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  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:   Tue, 26 Mar 2019 16:14:51 +0200
From:   Sakari Ailus <sakari.ailus@...ux.intel.com>
To:     Mickael GUENE <mickael.guene@...com>
Cc:     "linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
        Hugues FRUCHET <hugues.fruchet@...com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Matt Ranostay <matt.ranostay@...sulko.com>,
        Petr Cvek <petrcvekcz@...il.com>,
        Akinobu Mita <akinobu.mita@...il.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "David S. Miller" <davem@...emloft.net>,
        Ben Kao <ben.kao@...el.com>,
        Nicolas Ferre <nicolas.ferre@...rochip.com>,
        Todor Tomov <todor.tomov@...aro.org>,
        Rui Miguel Silva <rui.silva@...aro.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Hans Verkuil <hverkuil@...all.nl>,
        Ricardo Ribalda Delgado <ricardo@...alda.com>,
        Jacopo Mondi <jacopo+renesas@...ndi.org>,
        Tianshu Qiu <tian.shu.qiu@...el.com>,
        Bingbu Cao <bingbu.cao@...el.com>
Subject: Re: [PATCH v3 2/2] media:st-mipid02: MIPID02 CSI-2 to PARALLEL
 bridge driver

On Tue, Mar 26, 2019 at 02:12:52PM +0000, Mickael GUENE wrote:
> Sakari,
> 
> On 3/26/19 14:54, Sakari Ailus wrote:
> > Hi Mickael,
> > 
> > On Tue, Mar 26, 2019 at 12:57:03PM +0000, Mickael GUENE wrote:
> > ...
> >>>> +static int mipid02_set_fmt(struct v4l2_subdev *sd,
> >>>> +			   struct v4l2_subdev_pad_config *cfg,
> >>>> +			   struct v4l2_subdev_format *format)
> >>>> +{
> >>>> +	struct v4l2_mbus_framefmt *mbus_fmt = &format->format;
> >>>> +	struct mipid02_dev *bridge = to_mipid02_dev(sd);
> >>>> +	struct i2c_client *client = bridge->i2c_client;
> >>>> +	struct v4l2_mbus_framefmt *fmt;
> >>>> +	int ret;
> >>>> +
> >>>> +	dev_dbg(&client->dev, "%s for %d", __func__, format->pad);
> >>>> +
> >>>> +	if (format->pad >= MIPID02_PAD_NB)
> >>>> +		return -EINVAL;
> >>>> +	/* second CSI-2 pad not yet supported */
> >>>> +	if (format->pad == 1)
> >>>> +		return -EINVAL;
> >>>> +
> >>>> +	mutex_lock(&bridge->lock);
> >>>> +
> >>>> +	if (bridge->streaming) {
> >>>> +		ret = -EBUSY;
> >>>> +		goto error;
> >>>> +	}
> >>>> +
> >>>> +	if (format->which == V4L2_SUBDEV_FORMAT_TRY)
> >>>> +		fmt = v4l2_subdev_get_try_format(sd, cfg, format->pad);
> >>>> +	else
> >>>> +		fmt = &bridge->fmt;
> >>>> +
> >>>> +	*fmt = *mbus_fmt;
> >>>
> >>> What are the limits of the hardware regarding the size of the image? Aren't
> >>> there any?
> >>>
> >> There are no limits for image size.
> >>> The format on the sink pad needs to be propagated to the source pad as
> >>> well. I presume there's nothing to set on the sink pad for this device, is
> >>> there?
> >>  User only need to set format code so the driver can configure hardware.
> >>  In the mipid02 we have always the same format for sink and source. So I
> >>  only store one configuration during set_fmt when called from either pad0
> >> (sink) or pad2 (source). Is it the correct way to implement it ? or
> >> should I only accept set_fmt on pad0 ?
> >>  For get_fmt I return stored configuration for pad0 and pad2.
> > 
> > Only 76 or so characters per line, please.
> > 
> > For pad 0 (sink) the format must be settable freely (as you don't have any
> > hardware restrictions) and on the pad 2 (source) the driver converts the
> > format set on the pad 0 according to the hardware functionality. This is
> > what the link validation from the source pad onwards is based on.
> > 
>  So for a set_fmt on pad 2 I return current configuration ? (as
> in my case I have the same configuration for pad 0 and pad2)
>  And I only update format when set_fmt is called for pad 0 ?

Correct. Do note that the format is different in cases (uyvy8_1x16 vs. 2x8)
as you're dealing with two different kinds of busses.

-- 
Sakari Ailus
sakari.ailus@...ux.intel.com

Powered by blists - more mailing lists