| Message ID | 20240930-spi-v2-2-ed7f6bcbe0df@nxp.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | spi-nor: add optional vcc-supply and enable it for i.MX93 | expand |
Hi, peng.fan@oss.nxp.com wrote on Mon, 30 Sep 2024 17:22:25 +0800: > From: Peng Fan <peng.fan@nxp.com> > > SPI NOR flashes needs power supply to work properly. The power supply > maybe software controllable per board design. So add the support > for an vcc-supply regulator. > > Signed-off-by: Peng Fan <peng.fan@nxp.com> > --- > drivers/mtd/spi-nor/core.c | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c > index 9d6e85bf227b..5249c8b13916 100644 > --- a/drivers/mtd/spi-nor/core.c > +++ b/drivers/mtd/spi-nor/core.c > @@ -17,6 +17,7 @@ > #include <linux/mtd/spi-nor.h> > #include <linux/mutex.h> > #include <linux/of_platform.h> > +#include <linux/regulator/consumer.h> > #include <linux/sched/task_stack.h> > #include <linux/sizes.h> > #include <linux/slab.h> > @@ -3462,6 +3463,10 @@ int spi_nor_scan(struct spi_nor *nor, const char *name, > if (!nor->bouncebuf) > return -ENOMEM; > > + ret = devm_regulator_get_enable(dev, "vcc"); _optional ? > + if (ret) > + return ret; > + > ret = spi_nor_hw_reset(nor); > if (ret) > return ret; > Thanks, Miquèl
On 30.09.2024 17:22:25, Peng Fan (OSS) wrote: > From: Peng Fan <peng.fan@nxp.com> > > SPI NOR flashes needs power supply to work properly. The power supply > maybe software controllable per board design. So add the support > for an vcc-supply regulator. > > Signed-off-by: Peng Fan <peng.fan@nxp.com> > --- > drivers/mtd/spi-nor/core.c | 5 +++++ > 1 file changed, 5 insertions(+) > > diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c > index 9d6e85bf227b..5249c8b13916 100644 > --- a/drivers/mtd/spi-nor/core.c > +++ b/drivers/mtd/spi-nor/core.c > @@ -17,6 +17,7 @@ > #include <linux/mtd/spi-nor.h> > #include <linux/mutex.h> > #include <linux/of_platform.h> > +#include <linux/regulator/consumer.h> > #include <linux/sched/task_stack.h> > #include <linux/sizes.h> > #include <linux/slab.h> > @@ -3462,6 +3463,10 @@ int spi_nor_scan(struct spi_nor *nor, const char *name, > if (!nor->bouncebuf) > return -ENOMEM; > > + ret = devm_regulator_get_enable(dev, "vcc"); > + if (ret) > + return ret; > + What happens if the SPI-NOR doesn't have a "vcc" regulator? Marc
On 30.09.2024 11:21:27, Marc Kleine-Budde wrote: > On 30.09.2024 17:22:25, Peng Fan (OSS) wrote: > > From: Peng Fan <peng.fan@nxp.com> > > > > SPI NOR flashes needs power supply to work properly. The power supply > > maybe software controllable per board design. So add the support > > for an vcc-supply regulator. > > > > Signed-off-by: Peng Fan <peng.fan@nxp.com> > > --- > > drivers/mtd/spi-nor/core.c | 5 +++++ > > 1 file changed, 5 insertions(+) > > > > diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c > > index 9d6e85bf227b..5249c8b13916 100644 > > --- a/drivers/mtd/spi-nor/core.c > > +++ b/drivers/mtd/spi-nor/core.c > > @@ -17,6 +17,7 @@ > > #include <linux/mtd/spi-nor.h> > > #include <linux/mutex.h> > > #include <linux/of_platform.h> > > +#include <linux/regulator/consumer.h> > > #include <linux/sched/task_stack.h> > > #include <linux/sizes.h> > > #include <linux/slab.h> > > @@ -3462,6 +3463,10 @@ int spi_nor_scan(struct spi_nor *nor, const char *name, > > if (!nor->bouncebuf) > > return -ENOMEM; > > > > + ret = devm_regulator_get_enable(dev, "vcc"); > > + if (ret) > > + return ret; > > + > > What happens if the SPI-NOR doesn't have a "vcc" regulator? ...the SPI-NOR will use the dummy regulator. regards, Marc
