| Message ID | 1412199569-27958-1-git-send-email-robert.jarzmik@free.fr (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Quoting Robert Jarzmik (2014-10-01 14:39:29) > Fix the building of pxa clock drivers so that the files are actually > compiled if and only if COMMON_CLK was selected by the architecture. > > This prevents conflicts with mach-pxa clock legacy implementation. > > Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Applied. Thanks, Mike > --- > drivers/clk/Kconfig | 5 +++++ > drivers/clk/Makefile | 2 +- > 2 files changed, 6 insertions(+), 1 deletion(-) > > diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig > index cfd3af7..b110f33 100644 > --- a/drivers/clk/Kconfig > +++ b/drivers/clk/Kconfig > @@ -109,6 +109,11 @@ config COMMON_CLK_PALMAS > This driver supports TI Palmas devices 32KHz output KG and KG_AUDIO > using common clock framework. > > +config COMMON_CLK_PXA > + def_bool COMMON_CLK && ARCH_PXA > + ---help--- > + Sypport for the Marvell PXA SoC. > + > source "drivers/clk/qcom/Kconfig" > > endmenu > diff --git a/drivers/clk/Makefile b/drivers/clk/Makefile > index 7891e4e..b3a1da0 100644 > --- a/drivers/clk/Makefile > +++ b/drivers/clk/Makefile > @@ -48,7 +48,7 @@ obj-$(CONFIG_ARCH_MMP) += mmp/ > endif > obj-$(CONFIG_PLAT_ORION) += mvebu/ > obj-$(CONFIG_ARCH_MXS) += mxs/ > -obj-$(CONFIG_ARCH_PXA) += pxa/ > +obj-$(CONFIG_COMMON_CLK_PXA) += pxa/ > obj-$(CONFIG_COMMON_CLK_QCOM) += qcom/ > obj-$(CONFIG_ARCH_ROCKCHIP) += rockchip/ > obj-$(CONFIG_COMMON_CLK_SAMSUNG) += samsung/ > -- > 2.1.0 >
On Sunday 05 October 2014 13:43:58 Mike Turquette wrote: > Quoting Robert Jarzmik (2014-10-01 14:39:29) > > Fix the building of pxa clock drivers so that the files are actually > > compiled if and only if COMMON_CLK was selected by the architecture. > > > > This prevents conflicts with mach-pxa clock legacy implementation. > > > > Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> > I seem to still hit a build error with this applied in e.g. build/palmz72_defconfig, which enables both pxa25x and pxa27x, so I get both the pxa250 private clock code and the pxa270 common clock code, resulting in this error message: drivers/built-in.o:(___ksymtab_gpl+clk_enable+0x0): multiple definition of `__ksymtab_clk_enable' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_enable+0x0): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_get_rate+0x0): multiple definition of `__ksymtab_clk_get_rate' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_get_rate+0x0): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_round_rate+0x0): multiple definition of `__ksymtab_clk_round_rate' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_round_rate+0x0): first defined here drivers/built-in.o: In function `clk_set_rate': :(.text+0x86930): multiple definition of `clk_set_rate' arch/arm/mach-pxa/built-in.o::(.text+0x103c): first defined here drivers/built-in.o: In function `clk_set_parent': :(.text+0x869c4): multiple definition of `clk_set_parent' arch/arm/mach-pxa/built-in.o::(.text+0x127c): first defined here drivers/built-in.o: In function `clk_round_rate': :(.text+0x85c18): multiple definition of `clk_round_rate' arch/arm/mach-pxa/built-in.o::(.text+0x1248): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_set_parent+0x0): multiple definition of `__ksymtab_clk_set_parent' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_set_parent+0x0): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_disable+0x0): multiple definition of `__ksymtab_clk_disable' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_disable+0x0): first defined here drivers/built-in.o: In function `clk_get_rate': :(.text+0x85c64): multiple definition of `clk_get_rate' arch/arm/mach-pxa/built-in.o::(.text+0x1014): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_set_rate+0x0): multiple definition of `__ksymtab_clk_set_rate' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_set_rate+0x0): first defined here drivers/built-in.o: In function `clk_get_parent': :(.text+0x85c44): multiple definition of `clk_get_parent' arch/arm/mach-pxa/built-in.o::(.text+0x12b0): first defined here drivers/built-in.o: In function `clk_disable': :(.text+0x85b08): multiple definition of `clk_disable' arch/arm/mach-pxa/built-in.o::(.text+0x118c): first defined here drivers/built-in.o:(___ksymtab_gpl+clk_get_parent+0x0): multiple definition of `__ksymtab_clk_get_parent' arch/arm/mach-pxa/built-in.o:(___ksymtab+clk_get_parent+0x0): first defined here drivers/built-in.o: In function `clk_enable': :(.text+0x85808): multiple definition of `clk_enable' arch/arm/mach-pxa/built-in.o::(.text+0x10d8): first defined here Arnd
