[1] & 0x8000)) {
> -						dvb_ringbuffer_flush_spinlock_wakeup(&av7110->avout);
> -						dvb_ringbuffer_flush_spinlock_wakeup(&av7110->aout);
> -						ret = av7110_av_start_play(av7110, RP_AV);
> -						if (!ret)
> -							demux->playing = 1;
> -					}
> +				if (feed->pes_type >= 2 ||
> +				    (demux->pids[0] & 0x8000) ||
> +				    (demux->pids[1] & 0x8000))
> +					break;
> +
> +				dvb_ringbuffer_flush_spinlock_wakeup(&av7110->avout);
> +				dvb_ringbuffer_flush_spinlock_wakeup(&av7110->aout);
> +				ret = av7110_av_start_play(av7110, RP_AV);
> +				if (!ret)
> +					demux->playing = 1;
>  				break;
>  			default:
>  				ret = dvb_feed_start_pid(feed);

Re: [PATCH 1/2] staging: media: av7110: replace msleep with usleep_range

[Dan Carpenter]

On Fri, Jun 19, 2026 at 07:15:23PM -0300, André Moreira wrote:
> The msleep() function is not precise for short delays under 20ms.
> Replace it with usleep_range() to provide more accurate timing
> and avoid unnecessary scheduler overhead.
> 
> Signed-off-by: André Moreira <andrem.33333@gmail.com>
> ---

We don't accept this kind of patch without testing.  But also this
is not the current way to do it.  Do a search for "usleep_range staging"
on lore.kernel.org to find the current way.

regards,
dan carpenter