| Message ID | 1417524339-4604-1-git-send-email-yadi.brar@samsung.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On 02.12.2014 13:45, Yadwinder Singh Brar wrote: > rtc have different i2c client than power(pmic) block. So rtc device should > sit under its own i2c client in device hierarchy, which reflects in sysfs also. > This patch modifies code to register rtc cell with rtc->dev as parent. > > Without this patch : > # ls /sys/class/i2c-adapter/i2c-0/0-0009/ > driver max77686-pmic modalias power uevent > max77686-clk max77686-rtc name subsystem > > After applying patch : > # ls /sys/class/i2c-adapter/i2c-0/0-0006/ > driver/ modalias power/ uevent > max77686-rtc/ name subsystem/ > > Signed-off-by: Yadwinder Singh Brar <yadi.brar@samsung.com> > --- > > Or Can we follow another (exhaustive but more cleaner) approach, which will > be more like code refactoring and cleanup rather than only fix: > Since rtc uses i2c client, which gets created using i2c_new_dummy() and is not > shared by any other cell of max77686. So we can covert rtc platform driver > itself to i2c client driver. It will also allow to expilicitly describe > max77686-rtc in DT which we can't do now. > It can be applicable to some other existing and new mfd pmic drivers. > Any suggestion/comments ? Hi, What kind of problem is solved by this patch? Best regards, Krzysztof > > --- > drivers/mfd/max77686.c | 22 ++++++++++++++++++++-- > 1 files changed, 20 insertions(+), 2 deletions(-) > > diff --git a/drivers/mfd/max77686.c b/drivers/mfd/max77686.c > index 929795e..22c0948 100644 > --- a/drivers/mfd/max77686.c > +++ b/drivers/mfd/max77686.c > @@ -39,10 +39,13 @@ > > static const struct mfd_cell max77686_devs[] = { > { .name = "max77686-pmic", }, > - { .name = "max77686-rtc", }, > { .name = "max77686-clk", }, > }; > > +static const struct mfd_cell max77686_rtc_dev[] = { > + { .name = "max77686-rtc", }, > +}; > + > static const struct mfd_cell max77802_devs[] = { > { .name = "max77802-pmic", }, > { .name = "max77802-clk", }, > @@ -332,14 +335,27 @@ static int max77686_i2c_probe(struct i2c_client *i2c, > goto err_del_irqc; > } > > + if (max77686->type == TYPE_MAX77686) { > + ret = mfd_add_devices(&max77686->rtc->dev, -1, max77686_rtc_dev, > + 1, NULL, 0, NULL); > + if (ret < 0) { > + dev_err(&max77686->rtc->dev, > + "failed to add RTC device %d\n", ret); > + goto err_del_rtc_irqc; > + } > + } > + > ret = mfd_add_devices(max77686->dev, -1, cells, n_devs, NULL, 0, NULL); > if (ret < 0) { > dev_err(&i2c->dev, "failed to add MFD devices: %d\n", ret); > - goto err_del_rtc_irqc; > + goto err_del_rtc_dev; > } > > return 0; > > +err_del_rtc_dev: > + if (max77686->type == TYPE_MAX77686) > + mfd_remove_devices(&max77686->rtc->dev); > err_del_rtc_irqc: > regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); > err_del_irqc: > @@ -356,6 +372,8 @@ static int max77686_i2c_remove(struct i2c_client *i2c) > struct max77686_dev *max77686 = i2c_get_clientdata(i2c); > > mfd_remove_devices(max77686->dev); > + if (max77686->type == TYPE_MAX77686) > + mfd_remove_devices(&max77686->rtc->dev); > > regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); > regmap_del_irq_chip(max77686->irq, max77686->irq_data); > -- To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tuesday, December 02, 2014 7:04 PM, Krzysztof Koz?owski wrote: > On 02.12.2014 13:45, Yadwinder Singh Brar wrote: > > rtc have different i2c client than power(pmic) block. So rtc device > > should sit under its own i2c client in device hierarchy, which > reflects in sysfs also. > > This patch modifies code to register rtc cell with rtc->dev as > parent. > > > > Without this patch : > > # ls /sys/class/i2c-adapter/i2c-0/0-0009/ > > driver max77686-pmic modalias power uevent > > max77686-clk max77686-rtc name subsystem > > > > After applying patch : > > # ls /sys/class/i2c-adapter/i2c-0/0-0006/ > > driver/ modalias power/ uevent > > max77686-rtc/ name subsystem/ > > > > Signed-off-by: Yadwinder Singh Brar <yadi.brar@samsung.com> > > --- > > > > Or Can we follow another (exhaustive but more cleaner) approach, > which > > will be more like code refactoring and cleanup rather than only fix: > > Since rtc uses i2c client, which gets created using i2c_new_dummy() > > and is not shared by any other cell of max77686. So we can covert rtc > > platform driver itself to i2c client driver. It will also allow to > > expilicitly describe max77686-rtc in DT which we can't do now. > > It can be applicable to some other existing and new mfd pmic drivers. > > Any suggestion/comments ? > > Hi, > > What kind of problem is solved by this patch? > Let me try to explain once again :) After seeing a message "i2c i2c-0: .... , addr=0x06, .." in dmesg log, I was not able to find any such device in sysfs as well as device tree. There was no device under /sys/class/i2c-dev/i2c-0/device/0-0006/ Isn't something wrong or missing ? This patch fixes that missing parent child relation, which IMO should be correct always, though it causes any major problem or not. Still I am thinking, 0-0006 slave device(rtc) shouldn't also appear in DT? As DT should describe the hardware that we are using. Warm regards, Yadwinder > Best regards, > Krzysztof > > > > > > --- > > drivers/mfd/max77686.c | 22 ++++++++++++++++++++-- > > 1 files changed, 20 insertions(+), 2 deletions(-) > > > > diff --git a/drivers/mfd/max77686.c b/drivers/mfd/max77686.c index > > 929795e..22c0948 100644 > > --- a/drivers/mfd/max77686.c > > +++ b/drivers/mfd/max77686.c > > @@ -39,10 +39,13 @@ > > > > static const struct mfd_cell max77686_devs[] = { > > { .name = "max77686-pmic", }, > > - { .name = "max77686-rtc", }, > > { .name = "max77686-clk", }, > > }; > > > > +static const struct mfd_cell max77686_rtc_dev[] = { > > + { .name = "max77686-rtc", }, > > +}; > > + > > static const struct mfd_cell max77802_devs[] = { > > { .name = "max77802-pmic", }, > > { .name = "max77802-clk", }, > > @@ -332,14 +335,27 @@ static int max77686_i2c_probe(struct i2c_client > *i2c, > > goto err_del_irqc; > > } > > > > + if (max77686->type == TYPE_MAX77686) { > > + ret = mfd_add_devices(&max77686->rtc->dev, -1, > max77686_rtc_dev, > > + 1, NULL, 0, NULL); > > + if (ret < 0) { > > + dev_err(&max77686->rtc->dev, > > + "failed to add RTC device %d\n", ret); > > + goto err_del_rtc_irqc; > > + } > > + } > > + > > ret = mfd_add_devices(max77686->dev, -1, cells, n_devs, NULL, 0, > NULL); > > if (ret < 0) { > > dev_err(&i2c->dev, "failed to add MFD devices: %d\n", ret); > > - goto err_del_rtc_irqc; > > + goto err_del_rtc_dev; > > } > > > > return 0; > > > > +err_del_rtc_dev: > > + if (max77686->type == TYPE_MAX77686) > > + mfd_remove_devices(&max77686->rtc->dev); > > err_del_rtc_irqc: > > regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); > > err_del_irqc: > > @@ -356,6 +372,8 @@ static int max77686_i2c_remove(struct i2c_client > *i2c) > > struct max77686_dev *max77686 = i2c_get_clientdata(i2c); > > > > mfd_remove_devices(max77686->dev); > > + if (max77686->type == TYPE_MAX77686) > > + mfd_remove_devices(&max77686->rtc->dev); > > > > regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); > > regmap_del_irq_chip(max77686->irq, max77686->irq_data); > > -- To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On ?ro, 2014-12-03 at 14:32 +0530, Yadwinder Singh Brar wrote: > > On Tuesday, December 02, 2014 7:04 PM, Krzysztof Koz?owski wrote: > > On 02.12.2014 13:45, Yadwinder Singh Brar wrote: > > > rtc have different i2c client than power(pmic) block. So rtc device > > > should sit under its own i2c client in device hierarchy, which > > reflects in sysfs also. > > > This patch modifies code to register rtc cell with rtc->dev as > > parent. > > > > > > Without this patch : > > > # ls /sys/class/i2c-adapter/i2c-0/0-0009/ > > > driver max77686-pmic modalias power uevent > > > max77686-clk max77686-rtc name subsystem > > > > > > After applying patch : > > > # ls /sys/class/i2c-adapter/i2c-0/0-0006/ > > > driver/ modalias power/ uevent > > > max77686-rtc/ name subsystem/ > > > > > > Signed-off-by: Yadwinder Singh Brar <yadi.brar@samsung.com> > > > --- > > > > > > Or Can we follow another (exhaustive but more cleaner) approach, > > which > > > will be more like code refactoring and cleanup rather than only fix: > > > Since rtc uses i2c