* [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len()
@ 2014-03-21 21:04 Shuah Khan
2014-03-22 12:32 ` Frank Schäfer
0 siblings, 1 reply; 5+ messages in thread
From: Shuah Khan @ 2014-03-21 21:04 UTC (permalink / raw)
To: m.chehab; +Cc: Shuah Khan, linux-media, linux-kernel, shuahkhan
Change em28xx_scaler_set() to use em28xx_reg_len() to get register
lengths for EM28XX_R30_HSCALELOW and EM28XX_R32_VSCALELOW registers,
instead of hard-coding the length. Moved em28xx_reg_len() definition
for it to be visible to em28xx_scaler_set().
Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
---
drivers/media/usb/em28xx/em28xx-video.c | 29 ++++++++++++++++-------------
1 file changed, 16 insertions(+), 13 deletions(-)
diff --git a/drivers/media/usb/em28xx/em28xx-video.c b/drivers/media/usb/em28xx/em28xx-video.c
index 19af6b3..f8a91de 100644
--- a/drivers/media/usb/em28xx/em28xx-video.c
+++ b/drivers/media/usb/em28xx/em28xx-video.c
@@ -272,6 +272,18 @@ static void em28xx_capture_area_set(struct em28xx *dev, u8 hstart, u8 vstart,
}
}
+static int em28xx_reg_len(int reg)
+{
+ switch (reg) {
+ case EM28XX_R40_AC97LSB:
+ case EM28XX_R30_HSCALELOW:
+ case EM28XX_R32_VSCALELOW:
+ return 2;
+ default:
+ return 1;
+ }
+}
+
static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
{
u8 mode;
@@ -284,11 +296,13 @@ static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
buf[0] = h;
buf[1] = h >> 8;
- em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf, 2);
+ em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf,
+ em28xx_reg_len(EM28XX_R30_HSCALELOW));
buf[0] = v;
buf[1] = v >> 8;
- em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf, 2);
+ em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf,
+ em28xx_reg_len(EM28XX_R32_VSCALELOW));
/* it seems that both H and V scalers must be active
to work correctly */
mode = (h || v) ? 0x30 : 0x00;
@@ -1583,17 +1597,6 @@ static int vidioc_g_chip_info(struct file *file, void *priv,
return 0;
}
-static int em28xx_reg_len(int reg)
-{
- switch (reg) {
- case EM28XX_R40_AC97LSB:
- case EM28XX_R30_HSCALELOW:
- case EM28XX_R32_VSCALELOW:
- return 2;
- default:
- return 1;
- }
-}
static int vidioc_g_register(struct file *file, void *priv,
struct v4l2_dbg_register *reg)
--
1.7.10.4
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len()
2014-03-21 21:04 [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len() Shuah Khan
@ 2014-03-22 12:32 ` Frank Schäfer
2014-03-22 15:22 ` Shuah Khan
0 siblings, 1 reply; 5+ messages in thread
From: Frank Schäfer @ 2014-03-22 12:32 UTC (permalink / raw)
To: Shuah Khan, m.chehab; +Cc: linux-media, linux-kernel, shuahkhan
Am 21.03.2014 22:04, schrieb Shuah Khan:
> Change em28xx_scaler_set() to use em28xx_reg_len() to get register
> lengths for EM28XX_R30_HSCALELOW and EM28XX_R32_VSCALELOW registers,
> instead of hard-coding the length. Moved em28xx_reg_len() definition
> for it to be visible to em28xx_scaler_set().
>
> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
> ---
> drivers/media/usb/em28xx/em28xx-video.c | 29 ++++++++++++++++-------------
> 1 file changed, 16 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/media/usb/em28xx/em28xx-video.c b/drivers/media/usb/em28xx/em28xx-video.c
> index 19af6b3..f8a91de 100644
> --- a/drivers/media/usb/em28xx/em28xx-video.c
> +++ b/drivers/media/usb/em28xx/em28xx-video.c
> @@ -272,6 +272,18 @@ static void em28xx_capture_area_set(struct em28xx *dev, u8 hstart, u8 vstart,
> }
> }
>
> +static int em28xx_reg_len(int reg)
> +{
> + switch (reg) {
> + case EM28XX_R40_AC97LSB:
> + case EM28XX_R30_HSCALELOW:
> + case EM28XX_R32_VSCALELOW:
> + return 2;
> + default:
> + return 1;
> + }
> +}
> +
> static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
> {
> u8 mode;
> @@ -284,11 +296,13 @@ static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
>
> buf[0] = h;
> buf[1] = h >> 8;
> - em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf, 2);
> + em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf,
> + em28xx_reg_len(EM28XX_R30_HSCALELOW));
>
> buf[0] = v;
> buf[1] = v >> 8;
> - em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf, 2);
> + em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf,
> + em28xx_reg_len(EM28XX_R32_VSCALELOW));
Hmm... registers 0x30 and 0x32 are always 2 bytes long.
So this change would needlessly complicate the code.
Or do you know more than us, e.g. from looking into the datasheets ? ;-)
> /* it seems that both H and V scalers must be active
> to work correctly */
> mode = (h || v) ? 0x30 : 0x00;
> @@ -1583,17 +1597,6 @@ static int vidioc_g_chip_info(struct file *file, void *priv,
> return 0;
> }
>
> -static int em28xx_reg_len(int reg)
> -{
> - switch (reg) {
> - case EM28XX_R40_AC97LSB:
> - case EM28XX_R30_HSCALELOW:
> - case EM28XX_R32_VSCALELOW:
> - return 2;
> - default:
> - return 1;
> - }
> -}
>
> static int vidioc_g_register(struct file *file, void *priv,
> struct v4l2_dbg_register *reg)
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len()
2014-03-22 12:32 ` Frank Schäfer
@ 2014-03-22 15:22 ` Shuah Khan
2014-03-22 17:40 ` Frank Schäfer
0 siblings, 1 reply; 5+ messages in thread
From: Shuah Khan @ 2014-03-22 15:22 UTC (permalink / raw)
To: Frank Schäfer, m.chehab
Cc: linux-media, linux-kernel, shuahkhan, Shuah Khan
On 03/22/2014 06:32 AM, Frank Schäfer wrote:
>
> Am 21.03.2014 22:04, schrieb Shuah Khan:
>> Change em28xx_scaler_set() to use em28xx_reg_len() to get register
>> lengths for EM28XX_R30_HSCALELOW and EM28XX_R32_VSCALELOW registers,
>> instead of hard-coding the length. Moved em28xx_reg_len() definition
>> for it to be visible to em28xx_scaler_set().
>>
>> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
>> ---
>> drivers/media/usb/em28xx/em28xx-video.c | 29 ++++++++++++++++-------------
>> 1 file changed, 16 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/media/usb/em28xx/em28xx-video.c b/drivers/media/usb/em28xx/em28xx-video.c
>> index 19af6b3..f8a91de 100644
>> --- a/drivers/media/usb/em28xx/em28xx-video.c
>> +++ b/drivers/media/usb/em28xx/em28xx-video.c
>> @@ -272,6 +272,18 @@ static void em28xx_capture_area_set(struct em28xx *dev, u8 hstart, u8 vstart,
>> }
>> }
>>
>> +static int em28xx_reg_len(int reg)
>> +{
>> + switch (reg) {
>> + case EM28XX_R40_AC97LSB:
>> + case EM28XX_R30_HSCALELOW:
>> + case EM28XX_R32_VSCALELOW:
>> + return 2;
>> + default:
>> + return 1;
>> + }
>> +}
>> +
>> static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
>> {
>> u8 mode;
>> @@ -284,11 +296,13 @@ static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
>>
>> buf[0] = h;
>> buf[1] = h >> 8;
>> - em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf, 2);
>> + em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf,
>> + em28xx_reg_len(EM28XX_R30_HSCALELOW));
>>
>> buf[0] = v;
>> buf[1] = v >> 8;
>> - em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf, 2);
>> + em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf,
>> + em28xx_reg_len(EM28XX_R32_VSCALELOW));
> Hmm... registers 0x30 and 0x32 are always 2 bytes long.
> So this change would needlessly complicate the code.
>
The reason I made the change is that em28xx_reg_len() is handling these
two registers and I thought it would be good to make it consistent with
other writes to these registers and not hard-code the length.
I think it would help with maintenance later by avoiding hard-coding the
length and use the existing routine that returns the length for these
registers.
You are correct that it does add a function call in the code path. So if
you think the trade-off isn't worth it, I am not going to argue with it :)
-- Shuah
--
Shuah Khan
Senior Linux Kernel Developer - Open Source Group
Samsung Research America(Silicon Valley)
shuah.kh@samsung.com | (970) 672-0658
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len()
2014-03-22 15:22 ` Shuah Khan
@ 2014-03-22 17:40 ` Frank Schäfer
2014-03-22 22:03 ` Shuah Khan
0 siblings, 1 reply; 5+ messages in thread
From: Frank Schäfer @ 2014-03-22 17:40 UTC (permalink / raw)
To: shuah.kh, m.chehab; +Cc: linux-media, linux-kernel, shuahkhan
Am 22.03.2014 16:22, schrieb Shuah Khan:
> On 03/22/2014 06:32 AM, Frank Schäfer wrote:
>>
>> Am 21.03.2014 22:04, schrieb Shuah Khan:
>>> Change em28xx_scaler_set() to use em28xx_reg_len() to get register
>>> lengths for EM28XX_R30_HSCALELOW and EM28XX_R32_VSCALELOW registers,
>>> instead of hard-coding the length. Moved em28xx_reg_len() definition
>>> for it to be visible to em28xx_scaler_set().
