Message ID | 20170720155132.3250812-1-arnd@arndb.de |
---|---|
State | New |
Headers | show |
Series | [01/12] ARM: ixp4xx: fix ioport_unmap definition | expand |
Hi Arnd, On Thursday 20 July 2017 09:20 PM, Arnd Bergmann wrote: > Six ARM platforms still provide their own variant of the clk API > rather than using the generic COMMON_CLK API. This generally works, > but it causes some link errors with drivers using the clk_set_rate, > clk_get_parent, clk_set_parent or clk_round_rate functions when > a platform lacks those interfaces. > > OMAP1 implements the whole set already, but davinci, ep93xx, mmp, > sa1100 and w90x900 all lack at least one of the functions. > > This adds empty stub implementations for each of them, and I > don't even try to do something useful here but instead just > print a WARN() message to make it obvious what is going on > if they ever end up being called. > > The drivers that call these won't be used on these platforms > (otherwise we'd get a link error today), so the added code is harmless > bloat and will warn about accidental use. > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > arch/arm/mach-davinci/clock.c | 6 ++++++ > arch/arm/mach-ep93xx/clock.c | 21 +++++++++++++++++++++ > arch/arm/mach-mmp/clock.c | 22 ++++++++++++++++++++++ > arch/arm/mach-sa1100/clock.c | 29 +++++++++++++++++++++++++++++ > arch/arm/mach-w90x900/clock.c | 29 +++++++++++++++++++++++++++++ > 5 files changed, 107 insertions(+) > > diff --git a/arch/arm/mach-davinci/clock.c b/arch/arm/mach-davinci/clock.c > index f5dce9b4e617..89586779526c 100644 > --- a/arch/arm/mach-davinci/clock.c > +++ b/arch/arm/mach-davinci/clock.c > @@ -218,6 +218,12 @@ int clk_set_parent(struct clk *clk, struct clk *parent) > } > EXPORT_SYMBOL(clk_set_parent); > > +struct clk *clk_get_parent(struct clk *clk) > +{ > + return clk->parent; > +} > +EXPORT_SYMBOL(clk_get_parent); For mach-davinci change: Acked-by: Sekhar Nori <nsekhar@ti.com> > + > int clk_register(struct clk *clk) > { > if (clk == NULL || IS_ERR(clk)) > diff --git a/arch/arm/mach-ep93xx/clock.c b/arch/arm/mach-ep93xx/clock.c > index 39ef3b613912..c1c91fc6e178 100644 > --- a/arch/arm/mach-ep93xx/clock.c > +++ b/arch/arm/mach-ep93xx/clock.c > @@ -323,6 +323,27 @@ unsigned long clk_get_rate(struct clk *clk) > } > EXPORT_SYMBOL(clk_get_rate); > > +long clk_round_rate(struct clk *clk, unsigned long rate) > +{ > + WARN_ON(clk); > + return 0; > +} > +EXPORT_SYMBOL(clk_round_rate); Its probably better to WARN_ON_ONCE(). Also, since NULL clk is valid, it should be probably be WARN_ON_ONCE(1). Thanks, Sekhar
On Fri, Jul 21, 2017 at 10:05 AM, Sekhar Nori <nsekhar@ti.com> wrote: >> diff --git a/arch/arm/mach-davinci/clock.c b/arch/arm/mach-davinci/clock.c >> index f5dce9b4e617..89586779526c 100644 >> --- a/arch/arm/mach-davinci/clock.c >> +++ b/arch/arm/mach-davinci/clock.c >> @@ -218,6 +218,12 @@ int clk_set_parent(struct clk *clk, struct clk *parent) >> } >> EXPORT_SYMBOL(clk_set_parent); >> >> +struct clk *clk_get_parent(struct clk *clk) >> +{ >> + return clk->parent; >> +} >> +EXPORT_SYMBOL(clk_get_parent); > > For mach-davinci change: > > Acked-by: Sekhar Nori <nsekhar@ti.com> Thanks! >> diff --git a/arch/arm/mach-ep93xx/clock.c b/arch/arm/mach-ep93xx/clock.c >> index 39ef3b613912..c1c91fc6e178 100644 >> --- a/arch/arm/mach-ep93xx/clock.c >> +++ b/arch/arm/mach-ep93xx/clock.c >> @@ -323,6 +323,27 @@ unsigned long clk_get_rate(struct clk *clk) >> } >> EXPORT_SYMBOL(clk_get_rate); >> >> +long clk_round_rate(struct clk *clk, unsigned long rate) >> +{ >> + WARN_ON(clk); >> + return 