linux-kernel - Re: [Patch 1/3] media: ov5640: add PIXEL

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 2 Oct 2019 10:06:15 -0500
From:   Benoit Parrot <bparrot@...com>
To:     Jacopo Mondi <jacopo@...ndi.org>
CC:     Sakari Ailus <sakari.ailus@...ux.intel.com>,
        Hans Verkuil <hverkuil@...all.nl>,
        <linux-media@...r.kernel.org>, <devicetree@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, <hugues.fruchet@...com>
Subject: Re: [Patch 1/3] media: ov5640: add PIXEL_RATE control

Jacopo Mondi <jacopo@...ndi.org> wrote on Wed [2019-Oct-02 16:32:26 +0200]:
> Hi Benoit,
> 
> On Wed, Oct 02, 2019 at 07:14:38AM -0500, Benoit Parrot wrote:
> > Hi Jacopo,
> >
> > Maybe, I miss spoke when I mentioned a helper I did not intent a framework
> > level generic function. Just a function to help in this case :)
> 
> Yes indeed, the discussion thread I linked here was mostly interesting
> because Hugues tried to do the same for LINK_FREQ iirc, and there
> where some usefult pointers.
> 
> >
> > That being said, I re-read the thread you mentioned. And as Hughes pointed
> > out dynamically generating a "working" link frequency value which can be
> > used by a CSI2 receiver to properly configure its PHY is not trivial.
> >
> > When I created this patch, I also had another to add V4L2_CID_LINK_FREQ
> > support. I am testing this against the TI CAL CSI2 receiver, which already
> > uses the V4L2_CID_PIXEL_RATE value for that purpose, so I also had a patch
> > to add support for V4L2_CID_LINK_FREQ to that driver as well.
> >
> > Unfortunately, similar to Hughes' findings I was not able to make it "work"
> > with all supported resolution/framerate.
> 
> As reported by Hugues findings, the PLL calculation procedure might be
> faulty, and the actuall frequencies on the bus are different from the
> calculated ones.
> 
> I wish I had more time to re-look at that, as they worked for my and
> Sam's use case, but deserve some rework.
> 
> >
> > Unlike my V4L2_CID_PIXEL_RATE solution which now works in all mode with the
> > same receiver.
> >
> 
> It seems to me you're reporting a fixed rate. It might make your
> receiver happy, but does not report what is acutally put on the bus.
> Am I missing something ?

No it is not fixed, the only fixed value was the initial value (which
representative of the initial/default resolution and framerate), I
fixed this in v2. The reported PIXEL_RATE is re-calculated every time there
is a s_fmt and/or framerate change and the V4L2_CID_PIXEL_RATE control
value is updated accordingly.

> 
> > So long story short I dropped the V4L2_CID_LINK_FREQ patch and focused on
> > V4L2_CID_PIXEL_RATE instead.
> >
> 
> As Sakari pointed out, going from one to the other is trivial and
> could be done on top.

As you said it could be done on top. :)

Benoit

> 
> Thanks
>    j
> 
> > Regard,
> > Benoit
> >
> > Jacopo Mondi <jacopo@...ndi.org> wrote on Wed [2019-Oct-02 09:59:51 +0200]:
> > > Hi Benoit,
> > >   +Hugues
> > >
> > > If you're considering an helper, this thread might be useful to you:
> > > https://patchwork.kernel.org/patch/11019673/
> > >
> > > Thanks
> > >    j
> > >
> >
> >