Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:2 as permitted sender) client-ip=2620:137:e000::3:2;
Message-ID: <fc7d60ea-c48a-4666-8b69-0572ba2c66b1@samsung.com>
Date:   Wed, 15 Nov 2023 10:07:00 +0100
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
From:   Marek Szyprowski <m.szyprowski@samsung.com>
Subject: Re: [PATCH] PCI: exynos: Adapt to clk_bulk_* APIs
To:     Shradha Todi <shradha.t@samsung.com>,
        'Manivannan Sadhasivam' <manivannan.sadhasivam@linaro.org>
Cc:     jingoohan1@gmail.com, lpieralisi@kernel.org, kw@linux.com,
        robh@kernel.org, bhelgaas@google.com,
        krzysztof.kozlowski@linaro.org, alim.akhtar@samsung.com,
        linux-pci@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-samsung-soc@vger.kernel.org, linux-kernel@vger.kernel.org,
        pankaj.dubey@samsung.com
Content-Language: en-US
In-Reply-To: <000b01da178e$a43088d0$ec919a70$@samsung.com>
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 201P
References: <CGME20231009062222epcas5p36768b75c13c7c79965b5863521361a64@epcas5p3.samsung.com>
        <20231009062216.6729-1-shradha.t@samsung.com>
        <20231027134849.GA23716@thinkpad>
        <000b01da178e$a43088d0$ec919a70$@samsung.com>
Precedence: bulk

Hi!

On 15.11.2023 07:40, Shradha Todi wrote:
>> -----Original Message-----
>> From: Manivannan Sadhasivam [mailto:manivannan.sadhasivam@linaro.org]
>> Sent: 27 October 2023 19:19
>> To: Shradha Todi<shradha.t@samsung.com>
>> Cc:jingoohan1@gmail.com;lpieralisi@kernel.org;kw@linux.com;
>> robh@kernel.org;bhelgaas@google.com;krzysztof.kozlowski@linaro.org;
>> alim.akhtar@samsung.com;linux-pci@vger.kernel.org; linux-arm-
>> kernel@lists.infradead.org;linux-samsung-soc@vger.kernel.org; linux-
>> kernel@vger.kernel.org;pankaj.dubey@samsung.com
>> Subject: Re: [PATCH] PCI: exynos: Adapt to clk_bulk_* APIs
>>
>> On Mon, Oct 09, 2023 at 11:52:16AM +0530, Shradha Todi wrote:
>>> There is no need to hardcode the clock info in the driver as driver
>>> can rely on the devicetree to supply the clocks required for the
>>> functioning of the peripheral. Get rid of the static clock info and
>>> obtain the platform supplied clocks. The total number of clocks
>>> supplied is obtained using the devm_clk_bulk_get_all() API and used
>>> for the rest of the clk_bulk_* APIs.
>>>
>>> Signed-off-by: Shradha Todi<shradha.t@samsung.com>
>>> ---
>>>   drivers/pci/controller/dwc/pci-exynos.c | 46
>>> ++++++-------------------
>>>   1 file changed, 11 insertions(+), 35 deletions(-)
>>>
>>> diff --git a/drivers/pci/controller/dwc/pci-exynos.c
>>> b/drivers/pci/controller/dwc/pci-exynos.c
>>> index 9e42cfcd99cc..023cf41fccd7 100644
>>> --- a/drivers/pci/controller/dwc/pci-exynos.c
>>> +++ b/drivers/pci/controller/dwc/pci-exynos.c
>>> @@ -54,8 +54,8 @@
>>>   struct exynos_pcie {
>>>   	struct dw_pcie			pci;
>>>   	void __iomem			*elbi_base;
>>> -	struct clk			*clk;
>>> -	struct clk			*bus_clk;
>>> +	struct clk_bulk_data		*clks;
>>> +	int				clk_cnt;
>>>   	struct phy			*phy;
>>>   	struct regulator_bulk_data	supplies[2];
>>>   };
>>> @@ -65,30 +65,18 @@ static int exynos_pcie_init_clk_resources(struct
>> exynos_pcie *ep)
>>>   	struct device *dev = ep->pci.dev;
>>>   	int ret;
>>>
>>> -	ret = clk_prepare_enable(ep->clk);
>>> -	if (ret) {
>>> -		dev_err(dev, "cannot enable pcie rc clock");
>>> +	ret = devm_clk_bulk_get_all(dev, &ep->clks);
>>> +	if (ret < 0)
>> Please use !(ret) here and below to be consistent with the driver.
>>
> In this case, only negative values indicate failure. Hence we cannot use (!ret) here.
>
>>>   		return ret;
>>> -	}
>>>
>>> -	ret = clk_prepare_enable(ep->bus_clk);
>>> -	if (ret) {
>>> -		dev_err(dev, "cannot enable pcie bus clock");
>>> -		goto err_bus_clk;
>>> -	}
>>> +	ep->clk_cnt = ret;
>> Since clk_cnt is "int", you can just use it directly instead of "ret".
>>
> Thanks for this suggestion! Will take care in v2.
>
>>> -	return 0;
>>> -
>>> -err_bus_clk:
>>> -	clk_disable_unprepare(ep->clk);
>>> -
>>> -	return ret;
>>> +	return clk_bulk_prepare_enable(ep->clk_cnt, ep->clks);
>>>   }
>>>
>>>   static void exynos_pcie_deinit_clk_resources(struct exynos_pcie *ep)
>>> {
>>> -	clk_disable_unprepare(ep->bus_clk);
>>> -	clk_disable_unprepare(ep->clk);
>>> +	clk_bulk_disable_unprepare(ep->clk_cnt, ep->clks);
>>>   }
>>>
>>>   static void exynos_pcie_writel(void __iomem *base, u32 val, u32 reg)
>>> @@ -332,17 +320,9 @@ static int exynos_pcie_probe(struct
>> platform_device *pdev)
>>>   	if (IS_ERR(ep->elbi_base))
>>>   		return PTR_ERR(ep->elbi_base);
>>>
>>> -	ep->clk = devm_clk_get(dev, "pcie");
>>> -	if (IS_ERR(ep->clk)) {
>>> -		dev_err(dev, "Failed to get pcie rc clock\n");
>>> -		return PTR_ERR(ep->clk);
>>> -	}
>>> -
>>> -	ep->bus_clk = devm_clk_get(dev, "pcie_bus");
>>> -	if (IS_ERR(ep->bus_clk)) {
>>> -		dev_err(dev, "Failed to get pcie bus clock\n");
>>> -		return PTR_ERR(ep->bus_clk);
>>> -	}
>>> +	ret = exynos_pcie_init_clk_resources(ep);
>>> +	if (ret < 0)
>>> +		return ret;
>>>
>>>   	ep->supplies[0].supply = "vdd18";
>>>   	ep->supplies[1].supply = "vdd10";
>>> @@ -351,10 +331,6 @@ static int exynos_pcie_probe(struct
>> platform_device *pdev)
>>>   	if (ret)
>>>   		return ret;
>>>
>>> -	ret = exynos_pcie_init_clk_resources(ep);
>>> -	if (ret)
>>> -		return ret;
>>> -
>>>   	ret = regulator_bulk_enable(ARRAY_SIZE(ep->supplies), ep-
>>> supplies);
>>>   	if (ret)
>> You need to disable_unprepare() clocks in error path here and above.
>>
> Thanks for pointing out! Will take care in v2.


Maybe it would make sense to add devm_clk_bulk_get_all_enabled() to 
clock framework, similar to the existing devm_clk_get_enabled()?

It is really a common pattern to get all clocks and enable them for the 
time of driver operation.

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland