| Message ID | 20240220084046.23786-2-shradha.t@samsung.com |
|---|---|
| State | New |
| Headers | show |
| Series | Add helper function to get and enable all bulk clocks | expand |
Quoting Shradha Todi (2024-02-20 00:40:45) > Provide a managed devm_clk_bulk* wrapper to get and enable all > bulk clocks in order to simplify drivers that keeps all clocks > enabled for the time of driver operation. > > Suggested-by: Marek Szyprowski <m.szyprowski@samsung.com> > Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com> > Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> > Signed-off-by: Shradha Todi <shradha.t@samsung.com> > --- Applied to clk-next
On Tue, Feb 20, 2024 at 02:10:45PM +0530, Shradha Todi wrote: > +int __must_check devm_clk_bulk_get_all_enable(struct device *dev, > + struct clk_bulk_data **clks) > +{ > + struct clk_bulk_devres *devres; > + int ret; > + > + devres = devres_alloc(devm_clk_bulk_release_all_enable, > + sizeof(*devres), GFP_KERNEL); > + if (!devres) > + return -ENOMEM; > + > + ret = clk_bulk_get_all(dev, &devres->clks); > + if (ret > 0) { I feel like this should be >= instead of >. There aren't any callers of this function yet so we can't see what's in *clks at the start but it's easy to imagine a situation where it's bad data. > + *clks = devres->clks; > + devres->num_clks = ret; > + } else { > + devres_free(devres); > + return ret; When clk_bulk_get_all() returns zero then we return success here. regards, dan carpenter > + } > + > + ret = clk_bulk_prepare_enable(devres->num_clks, *clks); > + if (!ret) { > + devres_add(dev, devres); > + } else { > + clk_bulk_put_all(devres->num_clks, devres->clks); > + devres_free(devres); > + } > + > + return ret; > +} > +EXPORT_SYMBOL_GPL(devm_clk_bulk_get_all_enable); > +
> -----Original Message----- > From: Dan Carpenter <dan.carpenter@linaro.org> > Sent: 05 March 2024 14:20 > To: Shradha Todi <shradha.t@samsung.com> > Cc: linux-clk@vger.kernel.org; linux-kernel@vger.kernel.org; linux- > pci@vger.kernel.org; linux-arm-kernel@lists.infradead.org; linux-samsung- > soc@vger.kernel.org; mturquette@baylibre.com; sboyd@kernel.org; > jingoohan1@gmail.com; lpieralisi@kernel.org; kw@linux.com; robh@kernel.org; > bhelgaas@google.com; krzysztof.kozlowski@linaro.org; > alim.akhtar@samsung.com; linux@armlinux.org.uk; > m.szyprowski@samsung.com; manivannan.sadhasivam@linaro.org; > pankaj.dubey@samsung.com; gost.dev@samsung.com > Subject: Re: [PATCH v6 1/2] clk: Provide managed helper to get and enable bulk > clocks > > On Tue, Feb 20, 2024 at 02:10:45PM +0530, Shradha Todi wrote: > > +int __must_check devm_clk_bulk_get_all_enable(struct device *dev, > > + struct clk_bulk_data **clks) { > > + struct clk_bulk_devres *devres; > > + int ret; > > + > > + devres = devres_alloc(devm_clk_bulk_release_all_enable, > > + sizeof(*devres), GFP_KERNEL); > > + if (!devres) > > + return -ENOMEM; > > + > > + ret = clk_bulk_get_all(dev, &devres->clks); > > + if (ret > 0) { > > I feel like this should be >= instead of >. There aren't any callers of this function > yet so we can't see what's in *clks at the start but it's easy to imagine a situation > where it's bad data. > Reference for this piece of code has been taken from devm_clk_bulk_get_all() which has multiple callers, so it's safe. If we make this >=, it will hold on to the devres node even though there are no clocks. > > + *clks = devres->clks; > > + devres->num_clks = ret; > > + } else { > > + devres_free(devres); > > + return ret; > > When clk_bulk_get_all() returns zero then we return success here. > Yes, we are returning success in case there are no clocks as well. In case there are no clocks defined in the DT-node, then it is assumed that the driver does not need any clock manipulation for driver operation. So the intention here is to continue without throwing error. > regards, > dan carpenter > > > + } > > + > > + ret = clk_bulk_prepare_enable(devres->num_clks, *clks); > > + if (!ret) { > > + devres_add(dev, devres); > > + } else { > > + clk_bulk_put_all(devres->num_clks, devres->clks); > > + devres_free(devres); > > + } > > + > > + return ret; > > +} > > +EXPORT_SYMBOL_GPL(devm_clk_bulk_get_all_enable); > > +
Quoting Shradha Todi (2024-03-06 04:13:03) > > > > When clk_bulk_get_all() returns zero then we return success here. > > > > Yes, we are returning success in case there are no clocks as well. In case there > are no > clocks defined in the DT-node, then it is assumed that the driver does not need > any > clock manipulation for driver operation. So the intention here is to continue > without > throwing error. Maybe we shouldn't even return the clks to the caller. Do you have any use for the clk pointers?
> -----Original Message----- > From: Stephen Boyd <sboyd@kernel.org> > Sent: 09 March 2024 06:21 > To: 'Dan Carpenter' <dan.carpenter@linaro.org>; Shradha Todi > <shradha.t@samsung.com> > Cc: linux-clk@vger.kernel.org; linux-kernel@vger.kernel.org; linux- > pci@vger.kernel.org; linux-arm-kernel@lists.infradead.org; linux-samsung- > soc@vger.kernel.org; mturquette@baylibre.com; jingoohan1@gmail.com; > lpieralisi@kernel.org; kw@linux.com; robh@kernel.org; bhelgaas@google.com; > krzysztof.kozlowski@linaro.org; alim.akhtar@samsung.com; > linux@armlinux.org.uk; m.szyprowski@samsung.com; > manivannan.sadhasivam@linaro.org; pankaj.dubey@samsung.com; > gost.dev@samsung.com > Subject: RE: [PATCH v6 1/2] clk: Provide managed helper to get and enable bulk > clocks > > Quoting Shradha Todi (2024-03-06 04:13:03) > > > > > > When clk_bulk_get_all() returns zero then we return success here. > > > > > > > Yes, we are returning success in case there are no clocks as well. In > > case there are no clocks defined in the DT-node, then it is assumed > > that the driver does not need any clock manipulation for driver > > operation. So the intention here is to continue without throwing > > error. > > Maybe we shouldn't even return the clks to the caller. Do you have any use for > the clk pointers? The intention to return the clk pointers was in the case where caller wants to manipulate a particular clock in certain conditions. They can obtain the clock pointer and use clk_set_parent, clk_set_rate on those particular clocks. But I understand that in that case users can use existing clk_bulk_get_all() API. So, should I go ahead and send v7?
Quoting Shradha Todi (2024-03-15 04:34:44) > > > > Quoting Shradha Todi (2024-03-06 04:13:03) > > > > > > > > When clk_bulk_get_all() returns zero then we return success here. > > > > > > > > > > Yes, we are returning success in case there are no clocks as well. In > > > case there are no clocks defined in the DT-node, then it is assumed > > > that the driver does not need any clock manipulation for driver > > > operation. So the intention here is to continue without throwing > > > error. > > > > Maybe we shouldn't even return the clks to the caller. Do you have any use for > > the clk pointers? > > The intention to return the clk pointers was in the case where caller wants to > manipulate a particular clock in certain conditions. They can obtain the clock pointer > and use clk_set_parent, clk_set_rate on those particular clocks. > But I understand that in that case users can use existing clk_bulk_get_all() API. > So, should I go ahead and send v7? > No, I think this is fine.
diff --git a/drivers/clk/clk-devres.c b/drivers/clk/clk-devres.c index 4fb4fd4b06bd..cbbd2cc339c3 100644 --- a/drivers/clk/clk-devres.c +++ b/drivers/clk/clk-devres.c @@ -182,6 +182,46 @@ int __must_check devm_clk_bulk_get_all(struct device *dev, } EXPORT_SYMBOL_GPL(devm_clk_bulk_get_all); +static void devm_clk_bulk_release_all_enable(struct device *dev, void *res) +{ + struct clk_bulk_devres *devres = res; + + clk_bulk_disable_unprepare(devres->num_clks, devres->clks); + clk_bulk_put_all(devres->num_clks, devres->clks); +} + +int __must_check devm_clk_bulk_get_all_enable(struct device *dev, + struct clk_bulk_data **clks) +{ + struct clk_bulk_devres *devres; + int ret; + + devres = devres_alloc(devm_clk_bulk_release_all_enable, + sizeof(*devres), GFP_KERNEL); + if (!devres) + return -ENOMEM; + + ret = clk_bulk_get_all(dev, &devres->clks); + if (ret > 0) { + *clks = devres->clks; + devres->num_clks = ret; + } else { + devres_free(devres); + return ret; + } + + ret = clk_bulk_prepare_enable(devres->num_clks, *clks); + if (!ret) { + devres_add(dev, devres); + } else { + clk_bulk_put_all(devres->num_clks, devres->clks); + devres_free(devres); + } + + return ret; +} +EXPORT_SYMBOL_GPL(devm_clk_bulk_get_all_enable); + static int devm_clk_match(struct device *dev, void *res, void *data) { struct clk **c = res; diff --git a/include/linux/clk.h b/include/linux/clk.h index 1ef013324237..a8e6b045b848 100644 --- a/include/linux/clk.h +++ b/include/linux/clk.h @@ -438,6 +438,22 @@ int __must_check devm_clk_bulk_get_optional(struct device *dev, int num_clks, int __must_check devm_clk_bulk_get_all(struct device *dev, struct clk_bulk_data **clks); +/** + * devm_clk_bulk_get_all_enable - Get and enable all clocks of the consumer (managed) + * @dev: device for clock "consumer" + * @clks: pointer to the clk_bulk_data table of consumer + * + * Returns success (0) or negative errno. + * + * This helper function allows drivers to get all clocks of the + * consumer and enables them in one operation with management. + * The clks will automatically be disabled and freed when the device + * is unbound. + */ + +int __must_check devm_clk_bulk_get_all_enable(struct device *dev, + struct clk_bulk_data **clks); + /** * devm_clk_get - lookup and obtain a managed reference to a clock producer. * @dev: device for clock "consumer" @@ -960,6 +976,12 @@ static inline int __must_check devm_clk_bulk_get_all(struct device *dev, return 0; } +static inline int __must_check devm_clk_bulk_get_all_enable(struct device *dev, + struct clk_bulk_data **clks) +{ + return 0; +} + static inline struct clk *devm_get_clk_from_child(struct device *dev, struct device_node *np, const char *con_id) {