| Message ID | 1347429006-10879-1-git-send-email-r.sricharan@ti.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 9/12/2012 1:50 AM, R Sricharan wrote: > Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, > the feature is not getting selected. > Adding a string description in the Kconfig resolves this. > > But not sure if this is the right way to fix this. > Shouldn't you be selecting this in your 64-bit dma address capable arch instead? > Signed-off-by: R Sricharan <r.sricharan@ti.com> > Cc: Catalin Marinas <catalin.marinas@arm.com> > --- > arch/arm/mm/Kconfig | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig > index 101b968..90c5ea4 100644 > --- a/arch/arm/mm/Kconfig > +++ b/arch/arm/mm/Kconfig > @@ -601,7 +601,7 @@ config ARCH_PHYS_ADDR_T_64BIT > def_bool ARM_LPAE > > config ARCH_DMA_ADDR_T_64BIT > - bool > + bool "64bit capable DMA device" > > config ARM_THUMB > bool "Support Thumb user binaries" >
On Wed, Sep 12, 2012 at 01:46:18PM +0100, Cyril Chemparathy wrote: > On 9/12/2012 1:50 AM, R Sricharan wrote: > > Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, > > the feature is not getting selected. > > Adding a string description in the Kconfig resolves this. > > > > But not sure if this is the right way to fix this. > > Shouldn't you be selecting this in your 64-bit dma address capable arch > instead? Yes, that's how it was meant to be used.
On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: > > On 9/12/2012 1:50 AM, R Sricharan wrote: >> >> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, >> the feature is not getting selected. >> Adding a string description in the Kconfig resolves this. >> >> But not sure if this is the right way to fix this. >> > > Shouldn't you be selecting this in your 64-bit dma address capable arch > instead? > Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens in the arch which needs it. Since this option was made as simple bool, looks like because of some KCONFIG magic, it was not getting selected. Adding some help text does fix the issue and thats is the patch all about. Regards Santosh
On Wed, Sep 12, 2012 at 01:57:26PM +0100, Shilimkar, Santosh wrote: > On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: > > > > On 9/12/2012 1:50 AM, R Sricharan wrote: > >> > >> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, > >> the feature is not getting selected. > >> Adding a string description in the Kconfig resolves this. > >> > >> But not sure if this is the right way to fix this. > >> > > > > Shouldn't you be selecting this in your 64-bit dma address capable arch > > instead? > > > Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens > in the arch which needs it. Since this option was made as simple bool, > looks like because of some KCONFIG magic, it was not getting selected. > Adding some help text does fix the issue and thats is the patch all about. Are you sure you select this explicitly in Kconfig? It should not depend on a description. What I understood from the initial email was that this option was only added to the defconfig but ignored by the kernel as it's not a user selectable option (unless you add some text).
Hi, On Wed, Sep 12, 2012 at 6:27 PM, Shilimkar, Santosh <santosh.shilimkar@ti.com> wrote: > On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: >> >> On 9/12/2012 1:50 AM, R Sricharan wrote: >>> >>> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, >>> the feature is not getting selected. >>> Adding a string description in the Kconfig resolves this. >>> >>> But not sure if this is the right way to fix this. >>> >> >> Shouldn't you be selecting this in your 64-bit dma address capable arch >> instead? >> > Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens > in the arch which needs it. Since this option was made as simple bool, > looks like because of some KCONFIG magic, it was not getting selected. > Adding some help text does fix the issue and thats is the patch all about. > Correct. So even if you select this in the arch/defconfig, today this option is not getting enabled in the finally generated .config. So by adding this description text, this can really be selected from arch level. Thanks, Sricharan