0; >> +} >> +EXPORT_SYMBOL(clk_round_rate); > > Its probably better to WARN_ON_ONCE(). Also, since NULL clk is valid, it > should be probably be WARN_ON_ONCE(1). I thought about trying to make it as correct as possible, but then went for the opposite approach and instead just made the functions as small as possible to avoid bloat while having the warning in there along with an obviously wrong implementation. Apparently the davinci version unlike the others ended up being a correct one after all (let me know if you want to have a NULL pointer check there). I also remembered now that the mmp implementation is actually dead code (we always use the COMMON_CLK implementation for mmp these days). I think I'll just drop the mmp patch, and split out davinci, w90x900 and sa1100 into separate patches now, as there are only three of them left. Arnd
On Friday 21 July 2017 01:47 PM, Arnd Bergmann wrote: > On Fri, Jul 21, 2017 at 10:05 AM, Sekhar Nori <nsekhar@ti.com> wrote: > >>> diff --git a/arch/arm/mach-davinci/clock.c b/arch/arm/mach-davinci/clock.c >>> index f5dce9b4e617..89586779526c 100644 >>> --- a/arch/arm/mach-davinci/clock.c >>> +++ b/arch/arm/mach-davinci/clock.c >>> @@ -218,6 +218,12 @@ int clk_set_parent(struct clk *clk, struct clk *parent) >>> } >>> EXPORT_SYMBOL(clk_set_parent); >>> >>> +struct clk *clk_get_parent(struct clk *clk) >>> +{ >>> + return clk->parent; >>> +} >>> +EXPORT_SYMBOL(clk_get_parent); >> >> For mach-davinci change: >> >> Acked-by: Sekhar Nori <nsekhar@ti.com> > > Thanks! > >>> diff --git a/arch/arm/mach-ep93xx/clock.c b/arch/arm/mach-ep93xx/clock.c >>> index 39ef3b613912..c1c91fc6e178 100644 >>> --- a/arch/arm/mach-ep93xx/clock.c >>> +++ b/arch/arm/mach-ep93xx/clock.c >>> @@ -323,6 +323,27 @@ unsigned long clk_get_rate(struct clk *clk) >>> } >>> EXPORT_SYMBOL(clk_get_rate); >>> >>> +long clk_round_rate(struct clk *clk, unsigned long rate) >>> +{ >>> + WARN_ON(clk); >>> + return 0; >>> +} >>> +EXPORT_SYMBOL(clk_round_rate); >> >> Its probably better to WARN_ON_ONCE(). Also, since NULL clk is valid, it >> should be probably be WARN_ON_ONCE(1). > > I thought about trying to make it as correct as possible, but then went for > the opposite approach and instead just made the functions as small as > possible to avoid bloat while having the warning in there along with an > obviously wrong implementation. Apparently the davinci version unlike > the others ended up being a correct one after all (let me know if you > want to have a NULL pointer check there). Yeah, that would be better. Thanks! Regards, Sekhar
diff --git a/arch/arm/mach-davinci/clock.c b/arch/arm/mach-davinci/clock.c index f5dce9b4e617..89586779526c 100644 --- a/arch/arm/mach-davinci/clock.c +++ b/arch/arm/mach-davinci/clock.c @@ -218,6 +218,12 @@ int clk_set_parent(struct clk *clk, struct clk *parent) } EXPORT_SYMBOL(clk_set_parent); +struct clk *clk_get_parent(struct clk *clk) +{ + return clk->parent; +} +EXPORT_SYMBOL(clk_get_parent); + int clk_register(struct clk *clk) { if (clk == NULL || IS_ERR(clk)) diff --git a/arch/arm/mach-ep93xx/clock.c b/arch/arm/mach-ep93xx/clock.c index 39ef3b613912..c1c91fc6e178 100644 --- a/arch/arm/mach-ep93xx/clock.c +++ b/arch/arm/mach-ep93xx/clock.c @@ -323,6 +323,27 @@ unsigned long clk_get_rate(struct clk *clk) } EXPORT_SYMBOL(clk_get_rate); +long clk_round_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_round_rate); + +int clk_set_parent(struct clk *clk, struct clk *parent) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_parent); + +struct clk *clk_get_parent(struct clk *clk) +{ + WARN_ON(clk); + return