client, which gets created using i2c_new_dummy() > > > and is not shared by any other cell of max77686. So we can covert rtc > > > platform driver itself to i2c client driver. It will also allow to > > > expilicitly describe max77686-rtc in DT which we can't do now. > > > It can be applicable to some other existing and new mfd pmic drivers. > > > Any suggestion/comments ? > > > > Hi, > > > > What kind of problem is solved by this patch? > > > > Let me try to explain once again :) > After seeing a message "i2c i2c-0: .... , addr=0x06, .." in dmesg log, > I was not able to find any such device in sysfs as well as device tree. > There was no device under /sys/class/i2c-dev/i2c-0/device/0-0006/ > Isn't something wrong or missing ? > > This patch fixes that missing parent child relation, which IMO > should be correct always, though it causes any major problem or not. OK, I got your point. I'm fine with both solutions and the patch looks good, so: Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> > Still I am thinking, 0-0006 slave device(rtc) shouldn't also appear in DT? > As DT should describe the hardware that we are using. Your patch properly describes the hardware. However from driver perspective, the RTC here is not an standalone driver and depends on parent (MFD) driver. Although max77686 RTC could have its own DeviceTree node, I think it should still be put under main MFD driver's node, because the parent manages stuff like interrupts. Best regards, Krzysztof -- To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wednesday, December 03, 2014 2:56 PM, Krzysztof Kozlowski wrote: > > On ?ro, 2014-12-03 at 14:32 +0530, Yadwinder Singh Brar wrote: > > > > On Tuesday, December 02, 2014 7:04 PM, Krzysztof Koz?owski wrote: > > > On 02.12.2014 13:45, Yadwinder Singh Brar wrote: > > > > rtc have different i2c client than power(pmic) block. So rtc > > > > device should sit under its own i2c client in device hierarchy, > > > > which > > > reflects in sysfs also. > > > > This patch modifies code to register rtc cell with rtc->dev as > > > parent. > > > > > > > > Without this patch : > > > > # ls /sys/class/i2c-adapter/i2c-0/0-0009/ > > > > driver max77686-pmic modalias power > uevent > > > > max77686-clk max77686-rtc name subsystem > > > > > > > > After applying patch : > > > > # ls /sys/class/i2c-adapter/i2c-0/0-0006/ > > > > driver/ modalias power/ uevent > > > > max77686-rtc/ name subsystem/ > > > > > > > > Signed-off-by: Yadwinder Singh Brar <yadi.brar@samsung.com> > > > > --- > > > > > > > > Or Can we follow another (exhaustive but more cleaner) approach, > > > which > > > > will be more like code refactoring and cleanup rather than only > fix: > > > > Since rtc uses i2c client, which gets created using > > > > i2c_new_dummy() and is not shared by any other cell of max77686. > > > > So we can covert rtc platform driver itself to i2c client driver. > > > > It will also allow to expilicitly describe max77686-rtc in DT > which we can't do now. > > > > It can be applicable to some other existing and new mfd pmic > drivers. > > > > Any suggestion/comments ? > > > > > > Hi, > > > > > > What kind of problem is solved by this patch? > > > > > > > Let me try to explain once again :) > > After seeing a message "i2c i2c-0: .... , addr=0x06, .." in dmesg > log, > > I was not able to find any such device in sysfs as well as device > tree. > > There was no device under /sys/class/i2c-dev/i2c-0/device/0-0006/ > > Isn't something wrong or missing ? > > > > This patch fixes that missing parent child relation, which IMO should > > be correct always, though it causes any major problem or not. > > OK, I got your point. I'm fine with both solutions and the patch looks > good, so: > Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> > Thanks for reviewing :). > > Still I am thinking, 0-0006 slave device(rtc) shouldn't also appear > in DT? > > As DT should describe the hardware that we are using. > > Your patch properly describes the hardware. However from driver > perspective, the RTC here is not an standalone driver and depends on > parent (MFD) driver. > Hmm ... I can't see any hard dependency. I