Hi Miquel, On 24-09-30, Miquel Raynal wrote: > Hi, > > peng.fan@oss.nxp.com wrote on Mon, 30 Sep 2024 17:22:25 +0800: > > > From: Peng Fan <peng.fan@nxp.com> > > > > SPI NOR flashes needs power supply to work properly. The power supply > > maybe software controllable per board design. So add the support > > for an vcc-supply regulator. > > > > Signed-off-by: Peng Fan <peng.fan@nxp.com> > > --- > > drivers/mtd/spi-nor/core.c | 5 +++++ > > 1 file changed, 5 insertions(+) > > > > diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c > > index 9d6e85bf227b..5249c8b13916 100644 > > --- a/drivers/mtd/spi-nor/core.c > > +++ b/drivers/mtd/spi-nor/core.c > > @@ -17,6 +17,7 @@ > > #include <linux/mtd/spi-nor.h> > > #include <linux/mutex.h> > > #include <linux/of_platform.h> > > +#include <linux/regulator/consumer.h> > > #include <linux/sched/task_stack.h> > > #include <linux/sizes.h> > > #include <linux/slab.h> > > @@ -3462,6 +3463,10 @@ int spi_nor_scan(struct spi_nor *nor, const char *name, > > if (!nor->bouncebuf) > > return -ENOMEM; > > > > + ret = devm_regulator_get_enable(dev, "vcc"); > > _optional ? The regulator optional API is different compared to other optional APIs. If we would use optional here, we would need to check the returned error code. On the other hand if the non optional API is used and the regualtor is missing, a dummy regualtor is returned (as pointed out by Marc). Please see the _optional API doc to see more information about the _optional usage. Regards, Marco > > + if (ret) > > + return ret; > > + > > ret = spi_nor_hw_reset(nor); > > if (ret) > > return ret; > > > > > Thanks, > Miquèl >
Hi, > > > + ret = devm_regulator_get_enable(dev, "vcc"); > > > + if (ret) > > > + return ret; > > > + > > > > What happens if the SPI-NOR doesn't have a "vcc" regulator? > > ...the SPI-NOR will use the dummy regulator. Which then prints a warning "using dummy regulator". Is that the usual way to go? I mean the regulator is actually mandatory because it is the main voltage rail for the flash. To get rid of the warning one can add a fixed-regulator (which is correct anyway). But OTOH, the device tree lists it as optional (marking it as required isn't an option either because virtually all device trees won't have that property). -michael
On 10/7/24 8:50 AM, Michael Walle wrote: > Hi, > >>>> + ret = devm_regulator_get_enable(dev, "vcc"); >>>> + if (ret) >>>> + return ret; >>>> + >>> >>> What happens if the SPI-NOR doesn't have a "vcc" regulator? >> >> ...the SPI-NOR will use the dummy regulator. > > Which then prints a warning "using dummy regulator". Is that the > usual way to go? > I mean the regulator is actually mandatory because it is the main > voltage rail for the flash. To get rid of the warning one can add a > fixed-regulator (which is correct anyway). But OTOH, the device tree I think it's fine to have the warning. It will help people write a better device tree. > lists it as optional (marking it as required isn't an option either > because virtually all device trees won't have that property). > > -michael
diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c index 9d6e85bf227b..5249c8b13916 100644 --- a/drivers/mtd/spi-nor/core.c +++ b/drivers/mtd/spi-nor/core.c @@ -17,6 +17,7 @@ #include <linux/mtd/spi-nor.h> #include <linux/mutex.h> #include <linux/of_platform.h> +#include <linux/regulator/consumer.h> #include <linux/sched/task_stack.h> #include <linux/sizes.h> #include <linux/slab.h> @@ -3462,6 +3463,10 @@ int spi_nor_scan(struct spi_nor *nor, const char *name, if (!nor->bouncebuf) return -ENOMEM; + ret = devm_regulator_get_enable(dev, "vcc"); + if (ret) + return ret; + ret = spi_nor_hw_reset(nor); if (ret) return ret;