Re: [PATCH] media: staging: ipu3-imgu: Set fields before media_entity_pads_init()

[Sakari Ailus]

Hi Dan, Hidenori,

On Fri, Jan 05, 2024 at 10:34:20AM +0300, Dan Carpenter wrote:
> On Fri, Jan 05, 2024 at 11:19:23AM +0900, Hidenori Kobayashi wrote:
> > On Thu, Jan 04, 2024 at 01:04:27PM +0300, Dan Carpenter wrote:
> > > On Thu, Dec 28, 2023 at 06:39:25PM +0900, Hidenori Kobayashi wrote:
> > > > The pad's flags is checked in media_entity_pads_init(), so it has to be
> > > > initialized beforehand. The ops initialization is also moved together
> > > > for readability.
> > > > 
> > > 
> > > How does this bug look like to a user?  What is the Fixes tag?  Does
> > > this need to be backported to stable?
> > 
> > I suppose I should have included those in the commit message.
> > 
> > 1) To a user, the imgu driver fails to probe with the following message:
> > 
> > [   14.596315] ipu3-imgu 0000:00:05.0: failed initialize subdev media entity (-22)
> > [   14.596322] ipu3-imgu 0000:00:05.0: failed to register subdev0 ret (-22)
> > [   14.596327] ipu3-imgu 0000:00:05.0: failed to register pipes (-22)
> > [   14.596331] ipu3-imgu 0000:00:05.0: failed to create V4L2 devices (-22)
> > 
> 
> Yeah.  This is super useful information.
> 
> > 2) Re Fixes tag, I see that the first commit of imgu driver already
> > initializes the flags after media_entity_pads_init(). The documentation
> > of this API ( "Drivers must set the direction of every pad ... before
> > calling media_entity_pads_init") predates the first commit. So, I guess
> > 
> > Fixes: a0ca1627b450 ("media: staging/intel-ipu3: Add v4l2 driver based on media framework")
> > 
> > 3) Re stable, I was not sure. The probe failure only appears after a
> > check was added by Commit deb866f9e3a45ae058b21765feeffae6aea6a193. That
> > check is not in linux-6.6.y branch. So I was not sure if this counts as
> > "a real bug that bothers people" mentioned in the document.
> 
> Hm...  I don't know either.  Wait for a day and see if anyone else has
> an opinion then listen to your gut and resend with whatever your gut
> says?

The suggested commit message seems good to me, the Fixes: line is
important. Thanks for digging this up! The patch should go to v6.7, too, so
please add Cc: stable (for v6.7).

-- 
Kind regards,

Sakari Ailus