Arnd Bergmann <arnd@arndb.de> writes: > On Sunday 05 October 2014 13:43:58 Mike Turquette wrote: >> Quoting Robert Jarzmik (2014-10-01 14:39:29) >> > Fix the building of pxa clock drivers so that the files are actually >> > compiled if and only if COMMON_CLK was selected by the architecture. >> > >> > This prevents conflicts with mach-pxa clock legacy implementation. >> > >> > Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> >> > > I seem to still hit a build error with this applied in e.g. > build/palmz72_defconfig, which enables both pxa25x and pxa27x, > so I get both the pxa250 private clock code and the pxa270 > common clock code, resulting in this error message: Do you have the patch "Revert "arm: pxa: Transition pxa27x to clk framework" in your tree ? Normally the fix was made of : - the fixup patch "clk: pxa clocks build system fix" - the revert queued by Mike "arm: pxa: Transition pxa27x to clk framework" Cheers. -- Robert
On Monday 06 October 2014 23:25:13 Robert Jarzmik wrote: > Arnd Bergmann <arnd@arndb.de> writes: > > > On Sunday 05 October 2014 13:43:58 Mike Turquette wrote: > >> Quoting Robert Jarzmik (2014-10-01 14:39:29) > >> > Fix the building of pxa clock drivers so that the files are actually > >> > compiled if and only if COMMON_CLK was selected by the architecture. > >> > > >> > This prevents conflicts with mach-pxa clock legacy implementation. > >> > > >> > Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> > >> > > > > I seem to still hit a build error with this applied in e.g. > > build/palmz72_defconfig, which enables both pxa25x and pxa27x, > > so I get both the pxa250 private clock code and the pxa270 > > common clock code, resulting in this error message: > Do you have the patch "Revert "arm: pxa: Transition pxa27x to clk framework" in > your tree ? > > Normally the fix was made of : > - the fixup patch "clk: pxa clocks build system fix" > - the revert queued by Mike "arm: pxa: Transition pxa27x to clk framework" No, I was missing that one, sorry for the noise. It works now after I've applied it. I was using the latest -next kernel as a base and assumed that it had all the latest fixes from Mike's tree. Arnd
Quoting Arnd Bergmann (2014-10-06 14:36:43) > On Monday 06 October 2014 23:25:13 Robert Jarzmik wrote: > > Arnd Bergmann <arnd@arndb.de> writes: > > > > > On Sunday 05 October 2014 13:43:58 Mike Turquette wrote: > > >> Quoting Robert Jarzmik (2014-10-01 14:39:29) > > >> > Fix the building of pxa clock drivers so that the files are actually > > >> > compiled if and only if COMMON_CLK was selected by the architecture. > > >> > > > >> > This prevents conflicts with mach-pxa clock legacy implementation. > > >> > > > >> > Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> > > >> > > > > > > I seem to still hit a build error with this applied in e.g. > > > build/palmz72_defconfig, which enables both pxa25x and pxa27x, > > > so I get both the pxa250 private clock code and the pxa270 > > > common clock code, resulting in this error message: > > Do you have the patch "Revert "arm: pxa: Transition pxa27x to clk framework" in > > your tree ? > > > > Normally the fix was made of : > > - the fixup patch "clk: pxa clocks build system fix" > > - the revert queued by Mike "arm: pxa: Transition pxa27x to clk framework" > > No, I was missing that one, sorry for the noise. It works now after > I've applied it. I was using the latest -next kernel as a base and assumed > that it had all the latest fixes from Mike's tree. I pushed the changes to my mirror over the weekend. I guess -next hasn't picked it up. I'm waiting for a couple of -next cycles this week before I push my tag to Linus. Regards, Mike > > Arnd
On Monday 06 October 2014 15:15:16 Mike Turquette wrote: > > I pushed the changes to my mirror over the weekend. I guess -next hasn't > picked it up. I'm waiting for a couple of -next cycles this week before > I push my tag to Linus. I think there was no linux-next yesterday. Today's -next just arrived and it has the fix. Arnd
diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig index cfd3af7..b110f33 100644 --- a/drivers/clk/Kconfig +++ b/drivers/clk/Kconfig @@ -109,6 +109,11 @@ config COMMON_CLK_PALMAS This driver supports TI Palmas devices 32KHz output KG and KG_AUDIO using common clock framework. +config COMMON_CLK_PXA + def_bool COMMON_CLK && ARCH_PXA + ---help--- + Sypport for the Marvell PXA SoC. + source "drivers/clk/qcom/Kconfig" endmenu diff --git a/drivers/clk/Makefile b/drivers/clk/Makefile index 7891e4e..b3a1da0 100644 --- a/drivers/clk/Makefile +++ b/drivers/clk/Makefile @@ -48,7 +48,7 @@ obj-$(CONFIG_ARCH_MMP) += mmp/ endif obj-$(CONFIG_PLAT_ORION) += mvebu/ obj-$(CONFIG_ARCH_MXS) += mxs/ -obj-$(CONFIG_ARCH_PXA) += pxa/ +obj-$(CONFIG_COMMON_CLK_PXA) += pxa/ obj-$(CONFIG_COMMON_CLK_QCOM) += qcom/ obj-$(CONFIG_ARCH_ROCKCHIP) += rockchip/ obj-$(CONFIG_COMMON_CLK_SAMSUNG) += samsung/
Fix the building of pxa clock drivers so that the files are actually compiled if and only if COMMON_CLK was selected by the architecture. This prevents conflicts with mach-pxa clock legacy implementation. Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> --- drivers/clk/Kconfig | 5 +++++ drivers/clk/Makefile | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-)