| Message ID | 1355911460-19926-1-git-send-email-sachin.kamat@linaro.org |
|---|---|
| State | Accepted |
| Headers | show |
On Wed, 19 Dec 2012 15:34:20 +0530 Sachin Kamat <sachin.kamat@linaro.org> wrote: > This eliminates having an #ifdef returning NULL for the case > when OF is disabled. > > ... > > - .of_match_table = of_cmos_match, > + .of_match_table = of_match_ptr(of_cmos_match), Thanks. I suppose these guys might need the same treatment: akpm:/usr/src/25> grep of_match_table drivers/rtc/*.c|grep -v of_match_ptr drivers/rtc/rtc-mpc5121.c: .of_match_table = mpc5121_rtc_match, drivers/rtc/rtc-mv.c:static struct of_device_id rtc_mv_of_match_table[] = { drivers/rtc/rtc-sa1100.c: .of_match_table = sa1100_rtc_dt_ids, drivers/rtc/rtc-snvs.c: .of_match_table = snvs_dt_ids, drivers/rtc/rtc-stmp3xxx.c: .of_match_table = rtc_dt_ids, drivers/rtc/rtc-twl.c: .of_match_table = twl_rtc_of_match, It would be nice to have a single patch which addresses all of drivers/rtc/. Hint :) It's unneeded if the driver depends on OF in Kconfig. But even there, using of_match_ptr() is OK and keeps things nice and consistent.
On 20 December 2012 03:57, Andrew Morton <akpm@linux-foundation.org> wrote: > On Wed, 19 Dec 2012 15:34:20 +0530 > Sachin Kamat <sachin.kamat@linaro.org> wrote: > >> This eliminates having an #ifdef returning NULL for the case >> when OF is disabled. >> >> ... >> >> - .of_match_table = of_cmos_match, >> + .of_match_table = of_match_ptr(of_cmos_match), > > Thanks. I suppose these guys might need the same treatment: > > akpm:/usr/src/25> grep of_match_table drivers/rtc/*.c|grep -v of_match_ptr > drivers/rtc/rtc-mpc5121.c: .of_match_table = mpc5121_rtc_match, > drivers/rtc/rtc-mv.c:static struct of_device_id rtc_mv_of_match_table[] = { > drivers/rtc/rtc-sa1100.c: .of_match_table = sa1100_rtc_dt_ids, > drivers/rtc/rtc-snvs.c: .of_match_table = snvs_dt_ids, > drivers/rtc/rtc-stmp3xxx.c: .of_match_table = rtc_dt_ids, > drivers/rtc/rtc-twl.c: .of_match_table = twl_rtc_of_match, > > It would be nice to have a single patch which addresses all of > drivers/rtc/. Hint :) Sure, i will do it. > > It's unneeded if the driver depends on OF in Kconfig. But even there, > using of_match_ptr() is OK and keeps things nice and consistent. However, I get a lot of compilation errors when I select rtc-twl driver both in Linus' tree as well as linux-next (the errors are mostly in the files drivers/mfd/twl-core.c, drivers/mfd/twl4030-madc.c and drivers/mfd/twl4030-power.c). Hence I have not done any changes in drivers/rtc/rtc-twl.c for now. I will post the patch separately after this mail.
diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c index 4267789..2e84a64 100644 --- a/drivers/rtc/rtc-cmos.c +++ b/drivers/rtc/rtc-cmos.c @@ -1101,7 +1101,6 @@ static __init void cmos_of_init(struct platform_device *pdev) } #else static inline void cmos_of_init(struct platform_device *pdev) {} -#define of_cmos_match NULL #endif /*----------------------------------------------------------------*/ @@ -1143,7 +1142,7 @@ static struct platform_driver cmos_platform_driver = { #ifdef CONFIG_PM .pm = &cmos_pm_ops, #endif - .of_match_table = of_cmos_match, + .of_match_table = of_match_ptr(of_cmos_match), } };
This eliminates having an #ifdef returning NULL for the case when OF is disabled. Cc: David Brownell <david-b@pacbell.net> Cc: Alessandro Zummo <a.zummo@towertech.it> Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org> --- Compile tested using linux-next tree. --- drivers/rtc/rtc-cmos.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-)