>>>
>>> Signed-off-by: Shuah Khan <shuah.kh@samsung.com>
>>> ---
>>> drivers/media/usb/em28xx/em28xx-video.c | 29
>>> ++++++++++++++++-------------
>>> 1 file changed, 16 insertions(+), 13 deletions(-)
>>>
>>> diff --git a/drivers/media/usb/em28xx/em28xx-video.c
>>> b/drivers/media/usb/em28xx/em28xx-video.c
>>> index 19af6b3..f8a91de 100644
>>> --- a/drivers/media/usb/em28xx/em28xx-video.c
>>> +++ b/drivers/media/usb/em28xx/em28xx-video.c
>>> @@ -272,6 +272,18 @@ static void em28xx_capture_area_set(struct
>>> em28xx *dev, u8 hstart, u8 vstart,
>>> }
>>> }
>>>
>>> +static int em28xx_reg_len(int reg)
>>> +{
>>> + switch (reg) {
>>> + case EM28XX_R40_AC97LSB:
>>> + case EM28XX_R30_HSCALELOW:
>>> + case EM28XX_R32_VSCALELOW:
>>> + return 2;
>>> + default:
>>> + return 1;
>>> + }
>>> +}
>>> +
>>> static int em28xx_scaler_set(struct em28xx *dev, u16 h, u16 v)
>>> {
>>> u8 mode;
>>> @@ -284,11 +296,13 @@ static int em28xx_scaler_set(struct em28xx
>>> *dev, u16 h, u16 v)
>>>
>>> buf[0] = h;
>>> buf[1] = h >> 8;
>>> - em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf, 2);
>>> + em28xx_write_regs(dev, EM28XX_R30_HSCALELOW, (char *)buf,
>>> + em28xx_reg_len(EM28XX_R30_HSCALELOW));
>>>
>>> buf[0] = v;
>>> buf[1] = v >> 8;
>>> - em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf, 2);
>>> + em28xx_write_regs(dev, EM28XX_R32_VSCALELOW, (char *)buf,
>>> + em28xx_reg_len(EM28XX_R32_VSCALELOW));
>> Hmm... registers 0x30 and 0x32 are always 2 bytes long.
>> So this change would needlessly complicate the code.
>>
>
> The reason I made the change is that em28xx_reg_len() is handling
> these two registers and I thought it would be good to make it
> consistent with other writes to these registers and not hard-code the
> length.
Well, almost all register writes use a hardcoded length (most of the
time it's 1). ;-)
>
> I think it would help with maintenance later by avoiding hard-coding
> the length and use the existing routine that returns the length for
> these registers.
Yes, if the length would be variable, then this would be a good idea.
But AFAIK the length is always 2.
>
> You are correct that it does add a function call in the code path. So
> if you think the trade-off isn't worth it, I am not going to argue
> with it :)
I'm more concerned about the fact that readers of the code could think
that this is a write with a variable length, while the length is
actually always the same.
em28xx_reg_len() is a somewhat dirty hack for vidioc_[g,s]_register
debugging ioctls only.
Btw, what happens when you try to compile the code with this patch
applied and CONFIG_VIDEO_ADV_DEBUG disabled ? ;-)
Anyway, I very much appreciate your efforts to improve the em28xx
driver. :-)
There's much to do !
Regards,
Frank
>
> -- Shuah
>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len()
2014-03-22 17:40 ` Frank Schäfer
@ 2014-03-22 22:03 ` Shuah Khan
0 siblings, 0 replies; 5+ messages in thread
From: Shuah Khan @ 2014-03-22 22:03 UTC (permalink / raw)
To: Frank Schäfer, m.chehab; +Cc: linux-media, linux-kernel, shuahkhan
On 03/22/2014 11:40 AM, Frank Schäfer wrote:
>
> I'm more concerned about the fact that readers of the code could think
> that this is a write with a variable length, while the length is
> actually always the same.
Fair enough.
>
> em28xx_reg_len() is a somewhat dirty hack for vidioc_[g,s]_register
> debugging ioctls only.
I didn't realize that. In that case, it doesn't make sense to propagate
the change to non-debug code. This patch can be dropped. I thought
em28xx_reg_len() is good example of finding register length for these
registers.
> Btw, what happens when you try to compile the code with this patch
> applied and CONFIG_VIDEO_ADV_DEBUG disabled ? ;-)
CONFIG_VIDEO_ADV_DEBUG is disabled in my config.
-- Shuah
--
Shuah Khan
Senior Linux Kernel Developer - Open Source Group
Samsung Research America(Silicon Valley)
shuah.kh@samsung.com | (970) 672-0658
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2014-03-22 22:03 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-03-21 21:04 [PATCH] media: em28xx-video - change em28xx_scaler_set() to use em28xx_reg_len() Shuah Khan
2014-03-22 12:32 ` Frank Schäfer
2014-03-22 15:22 ` Shuah Khan
2014-03-22 17:40 ` Frank Schäfer
2014-03-22 22:03 ` Shuah Khan
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.