Re: [PATCH v3] media: renesas: vsp1: Add underrun debug print

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 15/02/2023 00:15, Laurent Pinchart wrote:
Hi Tomi,

Thank you for the patch.

On Tue, Feb 14, 2023 at 06:42:23PM +0200, Tomi Valkeinen wrote:
Print underrun interrupts with ratelimited print.

Note that we don't enable the underrun interrupt. If we have underruns,
we don't want to get flooded with interrupts about them. It's enough to
see that an underrun happened at the end of a frame.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@xxxxxxxxxxxxxxxx>
---

Changes in v3:
- Reset underrun counter when enabling VSP

I have to say I'm not familiar enough with the VSP driver to say if
these are the correct places where to reset the counters.

It's fine. We could factor it out to a clear function, but it's not
worth it if there's nothing else to factor out. It could be done later.

There's also a
possibility of a race, but my assumption is that we cannot get underrun
interrupts for the WPF we are currently enabling.

It should be fine.

Also, I realized the underrun counter could be moved to struct
vsp1_rwpf, but as that's used also for RPF, I didn't do that change.

Another option would be to store it in the pipeline structure, as a
pipeline has one and only one WPF. What do you think ?

Hmm, the pipe is allocated and assigned as needed, isn't it? So in the irq handler we might get an underflow with !pipe. We could skip the print in that case, of course.

Is a pipe allocated every time VSP is started? Or does the allocation normally happen only once? If the former, then if the counter was stored in the pipe, that would handle clearing the counter automatically.

 Tomi




[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux