Message ID | 1434211748-25572-4-git-send-email-hdegoede@redhat.com |
---|---|
State | New |
Headers | show |
> Add a cell for the usb power_supply part of the axp20x PMICs. Why are you duplicating the subject line? > Note that this cell is only for the usb power_supply part and not the > ac-power / battery-charger / rtc-backup-bat-charger bits. > > Depending on the board each of those must be enabled / disabled separately > in devicetree as most boards do not use all 4. So in dt each one needs its > own child-node of the axp20x node. Another reason for using separate child > nodes for each is so that other devicetree nodes can have a power-supply > property with a phandle referencing a node representing a single > power-supply. > > The decision to use a separate devicetree node for each is reflected on > the kernel side by each getting its own mfd-cell / platform_device and > platform-driver. You don't really need to say any of this, as this is the 'norm'. What you didn't mention however, is that you're taking the opportunity to fix some formatting issues and that there are no functional changes in these lines. > Cc: Bruno Prémont <bonbons@linux-vserver.org> > Signed-off-by: Hans de Goede <hdegoede@redhat.com> > --- > Changes in v2: > -Use DEFINE_RES_IRQ_NAMED > -Change indentation of axp20x_cells initializers to avoid line wrapping > --- > drivers/mfd/axp20x.c | 20 ++++++++++++++++---- > 1 file changed, 16 insertions(+), 4 deletions(-) Patch looks okay however: Acked-by: Lee Jones <lee.jones@linaro.org> > diff --git a/drivers/mfd/axp20x.c b/drivers/mfd/axp20x.c > index f9a3c2d..ca4a604 100644 > --- a/drivers/mfd/axp20x.c > +++ b/drivers/mfd/axp20x.c > @@ -113,6 +113,13 @@ static struct resource axp20x_pek_resources[] = { > }, > }; > > +static struct resource axp20x_usb_power_supply_resources[] = { > + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_PLUGIN, "VBUS_PLUGIN"), > + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_REMOVAL, "VBUS_REMOVAL"), > + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_VALID, "VBUS_VALID"), > + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_NOT_VALID, "VBUS_NOT_VALID"), > +}; > + > static struct resource axp22x_pek_resources[] = { > { > .name = "PEK_DBR", > @@ -363,11 +370,16 @@ static const struct regmap_irq_chip axp288_regmap_irq_chip = { > > static struct mfd_cell axp20x_cells[] = { > { > - .name = "axp20x-pek", > - .num_resources = ARRAY_SIZE(axp20x_pek_resources), > - .resources = axp20x_pek_resources, > + .name = "axp20x-pek", > + .num_resources = ARRAY_SIZE(axp20x_pek_resources), > + .resources = axp20x_pek_resources, > }, { > - .name = "axp20x-regulator", > + .name = "axp20x-regulator", > + }, { > + .name = "axp20x-usb-power-supply", > + .of_compatible = "x-powers,axp202-usb-power-supply", > + .num_resources = ARRAY_SIZE(axp20x_usb_power_supply_resources), > + .resources = axp20x_usb_power_supply_resources, > }, > }; >
Hi, On 24-06-15 13:23, Lee Jones wrote: >> Add a cell for the usb power_supply part of the axp20x PMICs. > > Why are you duplicating the subject line? Heh, because some maintainers insist that the main-body part of the commit message must stand by itself, without needing the subject to be understandable... >> Note that this cell is only for the usb power_supply part and not the >> ac-power / battery-charger / rtc-backup-bat-charger bits. >> >> Depending on the board each of those must be enabled / disabled separately >> in devicetree as most boards do not use all 4. So in dt each one needs its >> own child-node of the axp20x node. Another reason for using separate child >> nodes for each is so that other devicetree nodes can have a power-supply >> property with a phandle referencing a node representing a single >> power-supply. >> >> The decision to use a separate devicetree node for each is reflected on >> the kernel side by each getting its own mfd-cell / platform_device and >> platform-driver. > > You don't really need to say any of this, as this is the 'norm'. I agree it should be the norm, but I'm not sure if it actually is, while working on this I've seen several drivers which instantiate multiple power-supply class devices from a single mfd-cell. And this is what Bruno's original patches did, so I would prefer to keep this > What you didn't mention however, is that you're taking