Saturday 02 October 2010 07:47:47 Guennadi Liakhovetski napisaÅ(a):
> Ok, let's take this one, but, please, address the below couple of minor
> issues in an incremental patch.
>
> On Mon, 27 Sep 2010, Janusz Krzysztofik wrote:
> > +/* write a register */
> > +static int ov6650_reg_write(struct i2c_client *client, u8 reg, u8 val)
> > +{
> > + int ret;
> > + unsigned char data[2] = { reg, val };
> > + struct i2c_msg msg = {
> > + .addr = client->addr,
> > + .flags = 0,
> > + .len = 2,
> > + .buf = data,
> > + };
> > +
> > + ret = i2c_transfer(client->adapter, &msg, 1);
> > + usleep_range(100, 1000);
>
> So, 100us are enough? Then I'd just go with udelay(100).
Guennadi,
I already tried with udelay(100), as you had suggested before, and it worked,
but then checkpatch.pl --sctirct told me:
"CHECK: usleep_range is preferred over udelay; see
Documentation/timers/timers-howto.txt
+ udelay(100);"
So, I had read Documentation/timers/timers-howto.txt again and switched to
usleep_range, as it suggested.
Please confirm if you still prefere udelay(100) over usleep_range(), and I'll
change it back.
> > static bool is_unscaled_ok(int width, int height, struct v4l2_rect *rect)
> > +{
> > + return (width > rect->width >> 1 || height > rect->height >> 1);
> > +}
>
> Ok, just one more pair of brackets to remove;)
OK.
> > +
> > +static u8 to_clkrc(struct v4l2_fract *timeperframe,
> > + unsigned long pclk_limit, unsigned long pclk_max)
> > +{
> > + unsigned long pclk;
> > +
> > + if (timeperframe->numerator && timeperframe->denominator)
> > + pclk = pclk_max * timeperframe->denominator /
> > + (FRAME_RATE_MAX * timeperframe->numerator);
> > + else
> > + pclk = pclk_max;
> > +
> > + if (pclk_limit && pclk_limit < pclk)
> > + pclk = pclk_limit;
> > +
> > + return (pclk_max - 1) / pclk;
> > +}
> > +
> > +/* set the format we will capture in */
> > +static int ov6650_s_fmt(struct v4l2_subdev *sd, struct
> > v4l2_mbus_framefmt *mf) +{
> > + struct i2c_client *client = sd->priv;
> > + struct soc_camera_device *icd = client->dev.platform_data;
> > + struct soc_camera_sense *sense = icd->sense;
> > + struct ov6650 *priv = to_ov6650(client);
> > + bool half_scale = !is_unscaled_ok(mf->width, mf->height, &priv->rect);
> > + struct v4l2_crop a = {
> > + .type = V4L2_BUF_TYPE_VIDEO_CAPTURE,
> > + .c = {
> > + .left = priv->rect.left + (priv->rect.width >> 1) -
> > + (mf->width >> (1 - half_scale)),
> > + .top = priv->rect.top + (priv->rect.height >> 1) -
> > + (mf->height >> (1 - half_scale)),
> > + .width = mf->width << half_scale,
> > + .height = mf->height << half_scale,
> > + },
> > + };
>
> Hm, this seems wrong to me... You calculate left and top to preserve the
> center, right?
Exactly.
> This is good, but: if output is unscaled, you want
>
> .left = priv->rect.left + (priv->rect.width - mf->width) / 2;
== priv->rect.left + priv->rect.width / 2 - mf->width / 2
== priv->rect.left + (priv->rect.width >> 1) - ( mf->width >> 1)
== priv->rect.left + (priv->rect.width >> 1) - ( mf->width >> (1 - 0))
>
> in this case half_scale = 0 and the above is correct. Now, is the output
> is scaled, you want
>
> .left = priv->rect.left + (priv->rect.width - mf->width * 2) / 2;
== priv->rect.left + priv->rect.width / 2 - mf->width * 2 / 2
== priv->rect.left + (priv->rect.width >> 1) - ((mf->width << 1) >> 1)
== priv->rect.left + (priv->rect.width >> 1) - ( mf->width >> (1 - 1))
> which is not, what you have above. Am I missing anything?
One of us must be ;).
> > + case V4L2_MBUS_FMT_UYVY8_2X8:
> > + dev_dbg(&client->dev, "pixel format YUYV8_2X8_BE\n");
> > + if (half_scale) {
> > + coma_mask |= COMA_RGB | COMA_BW | COMA_WORD_SWAP;
> > + coma_set |= COMA_BYTE_SWAP;
> > + } else {
> > + coma_mask |= COMA_RGB | COMA_BW;
> > + coma_set |= COMA_BYTE_SWAP | COMA_WORD_SWAP;
> > + }
> > + break;
> > + case V4L2_MBUS_FMT_VYUY8_2X8:
> > + dev_dbg(&client->dev, "pixel format YVYU8_2X8_BE (untested)\n");
> > + if (half_scale) {
> > + coma_mask |= COMA_RGB | COMA_BW;
> > + coma_set |= COMA_BYTE_SWAP | COMA_WORD_SWAP;
> > + } else {
> > + coma_mask |= COMA_RGB | COMA_BW | COMA_WORD_SWAP;
> > + coma_set |= COMA_BYTE_SWAP;
> > + }
> > + break;
>
> ...since there anyway will be an incremental patch . what does
> word-swapping have to do with scaling?...
A hardware (firmware) bug perhaps? All I can say is that I had found out it
worked like this for me before, and I've just ensured it still does.
> > + case V4L2_MBUS_FMT_SBGGR8_1X8:
> > + dev_dbg(&client->dev, "pixel format SBGGR8_1X8 (untested)\n");
> > + coma_mask |= COMA_BW | COMA_BYTE_SWAP | COMA_WORD_SWAP;
> > + coma_set |= COMA_RAW_RGB | COMA_RGB;
> > + break;
> > + case 0:
> > + break;
>
> 0 is defined as reserved... What for do you need it here?
This I have mentioned in my v2 -> v3 changes list:
> - add support for geometry only change to s_fmt,
I've tried to follow a pattern copied from mt9v022.c:
+ case 0:
+ /* No format change, only geometry */
+ break;
but now I see I've missed the comment unfortunately. Please specify if you
prefere to have it dropped or documented.
Thanks,
Janusz
> Thanks
> Guennadi
> ---
> Guennadi Liakhovetski, Ph.D.
> Freelance Open-Source Software Developer
> http://www.open-technology.de/
> --
> To unsubscribe from this list: send the line "unsubscribe linux-media" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
