Message-ID: <20250615124752.GA1489213@ragnatech.se>
Date: Sun, 15 Jun 2025 14:47:52 +0200
From: Niklas Söderlund <niklas.soderlund+renesas@...natech.se>
To: Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc: Mauro Carvalho Chehab <mchehab@...nel.org>,
	Geert Uytterhoeven <geert+renesas@...der.be>,
	Sakari Ailus <sakari.ailus@...ux.intel.com>,
	linux-media@...r.kernel.org, linux-renesas-soc@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/3] media: rcar-vin: Check for correct capture interrupt
 event

Hi Laurent,

Thanks for your review!

On 2025-06-15 00:32:13 +0300, Laurent Pinchart wrote:
> Hi Niklas,
> 
> On Sat, Jun 14, 2025 at 04:15:44PM +0200, Niklas Söderlund wrote:
> > Depending on if the capture session deals with fields or whole frames
> > interrupts can be generated at an end of field, or end of frame event.
> > The interrupt mask is setup to generate an interrupt on one of the two
> > events depending on what is needed when the VIN is started. The end of
> > field bit is set in both cases so controlling the mask that generates an
> > interrupt have been enough to control the two use-cases.
> > 
> > Before extending the interrupt handler to deal with other types of
> > interrupt events it is needs to extended to "capture complete" check for
> > correct the use-case in operation. Without this the simplification in
> > the handler can result in corrupted frames when the mask on what type of
> > events can generate an interrupt generated can no longer be assumed to
> > only be an "capture complete" event.
> > 
> > Which bit is checked matches which bit is enabled at configuration time
> > as which event can generate an interrupt for "capture complete". There
> > is no functional change.
> > 
> > While at it switch to use the BIT() macro to describe the bit positions
> > for the interrupt functions.
> > 
> > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@...natech.se>
> > ---
> >  drivers/media/platform/renesas/rcar-vin/rcar-dma.c | 14 +++++++++-----
> >  1 file changed, 9 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/media/platform/renesas/rcar-vin/rcar-dma.c b/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > index 585b8b3dcfd8..85e44a00e0fc 100644
> > --- a/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > +++ b/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > @@ -115,11 +115,12 @@
> >  #define VNFC_S_FRAME		(1 << 0)
> >  
> >  /* Video n Interrupt Enable Register bits */
> > -#define VNIE_FIE		(1 << 4)
> > -#define VNIE_EFE		(1 << 1)
> > +#define VNIE_FIE		BIT(4)
> > +#define VNIE_EFE		BIT(1)
> >  
> >  /* Video n Interrupt Status Register bits */
> > -#define VNINTS_FIS		(1 << 4)
> > +#define VNINTS_FIS		BIT(4)
> > +#define VNINTS_EFS		BIT(1)
> >  
> >  /* Video n Data Mode Register bits */
> >  #define VNDMR_A8BIT(n)		(((n) & 0xff) << 24)
> > @@ -1034,7 +1035,7 @@ static void rvin_capture_stop(struct rvin_dev *vin)
> >  static irqreturn_t rvin_irq(int irq, void *data)
> >  {
> >  	struct rvin_dev *vin = data;
> > -	u32 status, vnms;
> > +	u32 capture, status, vnms;
> >  	int slot;
> >  	unsigned int handled = 0;
> >  	unsigned long flags;
> > @@ -1049,7 +1050,10 @@ static irqreturn_t rvin_irq(int irq, void *data)
> >  	handled = 1;
> >  
> >  	/* Nothing to do if nothing was captured. */
> > -	if (!(status & VNINTS_FIS))
> > +	capture = vin->format.field == V4L2_FIELD_NONE ||
> > +		vin->format.field == V4L2_FIELD_ALTERNATE ?
> > +		VNINTS_FIS : VNINTS_EFS;
> 
> I would find
> 
> 	capture = vin->format.field == V4L2_FIELD_NONE ||
> 		  vin->format.field == V4L2_FIELD_ALTERNATE
> 		? VNINTS_FIS : VNINTS_EFS;
> 
> easier to read, but it could be just me.

I agree it's easier to read, but my auto style checker for the file will 
not allow it. And if I make one exception from my format rules I fear 
the file will soon turn wild. The whole construct is ugly and I hope to 
be able to remove it all together in follow up work.

> 
> Do I read it correctly that you use the "Field Interrupt" with
> V4L2_FIELD_NONE ? That seems a bit weird to me, but maybe I don't get
> how the hardware operates.

I agree it's odd, and likely can be improved upon. But it mirrors what 
is done in the capture setup routine. The "Field Interrupt" is enabled 
for NONE and ALTERNATE, and the "End Frame Interrupt" is used for the 
other cases. See the 'progressive' variable in rvin_setup().

The historical reason for this oddity is that VIN placed to much logic 
in the kernel driver that should have been up to user-space, specially 
for Gen2. In this case if the video source was providing ALTERNATE then 
VIN would automatically default to INTERLACED the output. Some of this 
have been reworked and fixed over the years, and a lot of it is finally 
fixed in the Gen2 MC series on the list.

That work is however not a dependency of this work, and I feel like a 
broken record, but I have started work to refactor this and make it 
right. That work is however based on the Gen2 MC series as this allows 
me to add proper tests for all of this that covers all the different 
generations.

I did not want this series to depend on the cleanups and potentially 
grew to yet another large series that fix lots of unrelated things. I 
hope it's OK we live with this oddity a little while longer so we can 
address it properly as a separate thing.

> 
> > +	if (!(status & capture))
> >  		goto done;
> >  
> >  	/* Nothing to do if not running. */
> 
> -- 
> Regards,
> 
> Laurent Pinchart

-- 
Kind Regards,
Niklas Söderlund

Powered by blists - more mailing lists