the opportunity > to fix some formatting issues and that there are no functional changes > in these lines. That is the end result of your request to change the indentation to avoid line-wrapping :) > >> Cc: Bruno Prémont <bonbons@linux-vserver.org> >> Signed-off-by: Hans de Goede <hdegoede@redhat.com> >> --- >> Changes in v2: >> -Use DEFINE_RES_IRQ_NAMED >> -Change indentation of axp20x_cells initializers to avoid line wrapping >> --- >> drivers/mfd/axp20x.c | 20 ++++++++++++++++---- >> 1 file changed, 16 insertions(+), 4 deletions(-) > > Patch looks okay however: > > Acked-by: Lee Jones <lee.jones@linaro.org> Thanks. Regards, Hans > >> diff --git a/drivers/mfd/axp20x.c b/drivers/mfd/axp20x.c >> index f9a3c2d..ca4a604 100644 >> --- a/drivers/mfd/axp20x.c >> +++ b/drivers/mfd/axp20x.c >> @@ -113,6 +113,13 @@ static struct resource axp20x_pek_resources[] = { >> }, >> }; >> >> +static struct resource axp20x_usb_power_supply_resources[] = { >> + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_PLUGIN, "VBUS_PLUGIN"), >> + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_REMOVAL, "VBUS_REMOVAL"), >> + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_VALID, "VBUS_VALID"), >> + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_NOT_VALID, "VBUS_NOT_VALID"), >> +}; >> + >> static struct resource axp22x_pek_resources[] = { >> { >> .name = "PEK_DBR", >> @@ -363,11 +370,16 @@ static const struct regmap_irq_chip axp288_regmap_irq_chip = { >> >> static struct mfd_cell axp20x_cells[] = { >> { >> - .name = "axp20x-pek", >> - .num_resources = ARRAY_SIZE(axp20x_pek_resources), >> - .resources = axp20x_pek_resources, >> + .name = "axp20x-pek", >> + .num_resources = ARRAY_SIZE(axp20x_pek_resources), >> + .resources = axp20x_pek_resources, >> }, { >> - .name = "axp20x-regulator", >> + .name = "axp20x-regulator", >> + }, { >> + .name = "axp20x-usb-power-supply", >> + .of_compatible = "x-powers,axp202-usb-power-supply", >> + .num_resources = ARRAY_SIZE(axp20x_usb_power_supply_resources), >> + .resources = axp20x_usb_power_supply_resources, >> }, >> }; >> >
On Wed, 24 Jun 2015, Hans de Goede wrote: > On 24-06-15 13:23, Lee Jones wrote: > >>Add a cell for the usb power_supply part of the axp20x PMICs. > > > >Why are you duplicating the subject line? > > Heh, because some maintainers insist that the main-body part of the > commit message must stand by itself, without needing the subject to be > understandable... Some Maintainers are crazy. ;) > >>Note that this cell is only for the usb power_supply part and not the > >>ac-power / battery-charger / rtc-backup-bat-charger bits. > >> > >>Depending on the board each of those must be enabled / disabled separately > >>in devicetree as most boards do not use all 4. So in dt each one needs its > >>own child-node of the axp20x node. Another reason for using separate child > >>nodes for each is so that other devicetree nodes can have a power-supply > >>property with a phandle referencing a node representing a single > >>power-supply. > >> > >>The decision to use a separate devicetree node for each is reflected on > >>the kernel side by each getting its own mfd-cell / platform_device and > >>platform-driver. > > > >You don't really need to say any of this, as this is the 'norm'. > > I agree it should be the norm, but I'm not sure if it actually is, while > working on this I've seen several drivers which instantiate multiple > power-supply class devices from a single mfd-cell. And this is what > Bruno's original patches did, so I would prefer to keep this Up to you, but it's really not required. No one else says this stuff. > >What you didn't mention however, is that you're taking the opportunity > >to fix some formatting issues and that there are no functional changes > >in these lines. > > That is the end result of your request to change the indentation to > avoid line-wrapping :) Fine, you can add my Suggested-by too then [don't actually do that], but regardless of where the idea came from you should still tell us what the patch does in the commit log. I was looking for functional differences within those formatting changes. If I wanted to be really religious about it I would [rightly] say "don't mix functional changes with clean-ups", but instead I'll politely ask for a quick mention in the commit log when you re-submit. Aren't