| Message ID | 20220127105652.1063624-1-miquel.raynal@bootlin.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mtd/memory: omap: Fix build dependencies | expand |
On 11:56-20220127, Miquel Raynal wrote: > The Omap2 raw NAND controller (MTD_NAND_OMAP2) selects the Omap memory > controller (OMAP_GPMC) for the ARCH_K3 platforms and for compile > testing. > > The Omap memory controller (OMAP_GPMC) depends on OF_ADDRESS because the > driver uses of_address_to_resource(). > > This may produce the following warning: > WARNING: unmet direct dependencies detected for OMAP_GPMC > Depends on [n]: > - MEMORY [=y] && OF_ADDRESS [=n] > Selected by [m]: > - MTD_NAND_OMAP2 [=m] && MTD [=m] && MTD_RAW_NAND [=m] && > (ARCH_OMAP2PLUS || ARCH_KEYSTONE || ARCH_K3 || COMPILE_TEST [=y]) && > HAS_IOMEM [=y] > > As the of_address_to_resource() helper has a dummy implementation > returning an error when OF_ADDRESS is not selected, we do not need to > select OF_ADDRESS from the OMAP_GPMC Kconfig menu for compile testing > purposes. Nor we actually do for K3 platforms, but in order to not break > these platforms we can just select OF_ADDRESS from ARCH_K3 directly. > > Fixes: dbcb124acebd ("mtd: rawnand: omap2: Select GPMC device driver for ARCH_K3") > Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > Suggested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > --- > arch/arm64/Kconfig.platforms | 1 + > drivers/memory/Kconfig | 1 - > 2 files changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms > index 7d5d58800170..1a0d4b575b95 100644 > --- a/arch/arm64/Kconfig.platforms > +++ b/arch/arm64/Kconfig.platforms > @@ -123,6 +123,7 @@ config ARCH_K3 > select TI_SCI_INTR_IRQCHIP > select TI_SCI_INTA_IRQCHIP > select TI_K3_SOCINFO > + select OF_ADDRESS I understand the build failure, but, I don't see any other SoC selecting OF_ADDRESS because a driver has a problem? > help > This enables support for Texas Instruments' K3 multicore SoC > architecture. > diff --git a/drivers/memory/Kconfig b/drivers/memory/Kconfig > index 30bff6cb1b8d..241f0925cbf1 100644 > --- a/drivers/memory/Kconfig > +++ b/drivers/memory/Kconfig > @@ -104,7 +104,6 @@ config TI_EMIF > > config OMAP_GPMC > bool "Texas Instruments OMAP SoC GPMC driver" if COMPILE_TEST > - depends on OF_ADDRESS Thoughts Roger? > select GPIOLIB > help > This driver is for the General Purpose Memory Controller (GPMC) > -- > 2.27.0 > > > ______________________________________________________ > Linux MTD discussion mailing list > http://lists.infradead.org/mailman/listinfo/linux-mtd/
On 09/02/2022 14:57, Nishanth Menon wrote: > On 11:56-20220127, Miquel Raynal wrote: >> The Omap2 raw NAND controller (MTD_NAND_OMAP2) selects the Omap memory >> controller (OMAP_GPMC) for the ARCH_K3 platforms and for compile >> testing. >> >> The Omap memory controller (OMAP_GPMC) depends on OF_ADDRESS because the >> driver uses of_address_to_resource(). >> >> This may produce the following warning: >> WARNING: unmet direct dependencies detected for OMAP_GPMC >> Depends on [n]: >> - MEMORY [=y] && OF_ADDRESS [=n] >> Selected by [m]: >> - MTD_NAND_OMAP2 [=m] && MTD [=m] && MTD_RAW_NAND [=m] && >> (ARCH_OMAP2PLUS || ARCH_KEYSTONE || ARCH_K3 || COMPILE_TEST [=y]) && >> HAS_IOMEM [=y] >> >> As the of_address_to_resource() helper has a dummy implementation >> returning an error when OF_ADDRESS is not selected, we do not need to >> select OF_ADDRESS from the OMAP_GPMC Kconfig menu for compile testing >> purposes. Nor we actually do for K3 platforms, but in order to not break >> these