On Wed, Sep 12, 2012 at 6:49 PM, Catalin Marinas <catalin.marinas@arm.com> wrote: > On Wed, Sep 12, 2012 at 01:57:26PM +0100, Shilimkar, Santosh wrote: >> On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: >> > >> > On 9/12/2012 1:50 AM, R Sricharan wrote: >> >> >> >> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, >> >> the feature is not getting selected. >> >> Adding a string description in the Kconfig resolves this. >> >> >> >> But not sure if this is the right way to fix this. >> >> >> > >> > Shouldn't you be selecting this in your 64-bit dma address capable arch >> > instead? >> > >> Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens >> in the arch which needs it. Since this option was made as simple bool, >> looks like because of some KCONFIG magic, it was not getting selected. >> Adding some help text does fix the issue and thats is the patch all about. > > Are you sure you select this explicitly in Kconfig? It should not depend > on a description. > You are right. > What I understood from the initial email was that this option was only > added to the defconfig but ignored by the kernel as it's not a user > selectable option (unless you add some text). > That's right. Explicit KCONFIG select should work. Regards Santosh
Hi, On Wed, Sep 12, 2012 at 6:55 PM, Shilimkar, Santosh <santosh.shilimkar@ti.com> wrote: > On Wed, Sep 12, 2012 at 6:49 PM, Catalin Marinas > <catalin.marinas@arm.com> wrote: >> On Wed, Sep 12, 2012 at 01:57:26PM +0100, Shilimkar, Santosh wrote: >>> On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: >>> > >>> > On 9/12/2012 1:50 AM, R Sricharan wrote: >>> >> >>> >> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, >>> >> the feature is not getting selected. >>> >> Adding a string description in the Kconfig resolves this. >>> >> >>> >> But not sure if this is the right way to fix this. >>> >> >>> > >>> > Shouldn't you be selecting this in your 64-bit dma address capable arch >>> > instead? >>> > >>> Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens >>> in the arch which needs it. Since this option was made as simple bool, >>> looks like because of some KCONFIG magic, it was not getting selected. >>> Adding some help text does fix the issue and thats is the patch all about. >> >> Are you sure you select this explicitly in Kconfig? It should not depend >> on a description. >> > You are right. > >> What I understood from the initial email was that this option was only >> added to the defconfig but ignored by the kernel as it's not a user >> selectable option (unless you add some text). >> > That's right. Explicit KCONFIG select should work. yes. It works when enabled in arch/KCONFIG. Thanks. This patch is not needed then. Thanks, Sricharan
On Wed, Sep 12, 2012 at 06:50:00PM +0530, R, Sricharan wrote: > Hi, > On Wed, Sep 12, 2012 at 6:27 PM, Shilimkar, Santosh > <santosh.shilimkar@ti.com> wrote: > > On Wed, Sep 12, 2012 at 6:16 PM, Cyril Chemparathy <cyril@ti.com> wrote: > >> > >> On 9/12/2012 1:50 AM, R Sricharan wrote: > >>> > >>> Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, > >>> the feature is not getting selected. > >>> Adding a string description in the Kconfig resolves this. > >>> > >>> But not sure if this is the right way to fix this. > >>> > >> > >> Shouldn't you be selecting this in your 64-bit dma address capable arch > >> instead? > >> > > Yes. The selection to enable CONFIG_DMA_ADDR_64BIT_T happens > > in the arch which needs it. Since this option was made as simple bool, > > looks like because of some KCONFIG magic, it was not getting selected. > > Adding some help text does fix the issue and thats is the patch all about. > > > Correct. So even if you select this in the arch/defconfig, That is correct behaviour. You need an explicit "select DMA_ADDR_64BIT_T" somewhere in a Kconfig file to make it happen. Editing defconfigs is not how we enable these options.
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig index 101b968..90c5ea4 100644 --- a/arch/arm/mm/Kconfig +++ b/arch/arm/mm/Kconfig @@ -601,7 +601,7 @@ config ARCH_PHYS_ADDR_T_64BIT def_bool ARM_LPAE config ARCH_DMA_ADDR_T_64BIT - bool + bool "64bit capable DMA device" config ARM_THUMB bool "Support Thumb user binaries"
Even if CONFIG_DMA_ADDR_64BIT_T is enabled by the defconfig, the feature is not getting selected. Adding a string description in the Kconfig resolves this. But not sure if this is the right way to fix this. Signed-off-by: R Sricharan <r.sricharan@ti.com> Cc: Catalin Marinas <catalin.marinas@arm.com> --- arch/arm/mm/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)