I nice?! ;) > >>Cc: Bruno Prémont <bonbons@linux-vserver.org> > >>Signed-off-by: Hans de Goede <hdegoede@redhat.com> > >>--- > >>Changes in v2: > >>-Use DEFINE_RES_IRQ_NAMED > >>-Change indentation of axp20x_cells initializers to avoid line wrapping > >>--- > >> drivers/mfd/axp20x.c | 20 ++++++++++++++++---- > >> 1 file changed, 16 insertions(+), 4 deletions(-) > > > >Patch looks okay however: > > > >Acked-by: Lee Jones <lee.jones@linaro.org> > > Thanks. > > Regards, > > Hans > > > > > >>diff --git a/drivers/mfd/axp20x.c b/drivers/mfd/axp20x.c > >>index f9a3c2d..ca4a604 100644 > >>--- a/drivers/mfd/axp20x.c > >>+++ b/drivers/mfd/axp20x.c > >>@@ -113,6 +113,13 @@ static struct resource axp20x_pek_resources[] = { > >> }, > >> }; > >> > >>+static struct resource axp20x_usb_power_supply_resources[] = { > >>+ DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_PLUGIN, "VBUS_PLUGIN"), > >>+ DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_REMOVAL, "VBUS_REMOVAL"), > >>+ DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_VALID, "VBUS_VALID"), > >>+ DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_NOT_VALID, "VBUS_NOT_VALID"), > >>+}; > >>+ > >> static struct resource axp22x_pek_resources[] = { > >> { > >> .name = "PEK_DBR", > >>@@ -363,11 +370,16 @@ static const struct regmap_irq_chip axp288_regmap_irq_chip = { > >> > >> static struct mfd_cell axp20x_cells[] = { > >> { > >>- .name = "axp20x-pek", > >>- .num_resources = ARRAY_SIZE(axp20x_pek_resources), > >>- .resources = axp20x_pek_resources, > >>+ .name = "axp20x-pek", > >>+ .num_resources = ARRAY_SIZE(axp20x_pek_resources), > >>+ .resources = axp20x_pek_resources, > >> }, { > >>- .name = "axp20x-regulator", > >>+ .name = "axp20x-regulator", > >>+ }, { > >>+ .name = "axp20x-usb-power-supply", > >>+ .of_compatible = "x-powers,axp202-usb-power-supply", > >>+ .num_resources = ARRAY_SIZE(axp20x_usb_power_supply_resources), > >>+ .resources = axp20x_usb_power_supply_resources, > >> }, > >> }; > >> > >
diff --git a/drivers/mfd/axp20x.c b/drivers/mfd/axp20x.c index f9a3c2d..ca4a604 100644 --- a/drivers/mfd/axp20x.c +++ b/drivers/mfd/axp20x.c @@ -113,6 +113,13 @@ static struct resource axp20x_pek_resources[] = { }, }; +static struct resource axp20x_usb_power_supply_resources[] = { + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_PLUGIN, "VBUS_PLUGIN"), + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_REMOVAL, "VBUS_REMOVAL"), + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_VALID, "VBUS_VALID"), + DEFINE_RES_IRQ_NAMED(AXP20X_IRQ_VBUS_NOT_VALID, "VBUS_NOT_VALID"), +}; + static struct resource axp22x_pek_resources[] = { { .name = "PEK_DBR", @@ -363,11 +370,16 @@ static const struct regmap_irq_chip axp288_regmap_irq_chip = { static struct mfd_cell axp20x_cells[] = { { - .name = "axp20x-pek", - .num_resources = ARRAY_SIZE(axp20x_pek_resources), - .resources = axp20x_pek_resources, + .name = "axp20x-pek", + .num_resources = ARRAY_SIZE(axp20x_pek_resources), + .resources = axp20x_pek_resources, }, { - .name = "axp20x-regulator", + .name = "axp20x-regulator", + }, { + .name = "axp20x-usb-power-supply", + .of_compatible = "x-powers,axp202-usb-power-supply", + .num_resources = ARRAY_SIZE(axp20x_usb_power_supply_resources), + .resources = axp20x_usb_power_supply_resources, }, };
Add a cell for the usb power_supply part of the axp20x PMICs. Note that this cell is only for the usb power_supply part and not the ac-power / battery-charger / rtc-backup-bat-charger bits. Depending on the board each of those must be enabled / disabled separately in devicetree as most boards do not use all 4. So in dt each one needs its own child-node of the axp20x node. Another reason for using separate child nodes for each is so that other devicetree nodes can have a power-supply property with a phandle referencing a node representing a single power-supply. The decision to use a separate devicetree node for each is reflected on the kernel side by each getting its own mfd-cell / platform_device and platform-driver. Cc: Bruno Prémont <bonbons@linux-vserver.org> Signed-off-by: Hans de Goede <hdegoede@redhat.com> --- Changes in v2: -Use DEFINE_RES_IRQ_NAMED -Change indentation of axp20x_cells initializers to avoid line wrapping --- drivers/mfd/axp20x.c | 20 ++++++++++++++++---- 1 file changed, 16 insertions(+), 4 deletions(-)