think rtc can survive as independent i2c device. And rtc(@06) is more like a sibling of existing MFD driver(max77686@09) > Although max77686 RTC could have its own DeviceTree node, I think it > should still be put under main MFD driver's node, because the parent > manages stuff like interrupts. > I think it will not sound good to have something like max77686-rtc@06 node under max77686@09 node, since both are kind of siblings. And interrupts should not be a restricting factor, In case of max77686 interrupt source and mask registers for rtc also lie in rtc i2c register bank so we can add an irq_chip from rtc driver itself. Only thing I can see which can be little bit tricky, is providing alarm->pending info(pending interrupt) in alarm_read() callback in rtc as currently its based on MAX77686_REG_STATUS2 which is only common register in power(pmic) register bank of max77686. Though in practical scenarios it may not be non zero, since reading once source register clears this status also, but we may have to keep it for sake of functionality completeness. I can’t see any way(something similar to syscon) for sharing i2c regmap between independent drivers, other than getting reference to regmap from i2c device found by of_find_i2c_device_by_node(). Anyone can suggest better idea? Best Regards, Yadwinder -- To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/mfd/max77686.c b/drivers/mfd/max77686.c index 929795e..22c0948 100644 --- a/drivers/mfd/max77686.c +++ b/drivers/mfd/max77686.c @@ -39,10 +39,13 @@ static const struct mfd_cell max77686_devs[] = { { .name = "max77686-pmic", }, - { .name = "max77686-rtc", }, { .name = "max77686-clk", }, }; +static const struct mfd_cell max77686_rtc_dev[] = { + { .name = "max77686-rtc", }, +}; + static const struct mfd_cell max77802_devs[] = { { .name = "max77802-pmic", }, { .name = "max77802-clk", }, @@ -332,14 +335,27 @@ static int max77686_i2c_probe(struct i2c_client *i2c, goto err_del_irqc; } + if (max77686->type == TYPE_MAX77686) { + ret = mfd_add_devices(&max77686->rtc->dev, -1, max77686_rtc_dev, + 1, NULL, 0, NULL); + if (ret < 0) { + dev_err(&max77686->rtc->dev, + "failed to add RTC device %d\n", ret); + goto err_del_rtc_irqc; + } + } + ret = mfd_add_devices(max77686->dev, -1, cells, n_devs, NULL, 0, NULL); if (ret < 0) { dev_err(&i2c->dev, "failed to add MFD devices: %d\n", ret); - goto err_del_rtc_irqc; + goto err_del_rtc_dev; } return 0; +err_del_rtc_dev: + if (max77686->type == TYPE_MAX77686) + mfd_remove_devices(&max77686->rtc->dev); err_del_rtc_irqc: regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); err_del_irqc: @@ -356,6 +372,8 @@ static int max77686_i2c_remove(struct i2c_client *i2c) struct max77686_dev *max77686 = i2c_get_clientdata(i2c); mfd_remove_devices(max77686->dev); + if (max77686->type == TYPE_MAX77686) + mfd_remove_devices(&max77686->rtc->dev); regmap_del_irq_chip(max77686->irq, max77686->rtc_irq_data); regmap_del_irq_chip(max77686->irq, max77686->irq_data);
rtc have different i2c client than power(pmic) block. So rtc device should sit under its own i2c client in device hierarchy, which reflects in sysfs also. This patch modifies code to register rtc cell with rtc->dev as parent. Without this patch : # ls /sys/class/i2c-adapter/i2c-0/0-0009/ driver max77686-pmic modalias power uevent max77686-clk max77686-rtc name subsystem After applying patch : # ls /sys/class/i2c-adapter/i2c-0/0-0006/ driver/ modalias power/ uevent max77686-rtc/ name subsystem/ Signed-off-by: Yadwinder Singh Brar <yadi.brar@samsung.com> --- Or Can we follow another (exhaustive but more cleaner) approach, which will be more like code refactoring and cleanup rather than only fix: Since rtc uses i2c client, which gets created using i2c_new_dummy() and is not shared by any other cell of max77686. So we can covert rtc platform driver itself to i2c client driver. It will also allow to expilicitly describe max77686-rtc in DT which we can't do now. It can be applicable to some other existing and new mfd pmic drivers. Any suggestion/comments ? --- drivers/mfd/max77686.c | 22 ++++++++++++++++++++-- 1 files changed, 20 insertions(+), 2 deletions(-)