On Wed, 2008-07-30 at 18:23 -0700, Luis R. Rodriguez wrote:

On Wed, Jul 30, 2008 at 6:16 PM, Pavel Roskin [off-list ref] wrote:

quoted

On Mon, 2008-07-28 at 22:47 -0700, lrodriguez@atheros.com wrote:

quoted

From: Sujith Manoharan <redacted>

diff --git a/drivers/net/wireless/ath9k/ath9k.h b/drivers/net/wireless/ath9k/ath9k.h
index 6fa1eb6..c3294c0 100644
--- a/drivers/net/wireless/ath9k/ath9k.h
+++ b/drivers/net/wireless/ath9k/ath9k.h

@@ -622,17 +622,15 @@ struct hal_country_entry {

 #define SM(_v, _f)  (((_v) << _f##_S) & _f)
 #define MS(_v, _f)  (((_v) & _f) >> _f##_S)
-#define OS_REG_RMW(_a, _r, _set, _clr)    \
+#define REG_RMW(_a, _r, _set, _clr)    \
      REG_WRITE(_a, _r, (REG_READ(_a, _r) & ~(_clr)) | (_set))

I think these are pretty good candidates for inline functions.  I'm not
sure removing "OS_" is a useful intermediate step.  It's better to keep
such massive changes to the minimum to avoid potential conflicts with
other patches.

This was done in response to Johannes comments. This can be moved to
inline though.

It would be really nice.  Explicit return types tend to unmask some
issues in the calling code, if there are any.

I don't care if you are going to change the macros/functions themselves
in many steps, but let's avoid the need to rename all callers again.

-- 
Regards,
Pavel Roskin