On 04/23/2015 11:15 PM, Richard Fitzgerald wrote:
On Wed, Apr 22, 2015 at 07:20:09PM +0900, Chanwoo Choi wrote:
On 04/22/2015 06:19 PM, Richard Fitzgerald wrote:
On Wed, Apr 22, 2015 at 02:53:42PM +0900, Chanwoo Choi wrote:
Hi Richard,
@@ -1176,6 +1182,11 @@ static int arizona_extcon_probe(struct platform_device *pdev) break; } break;
- case WM8998:
- case WM1814:
info->micd_clamp = true;info->hpdet_ip = 2;What is meaning of '2'? I prefer to use the definition for '2'.
'2' is the version number of the hpdet ip block in silicon. We're already using it as a raw number '0', '1' or '2' all over extcon-arizona.c so changing it here would mean making other patches to the file that aren't really part of adding WM8998 support, so I'd prefer not to change that as a side-effect of adding WM8998.
I think that just you can define following definitions and use HPDET_IP_VER_V2 instead of '2'.
#define HPDET_IP_VER_V0 0 #define HPDET_IP_VER_V1 1 #define HPDET_IP_VER_V2 2
Can we deal with that as a separate patch from this series? Like I said, the code already uses '0' '1' and '2' for the existing codecs so making a change to use #define means patching the code for the other codecs. That is not part of adding WM8998 support and I don't like patches that make unexpected extra side-effect changes that are not relevant to the actual functionality being added by the patch. It's specially annoying when cherry-picking or reverting those patches if they included some extra code change.
If we can get this series submitted I can look at making a later patch to improve readbility, but since this really is just a version number I think it would be enough to rename the variable to hpdet_ip_version rather than effectively doing #define TWO 2
OK. I agree that rename the variable name (hpdet_ip_version) or add the definitions for version on later patch.
Acked-by: Chanwoo Choi cw00.choi@samsung.com
Thanks, Chanwoo Choi