NULL; +} +EXPORT_SYMBOL(clk_get_parent); + static int set_keytchclk_rate(struct clk *clk, unsigned long rate) { u32 val; diff --git a/arch/arm/mach-mmp/clock.c b/arch/arm/mach-mmp/clock.c index 28fe64c6e2f5..78513f74ae8d 100644 --- a/arch/arm/mach-mmp/clock.c +++ b/arch/arm/mach-mmp/clock.c @@ -106,3 +106,25 @@ int clk_set_rate(struct clk *clk, unsigned long rate) return ret; } EXPORT_SYMBOL(clk_set_rate); + +/* Dummy clk routine to build generic kernel parts that may be using them */ +long clk_round_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_round_rate); + +int clk_set_parent(struct clk *clk, struct clk *parent) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_parent); + +struct clk *clk_get_parent(struct clk *clk) +{ + WARN_ON(clk); + return NULL; +} +EXPORT_SYMBOL(clk_get_parent); diff --git a/arch/arm/mach-sa1100/clock.c b/arch/arm/mach-sa1100/clock.c index 0db46895c82a..a602d876c231 100644 --- a/arch/arm/mach-sa1100/clock.c +++ b/arch/arm/mach-sa1100/clock.c @@ -35,6 +35,35 @@ struct clk clk_##_name = { \ static DEFINE_SPINLOCK(clocks_lock); +/* Dummy clk routine to build generic kernel parts that may be using them */ +long clk_round_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_round_rate); + +int clk_set_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_rate); + +int clk_set_parent(struct clk *clk, struct clk *parent) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_parent); + +struct clk *clk_get_parent(struct clk *clk) +{ + WARN_ON(clk); + return NULL; +} +EXPORT_SYMBOL(clk_get_parent); + static void clk_gpio27_enable(struct clk *clk) { /* diff --git a/arch/arm/mach-w90x900/clock.c b/arch/arm/mach-w90x900/clock.c index ac6fd1a2cb59..3f93fac98d97 100644 --- a/arch/arm/mach-w90x900/clock.c +++ b/arch/arm/mach-w90x900/clock.c @@ -93,3 +93,32 @@ void nuc900_subclk_enable(struct clk *clk, int enable) __raw_writel(clken, W90X900_VA_CLKPWR + SUBCLK); } + +/* dummy functions, should not be called */ +long clk_round_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_round_rate); + +int clk_set_rate(struct clk *clk, unsigned long rate) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_rate); + +int clk_set_parent(struct clk *clk, struct clk *parent) +{ + WARN_ON(clk); + return 0; +} +EXPORT_SYMBOL(clk_set_parent); + +struct clk *clk_get_parent(struct clk *clk) +{ + WARN_ON(clk); + return NULL; +} +EXPORT_SYMBOL(clk_get_parent);
Six ARM platforms still provide their own variant of the clk API rather than using the generic COMMON_CLK API. This generally works, but it causes some link errors with drivers using the clk_set_rate, clk_get_parent, clk_set_parent or clk_round_rate functions when a platform lacks those interfaces. OMAP1 implements the whole set already, but davinci, ep93xx, mmp, sa1100 and w90x900 all lack at least one of the functions. This adds empty stub implementations for each of them, and I don't even try to do something useful here but instead just print a WARN() message to make it obvious what is going on if they ever end up being called. The drivers that call these won't be used on these platforms (otherwise we'd get a link error today), so the added code is harmless bloat and will warn about accidental use. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- arch/arm/mach-davinci/clock.c | 6 ++++++ arch/arm/mach-ep93xx/clock.c | 21 +++++++++++++++++++++ arch/arm/mach-mmp/clock.c | 22 ++++++++++++++++++++++ arch/arm/mach-sa1100/clock.c | 29 +++++++++++++++++++++++++++++ arch/arm/mach-w90x900/clock.c | 29 +++++++++++++++++++++++++++++ 5 files changed, 107 insertions(+) -- 2.9.0