platforms we can just select OF_ADDRESS from ARCH_K3 directly. >> >> Fixes: dbcb124acebd ("mtd: rawnand: omap2: Select GPMC device driver for ARCH_K3") >> Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> >> Suggested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> >> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> >> --- >> arch/arm64/Kconfig.platforms | 1 + >> drivers/memory/Kconfig | 1 - >> 2 files changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms >> index 7d5d58800170..1a0d4b575b95 100644 >> --- a/arch/arm64/Kconfig.platforms >> +++ b/arch/arm64/Kconfig.platforms >> @@ -123,6 +123,7 @@ config ARCH_K3 >> select TI_SCI_INTR_IRQCHIP >> select TI_SCI_INTA_IRQCHIP >> select TI_K3_SOCINFO >> + select OF_ADDRESS > > I understand the build failure, but, I don't see any other SoC selecting > OF_ADDRESS because a driver has a problem? > The ARM platforms do not need to select OF_ADDRESS because it is always there. The error looks like a compile test, so is this patch really a fix for reported error? Best regards, Krzysztof
Hi, On 09/02/2022 15:57, Nishanth Menon wrote: > On 11:56-20220127, Miquel Raynal wrote: >> The Omap2 raw NAND controller (MTD_NAND_OMAP2) selects the Omap memory >> controller (OMAP_GPMC) for the ARCH_K3 platforms and for compile >> testing. >> >> The Omap memory controller (OMAP_GPMC) depends on OF_ADDRESS because the >> driver uses of_address_to_resource(). >> >> This may produce the following warning: >> WARNING: unmet direct dependencies detected for OMAP_GPMC >> Depends on [n]: >> - MEMORY [=y] && OF_ADDRESS [=n] >> Selected by [m]: >> - MTD_NAND_OMAP2 [=m] && MTD [=m] && MTD_RAW_NAND [=m] && >> (ARCH_OMAP2PLUS || ARCH_KEYSTONE || ARCH_K3 || COMPILE_TEST [=y]) && >> HAS_IOMEM [=y] >> >> As the of_address_to_resource() helper has a dummy implementation >> returning an error when OF_ADDRESS is not selected, we do not need to >> select OF_ADDRESS from the OMAP_GPMC Kconfig menu for compile testing >> purposes. Nor we actually do for K3 platforms, but in order to not break >> these platforms we can just select OF_ADDRESS from ARCH_K3 directly. >> >> Fixes: dbcb124acebd ("mtd: rawnand: omap2: Select GPMC device driver for ARCH_K3") >> Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> >> Suggested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> >> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> >> --- >> arch/arm64/Kconfig.platforms | 1 + >> drivers/memory/Kconfig | 1 - >> 2 files changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms >> index 7d5d58800170..1a0d4b575b95 100644 >> --- a/arch/arm64/Kconfig.platforms >> +++ b/arch/arm64/Kconfig.platforms >> @@ -123,6 +123,7 @@ config ARCH_K3 >> select TI_SCI_INTR_IRQCHIP >> select TI_SCI_INTA_IRQCHIP >> select TI_K3_SOCINFO >> + select OF_ADDRESS > > I understand the build failure, but, I don't see any other SoC selecting > OF_ADDRESS because a driver has a problem? > > >> help >> This enables support for Texas Instruments' K3 multicore SoC >> architecture. >> diff --git a/drivers/memory/Kconfig b/drivers/memory/Kconfig >> index 30bff6cb1b8d..241f0925cbf1 100644 >> --- a/drivers/memory/Kconfig >> +++ b/drivers/memory/Kconfig >> @@ -104,7 +104,6 @@ config TI_EMIF >> >> config OMAP_GPMC >> bool "Texas Instruments OMAP SoC GPMC driver" if COMPILE_TEST >> - depends on OF_ADDRESS > > Thoughts Roger? OMAP_GPMC is not visible entry so someone has to select it. We first tried selecting it in ARCH_K3 but that was not accepted for fair reason. Then we tried to select it from MTD_NAND_OMAP2 