Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 25 Apr 2023 13:57:26 +0200
From:   Andrew Lunn <andrew@lunn.ch>
To:     Wang Zhang <silver_code@hust.edu.cn>
Cc:     Peter Korsgaard <peter@korsgaard.com>,
        hust-os-kernel-patches@googlegroups.com, linux-i2c@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] i2c: ocores: use devm_ managed clks
Message-ID: <cf543f92-af0d-4862-bea6-53a358ee9a31@lunn.ch>
References: <fc8ed989-68e6-4fd4-a818-ae077bf5e6aa@lunn.ch>
 <20230422123253.137368-1-silver_code@hust.edu.cn>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230422123253.137368-1-silver_code@hust.edu.cn>
Precedence: bulk

On Sat, Apr 22, 2023 at 08:32:53PM +0800, Wang Zhang wrote:
> If any wrong occurs in ocores_i2c_of_probe, the i2c->clk needs to be
> released. But the function returns directly in line 701 without freeing
> the clock. Even though we can fix it by freeing the clock manually if
> platform_get_irq_optional fails, it may not be following the best practice.
> The original code for this driver contains if (IS_ERR()) checks
> throughout, explicitly allowing the driver to continue loading even if
> devm_clk_get() fails.
> 
> While it is not entirely clear why the original author implemented this
> behavior, there may have been certain circumstances or issues that were not
> apparent to us. It's possible that they were trying to work around a bug by
> employing an unconventional solution.Using `devm_clk_get_enabled()` rather
> than devm_clk_get() can automatically track the usage of clocks and free
> them when they are no longer needed or an error occurs.
> 
> fixing it by changing `ocores_i2c_of_probe` to use
> `devm_clk_get_optional_enabled()` rather than `devm_clk_get()`, changing
> `goto err_clk' to direct return and removing `err_clk`.
> 
> Signed-off-by: Wang Zhang <silver_code@hust.edu.cn>

Reviewed-by: Andrew Lunn <andrew@lunn.ch>

    Andrew