and it caused the issue this patch is trying to fix. Only other option left is to make OMAP_GPMC a visible entry and just mark MTD_NAND_OMAP2 to depend on it? The relevant SoC defconfig files can enable OMAP_GPMC and MTD_NAND_OMAP2. Does this sound ok? >> select GPIOLIB >> help >> This driver is for the General Purpose Memory Controller (GPMC) >> -- >> 2.27.0 >> >> >> ______________________________________________________ >> Linux MTD discussion mailing list >> http://lists.infradead.org/mailman/listinfo/linux-mtd/ > cheers, -roger
On 16:39-20220209, Roger Quadros wrote: > >> config OMAP_GPMC > >> bool "Texas Instruments OMAP SoC GPMC driver" if COMPILE_TEST > >> - depends on OF_ADDRESS > > > > Thoughts Roger? > > OMAP_GPMC is not visible entry so someone has to select it. > We first tried selecting it in ARCH_K3 but that was not accepted for fair reason. > Then we tried to select it from MTD_NAND_OMAP2 and it caused the issue this patch is trying to fix. > > Only other option left is to make OMAP_GPMC a visible entry and just mark MTD_NAND_OMAP2 to depend on it? > The relevant SoC defconfig files can enable OMAP_GPMC and MTD_NAND_OMAP2. > > Does this sound ok? Given that of_address.h is handling OF_ADDRESS smartly, and the fact that there are quite the many drivers that are actually OF only, I am not entirely sure why the dependency is needed in Kconfig, but that is just me.. I'd suggest to see with the randconfig that Randy provided in [1] [1] https://lore.kernel.org/all/b18fc937-9cc2-bb7b-fb58-3ba2555371c7@infradead.org/
diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms index 7d5d58800170..1a0d4b575b95 100644 --- a/arch/arm64/Kconfig.platforms +++ b/arch/arm64/Kconfig.platforms @@ -123,6 +123,7 @@ config ARCH_K3 select TI_SCI_INTR_IRQCHIP select TI_SCI_INTA_IRQCHIP select TI_K3_SOCINFO + select OF_ADDRESS help This enables support for Texas Instruments' K3 multicore SoC architecture. diff --git a/drivers/memory/Kconfig b/drivers/memory/Kconfig index 30bff6cb1b8d..241f0925cbf1 100644 --- a/drivers/memory/Kconfig +++ b/drivers/memory/Kconfig @@ -104,7 +104,6 @@ config TI_EMIF config OMAP_GPMC bool "Texas Instruments OMAP SoC GPMC driver" if COMPILE_TEST - depends on OF_ADDRESS select GPIOLIB help This driver is for the General Purpose Memory Controller (GPMC)
The Omap2 raw NAND controller (MTD_NAND_OMAP2) selects the Omap memory controller (OMAP_GPMC) for the ARCH_K3 platforms and for compile testing. The Omap memory controller (OMAP_GPMC) depends on OF_ADDRESS because the driver uses of_address_to_resource(). This may produce the following warning: WARNING: unmet direct dependencies detected for OMAP_GPMC Depends on [n]: - MEMORY [=y] && OF_ADDRESS [=n] Selected by [m]: - MTD_NAND_OMAP2 [=m] && MTD [=m] && MTD_RAW_NAND [=m] && (ARCH_OMAP2PLUS || ARCH_KEYSTONE || ARCH_K3 || COMPILE_TEST [=y]) && HAS_IOMEM [=y] As the of_address_to_resource() helper has a dummy implementation returning an error when OF_ADDRESS is not selected, we do not need to select OF_ADDRESS from the OMAP_GPMC Kconfig menu for compile testing purposes. Nor we actually do for K3 platforms, but in order to not break these platforms we can just select OF_ADDRESS from ARCH_K3 directly. Fixes: dbcb124acebd ("mtd: rawnand: omap2: Select GPMC device driver for ARCH_K3") Reported-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Suggested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> --- arch/arm64/Kconfig.platforms | 1 + drivers/memory/Kconfig | 1 - 2 files changed, 1 insertion(+), 1 deletion(-)